THE PENROSE INQUIRY
Final Report

< Back to final report index page

Chapter 3

Statistics

Introduction

3.1 Term of Reference 4 required the Inquiry:

To investigate the systems for recording and monitoring the numbers of NHS patients in Scotland treated with blood and blood products, with particular reference to the numbers exposed to risk of infection with the Hepatitis C virus and HIV and the numbers contracting either or both such infections as a consequence of such treatment.

3.2 The National Health Service (NHS) patients who were treated with or received blood and blood products and were put at risk of infection with both HIV and the Hepatitis C virus (HCV), or HCV alone have been described in Chapter 2, Patients at Risk paragraph 2.1. This chapter attempts to estimate the numbers of patients infected with one or other or both viruses. For the purposes of this chapter, the patients are divided into four groups:

i. Patients with bleeding disorders who were infected with HCV as a result of treatment with blood products.

ii. Patients who were infected with HCV as a result of blood transfusion.

iii. Patients with bleeding disorders who were infected with HIV as a result of treatment with blood products.

iv. Patients who were infected with HIV as a result of blood transfusion.

Surgical and medical patients transfused with blood and blood components are the largest group of NHS patients exposed to infection by HCV.

3.3 Within the group of patients with blood coagulation disorders, some sub-groups, including female carriers of the disorders now recorded in the statistical registers, have become numerically significant in the past decade. Very few of these were known, and fewer still received blood products in the 1970s and 1980s. The numbers of those patients infected or potentially infected with HCV or HIV are very small indeed. They do not affect the overall picture and are not discussed separately in this chapter.^[1]

3.4 The search for reliable data for the numbers of patients treated by the NHS in Scotland with blood, blood components or blood products and thereby exposed to risks of transmission of HCV and HIV, and of the numbers who were infected, has proved for the most part to be extremely difficult. Contemporary records, whether of patients exposed to risk or of the incidence of infection, were either not maintained at all, or have become over time incomplete and unreliable. For much of the period it would not have been possible to maintain contemporaneous records of transmission of infection, given the dates when the viruses responsible for the diseases were first identified. This is particularly the case in respect of patients infected with HCV as a result of transfusion. The picture is clearer in the case of HIV infection where reports of infection arose over a relatively short and well-defined period of time. These records were recent and appear to have been reasonably reliable. Investigation of transmission was possible in many such cases.

3.5 However, even in the case of HIV, reports of the numbers of patients infected with the virus, or suffering from AIDS, in the UK generally and Scotland in particular have not always been consistent or comprehensive. Further, it became apparent during the Inquiry's investigations that information has not been disseminated equally among professional groups directly involved in patient care, other related professional groups, patient groups and the general public, with the result that there are inconsistencies among the sources of information that have been difficult, or even impossible, to resolve.

3.6 As a result of work carried out after the oral hearings, by the Scottish Haemophilia Directors and Dr Charles Hay and colleagues at the National Haemophilia Database, it is now possible to be confident of the number of patients with blood disorders who acquired HIV infection from therapy administered in Scotland. It has not been possible without some qualification to arrive at precise figures for the numbers of patients actually infected in any of the other three groups. Recorded, hard data contribute to the overall picture, but there are no comprehensive NHS records of unquestionable accuracy for either disease in any category of patients, extending over the whole of the period under review.

3.7 A short answer to Term of Reference 4 would be:

• In the case of each infection there was a period when the agent of transmission was not identified, and no diagnostic test for infection was available, so an effective recording of comprehensive data would therefore have been impossible.
• During the period in question there were no satisfactory regulatory or other central administrative systems for recording and monitoring the numbers of NHS patients in Scotland treated with blood and blood products, or the numbers exposed to the risk of infection with HCV or HIV, or the numbers contracting either or both such infections as a consequence of such treatment.
• Such provision as was devised for reporting and recording relevant data was not enforced and was ineffective.

3.8 However, notwithstanding the lack of systematic recording and monitoring of patients, it was recognised by the Inquiry that it might be possible for a retrospective analysis to be carried out, either by extraction and analysis of the material found, or using this basic data for statistical analysis and projection. The investigation has therefore been extended to include more complex study in order to determine whether relevant information can be deduced from available data using statistical modelling techniques.

Systems for reporting and recording infection

3.9 Infection of the liver with hepatotropic viruses has generally been seen as an issue only since the Second World War, partly because blood transfusion (in the broadest sense) has been practised widely only since then. It is only relatively recently, since the 1960s, that the impact on the human population of 'viral hepatitis' has been perceived at all.^[2] From 1932, there was a requirement to notify cases of 'infective jaundice' but 'infective jaundice' was defined exclusively as 'spirochaetosis ictero-haemorrhagica', indicating an 'icteric' or symptomatic condition.^[3] In 1968, that limited definition was removed, and the notifiable disease became 'infective jaundice', without further definition.^[4] One aim of this change was to assist Medical Officers of Health to obtain more precise information about the prevalence of hepatitis and the circumstances in which it was spread.^[5] Growing awareness of the prevalence of hepatitis (in the broadest sense of disease causing inflammation of the liver) in the UK, led to provision for notification of cases of 'viral hepatitis', which became reportable as an infectious disease under the Public Health (Infectious Diseases) (Scotland) Regulations 1975, which came into force on 2 April 1975.

3.10 Retrospective studies from America and modelling studies from France, in addition to data and studies from the health protection agencies, suggest that, in Western countries generally and in Scotland in particular, the rate of growth of NANB Hepatitis/HCV infection accelerated through the 1970s and 1980s, largely related to intravenous drug use.^[6] The rate of growth slowed thereafter, but the numbers infected continue to grow, and are expected to increase well beyond 2020.^[7] Until the second half of the 1980s NANB Hepatitis infection was generally thought not to be a potentially serious condition.^[8] So far as present purposes are concerned, this goes some way to explaining the lack of data on the incidence and prevalence of the disease, both generally and among haemophilia and other NHS patients. Furthermore, until discovery of the HCV virus in 1988 and subsequent epidemiological studies in the 1990s, NANB Hepatitis was thought probably to be due to more than one agent, and could not be identified by any serological test available. The background facts were not conducive to the development and implementation of an effective reporting strategy.

The Public Health (Infectious Diseases) (Scotland) Regulations 1975

3.11 Under the 1975 regulations a medical practitioner, on becoming aware that a patient was suffering from a notifiable disease, had a legal obligation to inform the chief administrative medical officer for the area health board 'forthwith' by means of a prescribed certificate.^[9] The regulations required only bare details to be given and did not address the possible or likely means by which the disease had been contracted.^[10] The chief administrative medical officer for each health board had an obligation, at the end of each week or as soon as practicable thereafter, to send to the Common Services Agency for the Scottish Health Service (the CSA) a return of the number of cases of each notifiable disease intimated to them during that week.^[11] In addition, they had an obligation to report immediately to the Chief Medical Officer any serious outbreak of any infectious disease which, to their knowledge, had occurred in their area.^[12]

3.12 The system was not effective. The regulations, which were unfortunately introduced just before the several distinct forms of viral hepatitis and their clinical manifestations began to be understood, did not promote epidemiological understanding of hepatitis. The existence of an NANB Hepatitis virus or viruses (in time, principally Hepatitis C) had only recently been postulated (in 1974) when the regulations came into force in 1975. Professor David Goldberg's group at Health Protection Scotland (HPS), (which collects and analyses statistical and epidemiological data relating to blood-borne viruses, sexually-transmitted infections and other diseases and infections), and its predecessor the Scottish Centre for Infection and Environmental Health (SCIEH), concluded that such data as were returned and registered were not reliable.^[13] The reports required of medical practitioners did not contain adequate information on the prevalence of infection.^[14] Professor Goldberg explained further that, following the 1991 introduction of reliable testing and screening for HCV, it soon became clear that the numbers reported by clinicians bore no relationship to the actual data HPS recovered from laboratories testing samples for evidence of infection. They found that data recorded by virology laboratories were over 90% complete and accurate, undermining the reliability of the different and inconsistent data held on the register. Use of the data reported under the regulations would have hindered HPS in developing its understanding of the epidemiology of HCV infection.^[15]

3.13 The regulations could never have been effective in securing full reporting of the incidence of viral hepatitis. There were several problems with the regulatory regime. 'Viral hepatitis' is not a specific disease. It can be caused by a number of hepatitis viruses.^[16] Until well into the reference period not enough was known about the forms of viral hepatitis to enable effective reporting requirements to be developed and enforced. Professor Goldberg said that, in reality, clinicians rarely reported the clinical entity, viral hepatitis, to health boards.^[17] Interpretation of the regulations as amended from time to time appears to have given rise to difficulties.

3.14 NANB Hepatitis/Hepatitis C infection could not be diagnosed without a definitive test for the virus or its antibody. Until 1989-90 there was no diagnostic blood test available generally, to demonstrate prior exposure to HCV. In the UK routine testing was not introduced until September 1991. When the regulations came into force in 1988, overt clinical jaundice was thought to be characteristic of viral hepatitis.^[18] In time it came to be recognised, however, that most patients who contracted NANB Hepatitis/HCV did not develop jaundice and that transmission of infection was largely a silent event which would not have alerted clinicians to the obligation to report.

3.15 Nevertheless, viral hepatitis remained in the scheduled list of notifiable infectious diseases under the Public Health (Notification of Infectious Diseases) (Scotland) Regulations 1988.^[19] These Regulations consolidated, with amendments, the 1975 Regulations relating to the notification and prevention of infectious disease.

European law

3.16 As regards blood products, Directive 2001/83/EC of the European Parliament and of the Council of 6 November 2001 on the Community Code relating to medicinal products for human use, included an obligation on Member States to take all appropriate measures to 'encourage' doctors and other healthcare professionals to report to a competent authority suspected adverse reactions.^[20] The Directive also obliged Member States to establish a pharmacovigilance system whereby those who were authorised to market a medicinal product were required to report to the competent authority all suspected serious adverse reactions to the product.^[21] Directive 2001/83/EC did not apply to whole blood, plasma or blood cells.^[22]

3.17 As regards whole blood and blood components (ie red cells, white cells, platelets and plasma), Directive 2002/98/EC introduced an obligation on Member States to ensure that any serious adverse events or reactions relating to the quality or safety of blood or blood components were notified to a competent authority.^[23] The Directive was implemented in the UK by the Blood Safety and Quality Regulations 2005,^[24] which required 'blood establishments'^[25] and hospital blood banks to notify the Secretary of State of any serious adverse events or reactions relating to the quality or safety of blood and blood components.^[26]

3.18 Before the Blood Safety and Quality Regulations 2005 came into force, there were no effective regulatory obligations on clinicians or others to report NANB Hepatitis/HCV or HIV infection from blood transfusion. The 2005 Regulations improved the regulatory regime, but came into force after the events that are relevant to this report.

3.19 The 2005 Regulations reflected growing knowledge of the characteristics of transmissible diseases and of the difficulty of tracing infected patients who had received transfusions many years previously, when there may have been differing standards, for example, between peripheral hospital blood banks and transfusion centres. While transfusion centres had been inspected for many years by the Medicines Inspectorate, hospital blood banks had not. The regulations stipulated for the first time that hospital blood banks and transfusion centres should be brought fully within the Medicines and Healthcare Regulatory Authority (MHRA) framework, and the role of the MHRA was strengthened accordingly. The provisions recognised a gap in the traceability of previous donations and therefore of the patients who received them. Traceability, with a proper 'trail', became entrenched in the new regime and an obligation was placed on clinicians, blood banks and the transfusion service to report formally to the MHRA when there was a serious adverse event in transfusion, such as transmission of virus infection.^[27]

Public Health etc. (Scotland) Act 2008

Hepatitis C

3.20 Since 2005, and following public consultation, the Public Health etc. (Scotland) Act 2008 added further requirements, effective from 1 January 2010.^[28] The Act involved significant reform and modernisation of the public health/health protection regime in Scotland. Testing laboratories are now required to report to health boards information about infection of individuals testing positive for hepatitis viruses.^[29] For Hepatitis B and Hepatitis A, that is important because there is a vaccine available for these infections and close contacts can be identified and offered vaccination. In the case of Hepatitis C the public health importance of reporting is less clear. There was, and is, no vaccine available. For that reason the close contacts of individuals who were infected with Hepatitis C are generally not contacted.^[30] However, there are now increasingly effective treatments available for HCV which may be offered to those found to be infected, which will generate demand for more effective tracing of people possibly infected.

3.21 In the 2008 Act, Hepatitis A, B, C and E viruses are scheduled as 'notifiable organisms', and the generic term 'viral hepatitis' has been removed from the list of notifiable diseases.^[31] The Act requires the director of a diagnostic laboratory which has identified a notifiable organism to notify, within 10 days, the health board in whose area the diagnostic laboratory is situated, as well as the CSA (presumably, in practice, Health Protection Scotland), of certain information.^[32] As in the previous legislation, only bare details require to be notified and there is no duty to include the possible or likely means by which the organism was transmitted.^[33] Fundamentally, notification depends on there being a listed organism identifiable by laboratory testing. Where diagnosis of disease depends on a laboratory test, linking surveillance of diseases to the laboratory testing process is likely to be a much more satisfactory procedure than that under the previous notifiable diseases regime. HCV infection is not a notifiable disease reportable by medical practitioners in terms of section 13 of the Act.

HIV/AIDS

3.22 HIV never became a reportable disease under the regulations at any stage.^[34] The AIDS (Control) Act 1987 required each health board to make periodic reports to the Secretary of State of information including the numbers diagnosed with AIDS, the numbers with AIDS who had died and the numbers of positive HIV antibody test results. The reports by health boards did not involve identification of individual patients and would have been based on the board's internal statistical records.^[35]

3.23 In 2006 the Scottish Executive consulted on public health legislation in Scotland. The analysis of responses to that consultation confirmed strong support for the repeal of the AIDS (Control) Act^[36] as well as a majority view in favour of including STIs (of which HIV is an example) in the new notification system.^[37] The arguments against a notification requirement for HIV or any other STIs appear to have been based on considerations of confidentiality and concern that it would have an overall negative effect on their surveillance. The point is not referred to in the Policy Memorandum to the subsequent Public Health (Scotland) Bill nor was it specifically explored during the Health and Sport Committee hearings, although there was some related discussion.^[38]

3.24 Part 2 of the Schedule to the Bill as introduced included on the list of notifiable organisms to be reported by diagnostic laboratories, 'any other clinically significant pathogen found in blood'. In its Stage 1 Report on the Bill, the Health and Sport Committee expressed its concern that this form of words could give rise to problems if directors of laboratories were, for example, to decide that HIV should be reported. In particular, it would compromise reassurances being given to people with STIs that confidentiality would be maintained. The words in question were removed by way of amendment during the passage of the Bill. The 1987 Act, in so far as it extended to Scotland, was repealed by the 2008 Act with effect from 1 January 2010. Under the 2008 Act, HIV/AIDS remained not listed either as a notifiable disease or as an organism.

3.25 The regime established by the 2008 Act is relatively untried. Some parts of it are novel and it will take time to develop experience of it. For example, the scope of the requirement to notify 'Health Risk States' in particular is novel and may become clearer in time. The Inquiry has not investigated the operation of the current legislation, and it would be inappropriate to comment on it further.

3.26 As matters stood during the material period up to 1991, the provisions for notification were not effective as a means of informing relevant medical authorities of the prevalence of infection with any relevant disease. For the purposes of the Inquiry it has been necessary to use other sources of information to attempt to paint a picture of the position as it emerged.

Patients with bleeding disorders infected with Hepatitis C as a result of treatment with blood products

3.27 The first group of patients to be considered is referred to in paragraph 3.2, item i, and consists of patients with bleeding disorders, primarily haemophilia, who were exposed to infection with Hepatitis C through their treatment with blood products. Determining the number of patients exposed to HCV as a result of their treatment by the NHS in Scotland is not straightforward because several sources of data are available, each reporting slightly different numbers. It has been possible to compare these and, to a great extent, reconcile the differences.

Databases of patients with bleeding disorders

3.28 From the early 1950s a UK haemophilia register of anonymised data was established and maintained in Oxford in the form of a card index. The register was initially funded by the Medical Research Council.^[39] It was taken over by the United Kingdom Haemophilia Centre Directors' (later 'Doctors') Organisation, (the UKHCDO) in about 1968. At the outset, there were widely-held reservations among haemophilia clinicians about the registration of information about patients. These reservations often related to confidentiality. The issue came to a head at a meeting of the UKHCDO in November 1979. It was resolved that the project should continue.^[40] The database comprises information collected from all UK haemophilia centres. In practice, Scottish Haemophilia Directors returned anonymised data to Oxford as part of the national database and that arrangement has continued.

3.29 In the meantime, it was proposed that the Scottish Home and Health Department (SHHD) should take over the similar register of patients with bleeding disorders that had been maintained for Scotland. There were objections, however, concerning the propriety of a central register containing the names and addresses of patients. Some patients had been concerned that there might be leakage of information, and it was agreed at a joint meeting of the SNBTS and Haemophilia Directors on 24 January 1977, that this desire for confidentiality should be respected. The cards were distributed to the Regional Haemophilia Directors to be held locally.^[41]

3.30 It appears that thereafter data were held by Haemophilia Centre Directors and Regional Transfusion Directors in Scotland on locally devised systems which were not necessarily compatible as between transfusion practitioners and haemophilia clinicians.^[42] It is likely that the Scottish Haemophilia Centres had held local databases of patients prior to 1977 and continued with these without modification in the new regime.

3.31 Thus, in the 1970s there came to be two parallel databases of patients: informal local indices at each hospital where blood disorder patients were treated and a national, increasingly anonymised, database maintained by the UKHCDO to which each treatment centre sent information. This state of affairs was formalised at UK level at the end of 1979. Locally, Haemophilia Directors and other clinicians continued to hold data relating to their own patients.

3.32 There were, and remain, limitations on the accuracy of the UKHCDO data. Capturing data concerning all of the patients proved to be a slow business at least until 1980. Quite apart from the resistance to disclosing information, already noted, patients often exercised a choice to be managed at local hospitals even where there was not a dedicated haemophilia facility. Some needed persuasion to accept the inconvenience of going to a bigger centre, which might involve travelling considerable distances, even though care might be better there.^[43] As explained by the UKHCDO, the rapid increase in numbers of registered patients in the first decade after the database was established, suggested that early registration data may be incomplete.^[44]

3.33 The problem relating to the earlier years was illustrated by Dr Hay with reference to Glasgow. Between 1970 and 1980, the Glasgow registrations for haemophilia at the UKHCDO, unadjusted for duplicate registrations, are as set out in Table 3.1.^[45]

Table 3.1: Count of Glasgow patients at five-yearly intervals

		Haemophilia A	Haemophilia B	Totals
1970	Royal Infirmary Royal Hospital for Sick Children	71 3	7 0	78 3 81
1975	Royal Infirmary Royal Hospital for Sick Children	102 8	15 0	117 8 125
1980	Royal Infirmary Royal Hospital for Sick Children	196 55	52 14	248 69 317

3.34 The increase in each case was far in excess of the birth-rate or any other factor indicating real growth over the period. This was a period when mortality was beginning to be offset by the benefits of concentrate therapy, and that may to some extent have increased the numbers of patients remaining on the register. However, few data were submitted during the early part of the period. Haemophilia care was fragmented in the west of Scotland at this time. Many patients were managed at smaller hospitals around the south west of Scotland at the beginning of the period but, with the advent of increasingly effective treatment, they gradually gravitated towards the centre. The Glasgow pattern is highly unlikely to have been unique, and the lack of reliability of early data must have been common.

3.35 It appears reasonable to avoid speculation as to probable numbers and to treat the 1970 registrations as wholly unreliable in estimating numbers of patients treated and therefore exposed to some level of risk. It also appears reasonable to accept that the number of 1975 registrations would require to be modified to reflect the probability that the process of gathering patients into the reporting system was continuing and had not completely captured all relevant patients. The numbers recorded in the UKHCDO database from 1980 (as adjusted) are reasonably firm.

3.36 A continuing problem relates to the 'hub and spoke' care model which was developed and which has resulted in some patients being registered and managed at more than one centre.^[46] The data required to be adjusted to exclude duplications. Investigation by the UKHCDO subsequent to the publication of the Preliminary Report uncovered inaccuracies in the data provided to, and published by, the Inquiry. The Inquiry is grateful to the UKHCDO for the work carried out to check the database, internally and against extant original records, and to provide adjusted data. The database, which has improved over time and now provides a wealth of relevant data, has become the primary source of published information on exposure to risk among this class of NHS patients, and their treatment thereafter.

UKHCDO and Scottish Haemophilia Directors' data on patients with bleeding disorders

3.37 The primary sources of published up-to-date data are the UKHCDO Bleeding Disorder Statistics for April 2010 to March 2011, and April 2011 to March 2012. The adjusted numbers of patients with clotting factor deficiencies in Scotland known to the UKHCDO and cross-checked with Scottish Haemophilia Centres, at April 2012 (subject to the reservations about the accuracy of data before 1980 already noted) are as set out in Table 3.2.^[47]

Table 3.2: Count of registered patients: All Scottish centres

	Haemophilia A	Haemophilia B	Females with F.VIII deficiency	Females with F.IX deficiency	vWD	Total
1970	159	17	0	0	2	178
1975	274	47	0	0	2	323
1980	408	88	5	3	44	548
1985	443	95	11	6	135	690
1990	462	102	20	9	185	778
1995	475	116	34	12	394	1031
2000	400	107	50	23	616	1196
2005	398	107	63	32	827	1427
2010	414	116	80	42	976	1628
2011	409	113	81	41	995	1639

3.38 The pattern of individuals registered with each condition is illustrated in Figure 3.1.

Figure 3.1: Individuals with bleeding disorders registered in Scotland

3.39 Setting aside early inaccuracies, the numbers of Haemophilia A and Haemophilia B patients registered are reasonably consistent over the period from 1980. The trends in registered female and von Willebrand's disease (vWD)^[48] patients cannot be explained in terms of natural growth in prevalence of the diseases. Three factors may have contributed:

• Increasing sophistication in the diagnosis of mild forms of the diseases.
• Greater diligence in reporting to UKHCDO.
• A growing tendency, in operating the hub and spoke model, for peripheral haemophilia centres and other, non-specialist, hospital units to refer patients with these less serious conditions to haemophilia centres in the major cities.

Isolating the numbers for individuals with Haemophilia A and B, the pattern is illustrated in Figure 3.2.

Figure 3.2: Individuals with Haemophilia A and Haemophilia B registered in Scotland

3.40 These data are accepted as accurate for the purposes of this report. The data for patients with Haemophilia A and Haemophilia B provide a reasonable view of the majority of patients potentially exposed to risk in and after 1980, since registration would have followed attendance at a haemophilia centre, whether the patient was treated or not. The numbers cannot be treated as completely accurate, since it is likely that some patients (particularly mildly and moderately affected patients) still preferred to receive the required treatment locally, rather than at what may have been a remote haematology clinic. Consequently, such patients were not referred to a haemophilia centre, and were not registered with UKHCDO.

3.41 The UKHCDO Bleeding Disorder Statistics for April 2010 to March 2011^[49] show the total numbers of UK patients currently on the register for that year, and provide a basis for comparison of the Scottish data in Table 3.2 with the position in the rest of the UK, as set out in Table 3.3.^[50]

Table 3.3: Count of registered patients: All Scottish centres and all other centres - 2011

	Haemophilia A	Haemophilia B	Females with F.VIII deficiency	Females with F.IX deficiency	vWD M + F	Total
All centres	5367	1133	1116	345	9112	17,073
Scotland	409	113	81	41	995	1639
Centres other than Scotland	4958	1020	1035	304	8117	15,434
Scotland/all centres %	7.62	9.97	7.26	11.88	10.92	9.6
Scotland/rest of UK %	8.25	11.08	7.83	13.49	12.6	10.62

3.42 It is not possible to make a similar comparison for earlier years, since the Scottish data before 2011 have been fully revised (for the purpose of informing this Inquiry) to avoid duplication, while the figures for the rest of the UK have not. The figures available for the UK as a whole are set out in Table 3.4.

Table 3.4: Count of registered patients: All UK centres

	Haemophilia A	Haemophilia B	von Willebrand's disease	Total
1987	5195	982	2215	8392
1990	5339	1055	2688	9082
1995	5413	1133	4105	10,651
2000	5385	1165	5955	12,505
2005	6012	1276	7718	15,006
2009	6480	1372	9265	17,117

3.43 The lack of fit of the von Willebrand's data with the data for Haemophilia A and B is as clear for the UK as it is for Scotland. The trends in growth of registered von Willebrand's disease appear generally to reflect the sensitivity of the results to variations in reporting and other factors referred to above. There is no other rational explanation for the apparent explosion in growth of the disease in Scotland, or in the UK as a whole. It is known that, historically, the disease was under-diagnosed. It is now recognised to be the most common bleeding disorder, but it is usually a mild disorder. Most patients were, and still are, more likely to be managed locally and outside the specialist haemophilia service. In that case they may not be reported to the National Haemophilia Database at all.^[51] These data continue to be revised. For present purposes they are accepted as an accurate reflection of the information available to the UKHCDO and the Scottish Haemophilia Directors.

Numbers of bleeding disorder patients infected with Hepatitis C

UKHCDO - Initial research

3.44 All patients with coagulation disorders were exposed to risk if they required treatment with blood, blood components or blood products, before the introduction of routine screening of blood donations for HCV and effective virus inactivation. Cryoprecipitate was commonly associated with transmission of HCV infection. Factor concentrates were almost universally associated with transmission of HCV. This was the case whether the concentrates were produced by commercial companies or by the NHS and whether a patient received a single treatment or multiple treatments.^[52] An SNBTS Factor IX concentrate, effectively heat-treated to inactivate the virus, was released in October 1985. By 1987 effective virus inactivation had been introduced for PFC's Factor VIII concentrate, and manufacture of concentrates by the older processes had stopped. In each case, until effective heat treatment was introduced it is likely that almost all haemophilia patients who received treatment in the 1970s and most of the 1980s with factor concentrates made from the large pools of plasma, contracted Hepatitis C.

3.45 Data of patients' use of factor concentrates were held in the National Haemophilia Database, and the UKHCDO provided an estimate of the number of patients exposed to Hepatitis C on the basis of their exposure to concentrate administered in a Scottish Haemophilia Centre. A UKHCDO statistics report was produced in April 2012 following additional work by the organisation and by the individual haemophilia centres in Scotland to enter additional data into the database, to correct inaccuracies in data previously entered and to cross-check all of the data entered against the UKHCDO's paper archive and against the information held by the Scottish centres.^[53] It was acknowledged that there were weaknesses in the exercise. Quite apart from inherent problems leading to duplication of entries, the data were only as good as the submissions on which they were based. Further, in his introduction to the updated report of April 2012, Dr Hay stated:

Inevitably, given the time that has elapsed since the period of interest, there remain gaps and many centres no longer have records stretching that far back and have not been able to cross-check everything. Nevertheless, we believe that the data presented is now as complete and as accurate as we are able to make it.^[54]

3.46 The UKHCDO extracted data from the National Haemophilia Database as adjusted, which indicated that 296 living patients and 216 deceased patients treated at Scottish centres had been exposed to HCV, giving a total of 512.^[55] Within that total, research uncovered a number of duplications. Excluding these, the adjusted totals provided by the UKHCDO were 254 living patients and 193 deceased patients, a total of 447. That was the position at April 2012 when the UKHCDO produced its updated report.^[56] Table 7 in the report provided a diagnostic breakdown of patients known to the UKHCDO as at 2011, both alive and dead, who were thought to have been infected with Hepatitis C by virtue of exposure to clotting factor concentrates, prior to effective virus inactivation (1987). The data are summarised below in Table 3.5.

Table 3.5: Estimate of the number of patients exposed to Hepatitis C based on historical clotting factor concentrate exposure from a Scottish Haemophilia centre: April 2012 Survey

Diagnosis	Alive	Dead
Haemophilia A	165	158
Haemophilia A with liver transplant	1	3
Acquired Haemophilia A	2	3
Haemophilia B	58	19
Females with VIII deficiency	7	0
Females with IX deficiency	4	0
von Willebrand's disease	16	10
Temporary coagulation defect, now normal	1	0
Totals	254	193
Total number exposed	447

3.47 These numbers reflected the adjustments to UKHCDO raw data required to exclude double counting where patients were treated at more than one centre and returned by each.^[57] The total number of deaths from all causes in the population, 193, has been cross-checked with the Office of National Statistics. The number of liver-related deaths appears to be 21.^[58]

3.48 The figures tabulated are based on known concentrate exposure, believed to be solid data. The strengths of the exercise are:

• The data were collected contemporaneously rather than retrospectively and so the documentary evidence on which they were based was as complete as it was ever going to be. In some cases the original local documentation had been destroyed and the National Haemophilia Database was the only available evidence.
• The Database was compiled from data for each patient submitted by the managing centre.
• The UKHCDO had data on patients who died before the advent of testing and data on patients subsequently lost to follow-up for which there was now no information available directly from the centres.^[59]

3.49 Dr Hay commented that the estimate of 447 patients exposed to Hepatitis C on the basis of concentrate therapy was probably an under-estimate of the total population at risk. It did not include provision for the number of patients likely to have been exposed through treatment with blood, blood components or cryoprecipitate, rather than concentrates.^[60] In oral evidence, Dr Hay expanded on this:

The other thing you need to recognise about these figures is that they are likely to be conservative because they are based on our records of concentrate exposure and some patients may only have been treated with blood components, plasma or cryoprecipitate, and we feel that our data on that exposure may be incomplete ...

So we have assumed that everyone who has had concentrate will have been exposed to Hepatitis C if they were treated during the period of risk. But they may in fact have contracted Hepatitis C before their first exposure to concentrate from treatment with blood components, particularly if they have a severe bleeding disorder and require regular treatment.^[61]

3.50 The results were not presented as definitive.

Scottish Haemophilia Centre Directors

3.51 The Scottish Haemophilia Centre Directors provided the Inquiry with separate estimates of the number of patients registered with Scottish Haemophilia Centres who were infected with the Hepatitis C virus as a result of their treatment with blood products (coagulation factor concentrate, cryoprecipitate or fresh frozen plasma) in February 2011 and March 2013. Dr Campbell Tait, Director of the Haemophilia Centre at Glasgow Royal Infirmary, explained that in February 2011 the Scottish Directors used UKHCDO data relating to patients at risk of having contracted Hepatitis C as a result of their treatment between 1970 and 1988,^[62] and HCV status data from patient records held at Scottish Haemophilia Centres.^[63] The Scottish Haemophilia Directors excluded from the UKHCDO data any patient whose first treatment was outwith Scotland, any patient known to have tested negative for Hepatitis C post-1991 (76 patients) and any patient whose sole treatment at a Scottish Haemophilia Centre during the relevant period was with non-plasma based products (for example synthetic agents such as Desamino-D-arginine-vasopressin (DDAVP)). The Directors then added a small number of patients known by Scottish Haemophilia Centres to be HCV positive, most likely from treatment in Scotland, but who did not appear on the list from the UKHCDO.

3.52 In summary, comparison of the outcomes of the Scottish Haemophilia Centre Directors' study in February 2011^[64] and the April 2012 UKHCDO survey appeared to show material differences in the numbers of HCV-infected patients with Haemophilia A or B and patients with von Willebrand's disease. This information is set out in Table 3.6.

Table 3.6: Comparison of the Scottish Haemophilia Centre Directors' study and the UKHCDO survey

Diagnosis	UKHCDO survey: April 2012	Scottish Directors' study: February 2011
Haemophilia A	332	303
Haemophilia B	77	68
von Willebrand's disease	26	62
Other coagulation disorders	12	26
Total	447	459

3.53 Dr Tait commented on weaknesses in the Scottish Directors' study at that stage:

Determining whether patients remaining on the list, who have never had an HCV antibody test undertaken, had NANB or HCV is very difficult. Some of the patients are dead and have no remaining blood sample that could be tested for HCV. Otherwise clarification will require detailed review of historical medical case notes (which in many centres no longer exist). It is possible that some of these patients have undiagnosed HCV .... Therefore it is the intention of the Scottish Haemophilia Centres to, where possible, trace these 'unknowns' and suggest HCV testing where appropriate.

Some of these patients only had a very few treatments in Scotland, and were probably visitors who if they did have HCV would likely have contracted it outwith Scotland - but we have insufficient information to be certain.

Therefore the final list supplied to [the Inquiry] represents the estimated maximum number of bleeding disorder patients who contracted HCV from treatment in Scotland.^[65] The accuracy of the data is limited by the assumptions made. There are small numbers of patients who received PFC Factor VIII before July 1987 who are HCV negative and the numbers of patients infected as a result of cryoprecipitate therapy only, are likely to be overstated.^[66]

3.54 HCV negative patients would comprise patients exposed to the virus but were not infected, and patients who were infected but had cleared the virus. In their methodology statement, the Scottish Haemophilia Directors indicated that they had assumed that all patients treated with coagulation factor concentrates or cryoprecipitate prior to July 1987 contracted Hepatitis C, unless they had subsequently been tested negative.^[67] On that basis, up to 62 von Willebrand's patients may have contracted HCV in Scotland. The UKHCDO estimated that 26 von Willebrand's patients treated at Scottish Haemophilia Centres contracted HCV infection.^[68]

3.55 In addition to estimating the maximum number of NHS coagulation disorder patients infected with Hepatitis C, the Scottish Haemophilia Directors provided an estimate of the minimum number infected as a result of treatment in Scotland. In their February 2011 study that minimum number was 314, on the basis that 314: 'represents the numbers that we know who are or have been Hepatitis C antibody positive, plus a small number who were never tested but we have evidence that they clinically suffered an episode of non-A non-B hepatitis'.^[69] That estimate was therefore based on hard data.

3.56 The totals advised by the UKHCDO and the Scottish Haemophilia Directors at this stage were broadly consistent with the advice received by Lord Ross' Expert Group^[70] that approximately 500 haemophilia patients were likely to have become infected with Hepatitis C as a result of treatment in Scotland with blood products.^[71] However, the initial exercises not only produced figures that differed numerically; more significantly it identified differences in the nature of the data available and the assumptions used. Dr Hay of the UKHCDO and Dr Tait, representing the Scottish Haemophilia Directors, were asked to reconsider their evidence. They did so and each produced further information.^[72]

3.57 Dr Tait's statement of 8 January 2013, revealed differences in underlying data among the various estimates provided. The UKHCDO based the estimate of likely HCV-infected patients (447) on the assumption that the virus was transmitted on the patient's first exposure to clotting factor concentrates. The UKHCDO database did not include information on patients' known HCV status, positive or negative. In his evidence, Professor Goldberg had explained that HPS was aware of 351 individuals who had received blood factor concentrates and who had been diagnosed as Hepatitis C antibody positive.^[73] These data related solely to bleeding disorder patients in Scotland who had tested positive for anti-HCV and who were likely to have acquired infection through blood product exposure. The HPS data did not distinguish patients according to the country in which it was likely they were infected, and did not take account of any patients who had died before antibody testing became available in about 1990. A recent review of the Scottish Haemophilia Directors' data had shown that among the cohort of 351 infected patients identified in Scotland there were 46 who were thought to have contracted HCV from prior treatment outwith Scotland. The Scottish Haemophilia Directors' own exercise had revealed 15 patients not included in the UKHCDO list; accordingly the UKHCDO figure could be adjusted to 462, which was similar to the Directors' figure of 459.

3.58 Having reviewed the data in 2013, Dr Tait advised on the range:

• A minimum number of 300 to 305 bleeding disorder patients known to have contracted HCV (as assessed by HCV antibody positivity) through haemostatic treatment in Scotland.
• An estimated upper figure of 459 to 462, based on the Scottish Haemophilia Directors' estimate of 459 patients and the UKHCDO figure of 447 increased by 15 patients not on the UKHCDO list who were known to the Scottish Directors to be anti-HCV positive, most likely due to treatment in Scotland.^[74]

3.59 On 26 March 2013 the Inquiry received an updated version of the Scottish Haemophilia Directors' calculations as prepared for publication.^[75] The Inquiry is grateful for the additional work carried out. In its final form, the paper reported that 455 patients with bleeding disorders were estimated to have been infected with HCV by coagulation factor provided by NHS Scotland before 1989. Of these, 440 were agreed with the updated UKHCDO database and 15 further patients were known to Scottish Haemophilia centres but did not appear in the list prepared by UKHCDO.

3.60 In order to chart outcomes after infection, Khan and colleagues reported on a second cohort of 302 patients, of whom 255 were included in the revised total of 455 patients infected in Scotland and 47 were infected elsewhere but followed up in Scotland.^[76] Of the 302 patients, 293 had Polymerase Chain Reaction (PCR) testing performed and of that 293, 51 patients (17.4%) cleared the infection naturally. Of the group of 293 patients who were tested, 243 were PCR positive at some stage. One of those 243 patients was documented as being positive on one occasion and tested negative thereafter despite having no treatment, giving a total of 51 patients with natural clearance. Further, of those 243 patients, 103 had by February 2012 completed combination anti-viral treatment with ribavarin and alpha-interferon or pegylated interferon. The overall sustained viral response for combination therapy was 38.8%.^[77] Fourteen of the 302 (4.6%) had a diagnosis of hepatocellular cancer, and 10 of those patients had died at February 2012. By the same date, 11 patients had undergone liver transplantation (seven for liver failure and four for liver cancer) and seven of them were still alive.^[78]

UKHCDO - Further research

3.61 Following the 2012 report (and as anticipated in it), the UKHCDO began a more detailed and systematic HCV look-back study, partly stimulated by the Inquiry. In collaboration with each haemophilia centre, including those in Scotland, data were collected from haemophilia centres on individual patients' Hepatitis C test results. In a letter dated 16 January 2013,^[79] Dr Hay explained that at January 2013, and with the UKHCDO HCV look-back still incomplete, a very preliminary overall estimate of the numbers infected with HCV should first include the 447 patients (alive and dead) exposed to infection by concentrate, as set out in the earlier exercise. Dr Tait's review then noted that there were 15 patients known by the Scottish Haemophilia Directors to be HCV positive, most likely due to treatment in Scotland, who were not included in the UKHCDO list of 447 patients. Adjusting for those increased the UKHCDO figure to 462. The amended and additional data provided by the UKHCDO was not available when the Scottish Haemophilia Directors updated their estimates as set out in the paper by Khan and colleagues.

3.62 Dr Hay provided a further table showing a diagnostic breakdown of 85 Scottish patients exposed only to blood components (plasma, cryoprecipitate and blood or platelet concentrates) and not exposed to (factor) concentrates during the period of risk.^[80] Not all of the patients in this group would have become infected with HCV.

Table 3.7: Estimate of the number of patients in each diagnostic group exposed to Hepatitis C based on historical plasma exposure from a Scottish Haemophilia centre (January 2013; based on first 20% data only)

	Alive	Dead
Haemophilia A	18	9
Haemophilia B	1	2
Females with Factor VIII deficiency	4	0
Females with Factor IX deficiency	3	0
von Willebrand's disease	33	6
Factor VII deficiency	1	0
Factor IX deficiency	1	0
Factor XII (Hageman) defect	1	0
Factor XIII deficiency	1	0
Fibrinogen deficiency	1	0
Platelet defects (misc)	2	0
Unclassified	0	1
Temporary coagulation defect, now normal	1	0
Total	67	18

3.63 On the basis of data available in the look-back exercise as a whole, Dr Hay estimated that patients selected in this way would have an 18.6% probability of HCV infection. This would add a further 16 patients to the UKHCDO number. Of that group of 16, chronic HCV infection would be expected in 10-12.^[81] Whilst this approach permitted him to estimate the proportion of that group exposed to HCV, it did not permit him to identify individual patients likely to have been infected.^[82] This look-back study was ongoing at the time and Scottish centres had by then submitted only 20% of the data requested. More accurate data might therefore become available on completion of the exercise.^[83]

3.64 As noted already, it is highly likely that almost all of the patients exposed to risk in the course of their treatment with factor concentrates before effective heat treatment was introduced, will in fact have had Hepatitis C virus transmitted to them at some time. However, current understanding of the natural history of infection with HCV indicates that a percentage of patients will either not have developed an HCV infection or will have cleared the virus spontaneously. Where that is due to a protective immunity or to the individual's other genetic characteristics, the same outcome will follow a subsequent infection.^[84] Estimates vary, as indicated by the evidence noted so far.

Discussion of the reported estimates

3.65 It is still not possible to achieve a satisfactory reconciliation of all available sources of data. The additional evidence of Dr Hay and of Dr Tait is accepted as providing the best estimates that can be made on the sources of information available. They provide a measure of reconciliation of the upper figures and therefore an indication of the likely extent of the problem. However, each has stated that the exercises now reported are continuing. The UKHCDO continues to refine its exercise. The Scottish Haemophilia Directors group continues to seek out and assess patients who have or may have received coagulation treatments prior to 1989-91 and who remain untested for HCV. It is a slow ongoing process.

3.66 In the circumstances it is important to avoid underestimating the prevalence of HCV infection among blood disorder patients. While some percentage of those at risk may have avoided infection and others will have cleared infection spontaneously, these numbers cannot be known with any certainty. It seems safer to proceed on the basis that all of those at risk were infected. The UKHCDO estimate (as adjusted in conjunction with the additional data provided by Dr Tait) of 462 patients infected by concentrates is accepted on that basis. The number infected by blood components and plasma products is speculative at this stage. It is possible to estimate a number arithmetically, but any exercise is based on the selection of assumptions required for the calculation. The data gathered to date (20% of the data sought) are not a random sample and cannot be assumed to represent the whole class. So far, it appears that 16 patients can be added to the 462. It is not known what the remaining 80% of data will reveal and no provision is therefore made for any additional numbers of patients which may be generated.

Patients with bleeding disorders known to have been infected with Hepatitis C

Health Protection Scotland

3.67 Agencies other than UKHCDO recorded lower numbers of patients said to have been infected with HCV. As previously observed, Professor Goldberg gave evidence that HPS is aware of 351 individuals who had received blood factor concentrates and been diagnosed as Hepatitis C antibody positive.^[85] For each of these individuals there was no information to indicate that blood factor concentrates had been received outside Scotland.

3.68 Dr Tait suggested that the HPS data could be reconciled with the Scottish Haemophilia Directors' data. A review carried out in February 2012 by Dr Henry Watson of the Directors' 2007 review of HCV and its treatment in Scotland had identified, within the cohort of HCV-positive bleeding disorder patients, 46 persons who were thought to have contracted HCV from prior treatment outwith Scotland.^[86] Adjusted for these patients, the HPS figure was reduced to 305, a close similarity to the Scottish Haemophilia Directors' figure for bleeding disorder patients known to be HCV positive from treatment in Scotland.^[87]

3.69 That is accepted. Yet it indicates that the HPS database presently captures, and is able to capture, only part of the relevant data. It provides clear evidence of the numbers found to be infected on testing at virology laboratories reporting to HPS. Of necessity that cannot cover patients who were never tested while they were alive. As shown in Table 3.5, UKHCDO data suggest that, by April 2012, 193 Scottish blood disorder patients exposed to clotting factor concentrates before 1988 had died. Some might have been collected by retrospective testing of stored samples, but that would be of assistance only if there had been a systematic study of stored samples. There is no evidence of that having happened. Finally, as Professor Goldberg indicated, clinical information available to HPS is very limited in respect of the underlying diagnoses of those tested.^[88]

The Skipton Fund

3.70 The Skipton Fund^[89] advised that 231 patients with bleeding disorders living in Scotland who had received treatment with blood products and who were alive as at August 2003 had received payments from the Fund.

3.71 HPS and Skipton data might suggest that the actual numbers infected may have been lower than the estimates already mentioned. However, some patients may well have been infected by the virus but cleared it prior to the availability of testing. Others may have died without infection with Hepatitis C being diagnosed. Some may survive without signs or symptoms of infection having been noted or brought to the attention of those keeping the relevant records.

3.72 In the case of the Skipton Fund, it cannot be assumed that every person entitled to make a claim did so, and there is some concern about the reliability of a very small number of the claims admitted.^[90] It would not be appropriate to adjust downwards the numbers of patients infected, as inferred from exposure to risk, on the basis of these recorded data.

3.73 The Skipton data are consistent with the UKHCDO data on living patients: given the limits on accuracy in the exercise as a whole, the difference between 231 and 254 living patients is not significant.

Hepatitis C - secondary transmission

3.74 As discussed elsewhere in this Report, secondary transmission of Hepatitis C occurs relatively rarely from mother to baby and, even less commonly, as a result of sexual transmission.

3.75 As regards possible secondary transmission of HCV from patients with bleeding disorders to their partners, Dr Tait said:

Where appropriate, during counselling, HCV infected haemophilia patients were informed about the suspected low risk of sexual transmission and that partners could be tested. This testing could be undertaken by the partner's GP, by the Haemophilia Centre or by the Infectious Disease Units in Glasgow and Aberdeen. In reality, only a small number of partners attended the Haemophilia Centres for testing (approximately 40 between Edinburgh and Glasgow), and I understand from colleagues that only 1 partner was found to be positive. Since the Haemophilia Centre staff had no direct contact with other partners, nor necessarily knew their identity, we have no information on their HCV status.^[91]

3.76 Beyond noting that a minority of patients who contracted Hepatitis C as a result of infected blood or blood products may in turn have inadvertently infected others, it is not possible for the Inquiry to make an estimate of the number of individuals who may have contracted Hepatitis C as a result of secondary infection. It seems reasonable, however, to conclude that the number will have been very small and given the margins of error inherent in the overall estimation of patients at risk, would not make any material difference to assessing the scale of infection.^[92]

Conclusions on HCV infection among patients with bleeding disorders

3.77 The most reliable estimates of the number of patients with bleeding disorders who were exposed to risk of infection with Hepatitis C as a result of their treatment in Scotland are those provided by the UKHCDO in collaboration with the Scottish Haemophilia Directors. Together, they have the best relevant data. Given their expertise in this area and the detailed investigation they have made of their respective records, their evidence is accepted as sufficiently reliable to support conclusions on numbers overall. Taking on board this ongoing work and using the data from the UKHCDO's ongoing look-back exercise, so far the best estimate appears to be about 478.

Hepatitis C as a cause of death among patients with bleeding disorders

3.78. The Inquiry attempted to establish how many patients with bleeding disorders had died as a result of infection with HCV. The Inquiry was assisted by evidence from the Scottish Haemophilia Directors, the UKHCDO and Professor Goldberg of HPS.

3.79 Dr Campbell Tait, on behalf of the Scottish Haemophilia Directors, stated:

To date we have not been able to establish with certainty the current status (alive or deceased) for many patients on the list. This work is ongoing and may well be updated following the cross-referencing exercise with the HPS database. We also have no readily available information on causes of death, but again efforts will be made to establish this detail. Thus at present we can only determine that of 314 cases known to be HCV positive or likely to have had NANB hepatitis, 88 are known to be no longer alive, and of these 88 liver disease was a major contributor to death in 29 out of the 65 for which we currently have 'cause of death' details.^[93]

3.80 Dr Tait confirmed that for the 65 patients for whom cause of death details were available, the data were provided largely by HPS from the information noted on death certificates, albeit some data may also have been provided by the UKHCDO.^[94] To date, Dr Tait and colleagues have not been able to undertake further work on the causes of death of coagulation disorder patients infected with HCV.^[95]

3.81 Dr Tait discussed a separate exercise undertaken by the Scottish Haemophilia Directors to document key features relating to HCV infection in patients looked after in the Scottish centres. Dr Tait explained: 'The patients included were predominantly infected in Scotland but include small numbers of HCV-infected patients who had moved to Scotland and were being cared for in a Scottish centre.'^[96] The results of that exercise were set out in an Abstract which noted that, overall, 248 of 291 patients infected with Hepatitis C (85%) were still alive as at summer 2007. Death had occurred in 11 out of 33 patients (33%) who were co-infected with HIV, against 32 out of 260 patients (12%) who were infected with Hepatitis C only.^[97]

3.82 The updated statistics report provided by the UKHCDO included data on the cause of death of the 193 patients who had been exposed to Hepatitis C as a result of their treatment with factor concentrate and who, as at 2011, were known to be dead. The report explained:

The number and causes of death data are derived from reports from centres and reports from the Office of National Statistics. We provide them with details of all (25,000+) patients registered with the database and they flag them up when they die and send us death certification data.^[98]

3.83 The UKHCDO advised that of the 193 patients exposed to Hepatitis C as a result of their treatment with factor concentrate who were known to have died as at 2011, 21 patients had a liver-related cause of death.^[99]

3.84 Professor Goldberg gave evidence that, as at December 2009, of the 351 patients with blood disorders recorded by HPS as having contracted Hepatitis C as a result of treatment with a blood factor, 78 patients were known to have died.^[100] Of these 78 individuals, 15 had a primary liver-related cause of death recorded on their death certificate and 15 had a secondary liver-related cause of death recorded. Professor Goldberg added the caveat that 'it is not possible to conclude, from this information alone, if Hepatitis C materially contributed to death in these instances'.^[101] As will be seen in the discussion of deaths among transfusion patients at paragraph 3.260, Professor Goldberg did not regard death certificates as a reliable means of identifying those whose cause of death related to infection with Hepatitis C.

3.85 As can be seen from the discussion above, it has not been possible to arrive at an exact number of deaths resulting from HCV infection among patients with bleeding disorders. The current status of some patients is unknown, and even where a patient is known to have died, determining the role of HCV from the information on the death certificate involves some guesswork. The following figures emerged from the various investigations reported to the Inquiry:

Source	Population in study	Deaths with liver-related cause
Scottish Haemophilia Directors (reported by Dr Tait)	314 cases known to be HCV +ve or likely to have NANBH	29
UKHCDO	193 patients known to be HCV +ve and known to have died at 2011	21
HPS (reported by Professor Goldberg)	351 cases recorded as HCV +ve	30

Patients exposed to risk of transmission of Hepatitis C as a result of blood transfusion

3.86 The second group to be considered is referred to in paragraph 3.2, item ii, and consists of patients who were exposed to infection with Hepatitis C through transfusion with blood or blood components.

3.87 There are several particular issues relating to transmission of HCV by transfusion around:

• the number of patients exposed to risk of infection;
• the number of individuals infected with HCV;
• the number of those infected who have been seriously affected by disease;
• the mortality from liver disease among those infected; and
• the numbers of infected individuals who might still be alive.

3.88 Finding the answers to some of the questions that arise remains problematic. For example, the Inquiry has gathered evidence of the adverse consequence of infection for specific individuals, reflected in the narrative of patients' experiences, which demonstrates that in some cases the progression of disease has been associated with severe morbidity. These patients essentially make up a large but self-selected group of individuals prepared to provide evidence to the Inquiry. It cannot be assumed that they are representative of a wider population of patients, and inferences about total numbers affected to a similar extent cannot be valid. At most a general inference that significant numbers of patients are probably severely affected would be justified.

3.89 It is not known how many patients in Scotland contracted Hepatitis C as a result of blood transfusion, following the increasing use of transfusion between the end of the Second World War and the introduction of the screening of donors for Hepatitis C in September 1991. There are several reasons for that. The size of the total population at risk is not known - that is the numbers of patients receiving transfusions of whole blood, or blood components which might have transmitted the virus, in the course of surgical or medical treatment, over the period when there was in fact a risk of transmission of infection. That period started before the reference period of the Inquiry. Material risk was probably low in the early 1970s. It grew thereafter with the increased prevalence of HCV infection in the general population, and by inference to some extent in the donor population. That increase can be estimated statistically, but is essentially unquantifiable. This major building block in any statistical analysis of the extent of HCV infection in transfusion recipients is at best incomplete.

3.90 Inevitably, much of the evidence received has related to retrospective studies. In general terms, regardless of the infection concerned, the length of time over which the interval of retrospection extends may be a factor affecting the availability and reliability of the data available for study. There may be insufficient biological material (usually stored serum) from possibly infected individuals available to researchers for study. The virulence of the agent affects the reliability of studies. Very serious, often fatal, diseases may have left few survivors by the time the retrospective analysis is attempted. The age and co-morbidity of the infected individuals at the time of transmission also affects the population available for study. The longer the time that has elapsed since the putative transmission event, the more likely it is that the recipient has died. The cause of death may be related to the infection transmitted, or a cause other than the infection itself.^[102] Older individuals affected by other unconnected but serious illnesses, may have died before the retrospective study is carried out, unaffected by the infection transmitted or where the contribution of that infection had not been identified.

3.91 The infection may be distributed within a very large, heterogeneous, population. It may be impractical to test individuals from all cohorts within the population and extrapolation from one particular group to the population as a whole may lead to serious inaccuracies. Similarly, attempts to identify the incidence and prevalence of an infection from symptoms, may be vitiated when only a small proportion of those infected display recognisable symptoms within weeks, months or even years of acquiring the infection. The study of Hepatitis C in NHS patients provides a particular example of these difficulties.

3.92 In contrast to patients with coagulation defects, most of whom benefit from regular monitoring and treatment as part of the management of their primary disease, the follow-up of patients receiving blood transfusion in the course of medical or surgical procedures has been less structured and less comprehensive. In addition, the vast majority of patients who have contracted Hepatitis C through transfusion are likely to have done so at a time before the virus was identified. By the time tests became available to diagnose the disease, many of those patients will have died untested for the virus (mostly as a result of age-related conditions or as a result of the condition which necessitated the transfusion). Of those patients who contracted Hepatitis C through transfusion and who were still alive when diagnostic tests became available, many have not been tested for Hepatitis C. Most of these patients will have been unaware that they may have been infected with the virus at the time of transfusion, since only a small minority of patients experience significant symptoms when first infected. For the majority of patients who are infected, symptoms readily related to the infection are unlikely to appear until decades later.^[103] On the other hand, a proportion of those exposed to the virus either do not become infected or, if infected, spontaneously clear the virus and in either case would test negative for the virus. They would not go on to develop any Hepatitis C disease manifestations.

3.93 Of those patients who develop symptoms and who test positive for Hepatitis C, a transfusion (possibly many years ago) may not be identified by the patient or the patient's clinician as the likely source of infection. In those cases where transfusion is identified as a possible route of transmission, matters may be complicated by the presence of other possible routes of transmission which cannot be excluded, such as intravenous drug use, which is by far the most common route of transmission overall in Scotland and in the rest of the UK.

3.94 Despite these considerable difficulties, the Inquiry has attempted to investigate the number of patients who may have contracted HCV as a result of blood transfusion in Scotland. Several sources of evidence have been drawn to the attention of the Inquiry. There were radical changes in reporting structures in 1995 and 1996 which complicate discussion of the sources and of the data contributed. It is appropriate, however, to note that the problem facing the Inquiry is not new: look-back studies were initiated to find and inform those who had been infected, out of a sense of responsibility for their welfare. The investigation of the numbers affected arose incidentally in that context, and it is to the conduct and outcome of those studies that it is necessary first to turn.

The look-back studies

3.95 Targeted HCV look-back studies, first in the Edinburgh and south east region of the SNBTS and later throughout the UK, are discussed generally in Chapter 35, An Investigation into the Steps Taken to Identify the Individuals who were Infected (Look-back). In 'targeted look-back' the donation histories of individuals found to be infected when screened for a specific pathogen are followed with a view to tracing possible recipients of blood, blood components or products prepared from previous donations by those individuals, in order to ascertain whether any of the recipients may have contracted the virus. For the purposes of this chapter, the outcome of the national HCV look-back discussed in Chapter 35 at paragraphs 35.182 to 35.193 is relevant.

3.96 The national look-back was initiated in 1995 and ran until 1998, using data available following the introduction of anti-HCV screening in September 1991. Prospective donors who tested positive comprised individuals who had not been dissuaded from offering blood by self-exclusion literature, which they may have ignored or not known of, or which they may have thought irrelevant to them. For present purposes, those found to be anti-HCV positive can be assumed to have been apparently healthy individuals. The results of the first few months of testing in England and Scotland, showed anti-HCV prevalence values among the blood donors tested, of 0.066% and 0.088% respectively.

3.97 In the case of return donors found to be HCV positive, prior donations were traced and, where available, stored samples were tested for infection. The destinations of infected donations were then traced, and those used for clinical purposes identified. The hospitals to which blood or blood components - red cells or platelets, for example - were issued were contacted through formal routes, informed of the facts and asked for details of the status of the individual component. They were asked in particular whether a patient had received it and, if so, whether the recipient patient could be identified.^[104]

3.98 On 31 January 2006 the Health Minister of the Scottish Executive provided the results of the look-back exercise to the Health Committee of the Scottish Parliament.^[105] The results, taken from the final report of the look-back exercise, are summarised in Tables 3.8 and 3.9.

Table 3.8: Results of Scottish Hepatitis C look-back

Hepatitis C positive donors who had given blood before 1991	360
Donations by those donors	1658
Components prepared from those donations Of which - Traced Not traceable	2026 1356 670
Number of recipients identified by hospitals Potentially eligible for counselling and testing Of which - Counselled and tested positive Counselled and tested negative Other - Declined; not appropriate for testing; results not reported back to SNBTS	880 266 133 70 63
Deceased	536
Not traceable	78

3.99 The number identified in the Scottish look-back report may have substantially understated the full extent of the problem. Screening and enquiry after the commencement of HCV testing in September 1991 identified infected blood donors who had given donations before 1991. Previous donors who were infected with HCV but did not present again after the introduction of HCV testing in September 1991 could not be identified, nor could their prior infective donations. Among recipients of blood from donors in this category there may have been individuals who developed the clinical manifestations of HCV infection and were treated by the NHS, before or after September 1991. The look-back exercise did not target those individuals.

3.100 As shown in Table 3.8, in total the group identified had given 1658 donations which were available for clinical use. The data recorded showed that an average of 1.2 components were prepared from each donation. The destinations of 1356 of the total of 2026 components were traced. The destinations of the remaining 670 were not traced, and therefore whether they were applied in treating patients, or discarded, could not be determined. Only 880 recipients of the 1356 components traced were identified. Five hundred and thirty-six of these recipients (60%) were already dead by 1995 and could not, therefore, be tested.^[106] Of the remaining recipients, some were not traceable, some were unwilling to be tested, some patients were very elderly, very ill or had a low life expectancy, or were incapable of consenting to testing, and some results were not returned for recording. In the end, in Scotland, 133 individuals were counselled and tested positive, 70 were counselled and tested negative, and the remainder fell out of the scope of the project. Only 203 of the 880 recipients contributed relevant data.^[107]

3.101 Data recovered and held by the HPA on the National Hepatitis C Register in respect of 103 of these patients show known dates of exposure as set out in Table 3.9.^[108]

Table 3.9: Known dates of exposure to HCV

1977	1
1983	3
1984	4
1985	8
1986	11
1987	9
1988	14
1989	19
1990	15
1991	19
Total	103

3.102 The range of values over the period probably reflects the reality that for earlier periods more patients would have died or have disappeared without having been identified.^[109] Furthermore, it is likely that the proportion of donors found to be HCV positive in 1991 who would also have been donors in the early 1980s, is smaller than in the later years of the decade. Forty nine of those 103 patients were known to be alive at January 2011. In Dr John Gillon's view, Hepatitis C caused or materially contributed to the deaths of eight of the 54 patients known to have died. In one case no cause of death is known.^[110] Of the 53 for whom death certificates were available, Hepatitis C was recorded on 14 only, though all 53 were known to have been infected.^[111] Dr Gillon explained:

I should also add that these data are so far not analysed by HPA and haven't been published. So this was kindly made available to us by Dr Helen Harris for the purposes of the Inquiry. The interpretation which I have given is a personal interpretation, and may not reflect what they finally decide when they analyse these things formally through HPA and publish the data ....

I really categorised people into those who had clear evidence that their final demise was fairly directly attributable to hepatic disease. In other words they have liver failure or a complication such as sepsis, or they had hepatocellular carcinoma. Again, there may be a primary cause of death such as bronchial pneumonia, but hepatocellular carcinoma ... that is clearly attributable to the Hepatitis C.

So my reading is that eight of these 53 had a final illness where Hepatitis C was the significant factor.

....

I think that by and large at this sort of length of follow-up, which is 20 years plus now, that's broadly in line with what's in the published literature.^[112]

3.103 In response to a request by the Inquiry, the SNBTS produced a paper on look-back^[113] which sought to explain the relative ineffectiveness of the procedure in relation to HCV. The paper highlighted:

• In contrast to HIV, HCV was present in the population for several decades before a test was implemented;
• HCV typically remained silent following infection in most individuals for many years. HIV may progress rapidly;
• the epidemiology of HCV infection was less well understood than that of HIV, and measures to exclude at-risk donors in the period before development of a test were less effective, though exclusion measures related to HIV infection presumably reduced the numbers of donors with HCV;
• the greater the period elapsed between transfusion and the look-back exercise, the more likely it is that the patient will have died from a cause other than Hepatitis C;
• the increasing use of component therapy put more recipients at risk from a single donation;
• blood bank and hospital records were seldom available from the pre-computer era, and it was almost impossible to trace the fate of donations from before the early 1980s. After the early 1980s many blood banks relied on paper records which could not be searched systematically. It was therefore frequently impossible to establish whether, or to whom, a blood component was transfused; and
• the highly efficient patient tracking systems available today are a relatively recent development.^[114]

3.104 In oral evidence, Dr Gillon expanded on the difficulties with the Hepatitis C look-back exercise.^[115]

• Early donor card records were not searchable.^[116]
• It was not possible for the SNBTS to identify all prior donations by an infected donor.
• It was not until 1983 or 1985 (depending on the region involved) that a computer system was in place by which each donation could be linked to each donor and, importantly, to each recipient.
• Even where previous donations by an infected donor could be identified, there were difficulties in tracing the recipients of each donation.
• Different transfusion centres and hospitals had different record keeping systems.
• There were difficulties in cross-referencing donation numbers with recipients and vice versa.
• Medical records were lost or destroyed.
• Over time, recipients may have married (and changed their surname) or moved address.

While the look-back exercise was undoubtedly worthwhile, in Dr Gillon's view, as a means of trying to identify, counsel, test and treat those patients at risk of having contracted Hepatitis C as a result of blood transfusion, it was not a reliable guide to the number of patients likely to have become infected with Hepatitis C through transfusion.

3.105 Dr Gillon's evidence is accepted. The HCV look-back exercise cannot be relied on as providing an accurate measure of the number of NHS patients exposed to the risk of transfusion-transmitted Hepatitis C virus. Nor does it provide, of itself, a basis for estimating total exposure.

SNBTS donor data: reporting of transfusion-transmitted infection

3.106 During the reference period, clinicians were encouraged to report to the blood transfusion service any adverse reactions arising from the transfusion of blood and blood products, including the transmission of infection.^[117] However, as previously discussed at paragraph 3.11, there was no effective legal obligation on clinicians to report, nor was there any legal obligation on the transfusion service to report to the Secretary of State or any other central authority, any adverse reactions of which they were aware. It is unlikely that surgeons or other clinicians who were not in a professional relationship with the SNBTS would have read the relevant guidance.^[118]

3.107 A Scotland-wide Hepatitis C diagnosis database, now maintained by HPS, was set up in 1996. The SNBTS and HPS now share basic data on HCV infection. The data are held by the National Microbiology Reference Unit for HPS.^[119]

3.108 The UK transfusion services set up a voluntary informal reporting system in 1996, the Serious Hazards of Transmission (SHOT) system, under which a central body collects and analyses anonymised information on adverse events and reactions in blood transfusion, from all healthcare organisations involved in the transfusion of blood and blood components in the UK and makes recommendations to improve patients' safety.^[120] Since then Dr Brian Dow has collated all look-back cases and reports from Scotland and forwarded them to SHOT, providing a database from 1998.^[121]

HPS Hepatitis C Diagnosis Database

3.109 Anonymised data acquired from Hepatitis C testing laboratories in Scotland is recorded by HPS in the Hepatitis C Diagnosis Database.^[122] The primary objectives of the database are (1) to determine the total number of persons diagnosed with HCV infection in Scotland (from whatever source), by reference to year, NHS board, age, gender, risk group and referral source and (2) to monitor trends in diagnosed HCV infection in the Scottish population.^[123]

3.110 Professor Goldberg explained that virology laboratories (in hospitals throughout Scotland) hold information obtained from the Hepatitis C test request form that accompanies a blood sample to the laboratory.^[124] Unlike the test request form for HIV, there is no standardised Hepatitis C test request form for use throughout Scotland.^[125] All forms contain a space for 'additional information' or 'clinical information'. It is by interrogating the test request forms sent to the laboratory that HPS is able to access the information they need for their health surveillance and epidemiological purposes.^[126] In his written statement Professor Goldberg said that: 'Hepatitis C information available to HPS ... is not as comprehensive or accurate' as in the case of HIV, due to the use of forms of varying design.^[127] It is apparent that there are two deficiencies in the existing arrangements: the lack of consistency in form design and lack of consistency in returning relevant information. Professor Goldberg said that 'sometimes' clinicians provide the laboratory with relevant information in the space for 'additional information' or 'clinical information'.

3.111 HPS took over data collection from the SNBTS, and the numbers generated by the database are discussed below at paragraphs 3.131-3.139.

The numerical evidence: known or recorded cases of transfusion-transmitted HCV

SNBTS donor data

3.112 In the period between 1 September 1991 and 31 December 2009, for which detailed data are available, the SNBTS conducted anti-HCV testing on 546,843 donations from new donors and 4,663,441 donations from repeat donors.^[128] It was found that 450 new donors and 417 repeat donors were infected with HCV (that is to say were anti-HCV positive), a total of 867. Though the aggregate numbers were found to be similar, the prevalence rates differed widely. On average over the period the rate of positive donations from repeat donors was 8.94 per 100,000. The rate for new donors was 82.29 per 100,000. The rate overall was 16.64 positive donations per 100,000. The numerical distribution of infected donations over the period is reflected in Figure 3.3. The relative rates per 100,000 donations are shown in Figure 3.4.

3.113 The data for 1991 are not comparable with the data for the rest of the period, since four months only are reflected in the numbers. For the whole of 1991, SNBTS National Statistics records 332,731 donations (approximately three times the total for the part year). The total for 1992 was 332,825. For the full years from 1992 onwards, a downward trend in positive results has been apparent. It is highly likely that a proportion of the return donors found to be anti-HCV positive in 1992 would have tested positive in the first eight months of 1991. Many donate annually. It is also likely that some donors in that eight-month period who did not return in 1992 would have tested positive. While it might be speculated that the result would be to shift the peak of infected donations back one year, it is impossible to say what the peak values would have been. Individuals' donation patterns vary widely, though the return donor population as a whole is reasonably stable. The trend is reliable only from 1992 onwards.^[129]

Figure 3.3: HCV-infected blood donations: Scotland

Figure 3.4: HCV-infected blood donations: Scotland: Rate per 100,000

3.114 As would be expected, in the case of return donors, the numerical trend reflects the relatively high initial effectiveness of a new screening test on a largely stable donor population. Dr Gillon identified two further possible contributory factors to the trend: a change in population prevalence of infection, and better donor selection.^[130] Of the total of 867 HCV-positive blood donors, 59 (6.8%) stated that blood transfusion was the only risk factor for exposure to infection.^[131] That number might include individuals who were transfused outside Scotland. Their cases have not been investigated individually.^[132] Initial enquiries known to Dr Gillon established a date of transfusion in nine of 59 cases.

3.115 The reliability of these reports of an association with transfusion is open to question and it is unclear what inferences can be drawn from the information generated. The following factors demonstrate this lack of clarity:

• The 59 individuals were people who, on receiving counselling, indicated that they thought they had received a blood transfusion in circumstances that suggested that the transfusion might be responsible for their infection. There was no objective verification of their statements.
• The information depended entirely upon reporting by the individual.^[133]
• The reporting may have understated their experience of transfusion, as in the case of patients who had forgotten that it had happened, or who may have been transfused under anaesthetic, or who may not have understood the procedure carried out to be a transfusion.
• A clinician responsible for an injection of platelets might not always have informed the patient that he or she had been transfused. For example, in one study in Italy of patients who had well-documented transfusion-transmitted HCV infection, none of the patients knew they had been transfused.^[134]
• The responsibility of transfusion may have been overstated by patients who concealed, innocently or otherwise, previous conduct exposing them to risk, such as intravenous drug use, body piercing or tattoos.^[135]
• The process will not have identified individuals who were transfused in Scotland but diagnosed as infected in England or Wales or abroad.^[136]

3.116 The differences in prevalence values between return and new donors are significant. The relatively high, and in percentage terms sustained, levels of HCV-positive donations from new donors supports the view that donor demographics and donor selection practices contributed to a relatively low prevalence of HCV-positivity among the established donor population as compared to the general population (represented by the new donors). On the other hand, the impact on total prevalence of HCV-positive donations was low because throughout the period the pattern was dominated by donations from return donors.

Patients reported to SNBTS by clinicians as possible cases of transfusion-transmitted (TT) HCV

3.117 The next group of patients identified by the SNBTS comprised individuals reported to the SNBTS by clinicians, for the most part hepatologists and gastroenterologists, who had treated a patient infected with Hepatitis C where the information available suggested that the likely mode of infection had been transfusion. These cases were investigated individually. It was not always possible to establish a diagnosis of transfusion-transmitted Hepatitis C with certainty. Archives of frozen plasma or serum samples mostly dated back only as far as 1986 and it was often not possible to trace the implicated donors for testing.^[137]

3.118 This exercise was carried out by SNBTS retrospectively to attempt to answer the Inquiry's questions about the incidence of transfusion-transmitted HCV infection, and was dependent on finding and researching individual patients' records in transfusion centres.^[138] After that it became a question of judgement. Dr Gillon said:

[T]he information is open to interpretation. It can be difficult to know whether to say, "Well, we accept this as a case of transfusion transmission. This one probably is but we really don't have enough documentary evidence to say with certainty".

On a number of occasions we can rule it out. We can say the blood that this person received was tested, all the donations were negative, all of the donors have come back and tested negative subsequently. We can be confident that transfusion did not transmit that infection. There are some cases where you can feel that the information that we have is a bit skimpy and therefore you would hesitate to say that this is likely to be a transfusion transmission. But we know there are other ways of picking up transmissible viruses in hospital environments, as we will see in some of the data from the renal units, for instance. Therefore, unless we can identify a donor and establish that link with certainty, there is always a bit of interpretation that's necessary here. I have tried to be inclusive here. In other words, not to wish to minimise the figures in any way, but there is this caveat that, we can never be certain unless we make the link.^[139]

3.119 After investigation of the reports, the numbers accepted as 'known' cases of transfusion-transmitted HCV, as detected by these means, were:

Glasgow and west of Scotland	6
Edinburgh and east of Scotland	15
Dundee (east of Scotland)	4
Aberdeen (north east Scotland)	3
Inverness (north of Scotland)	0

3.120 It is apparent that the numbers for the two major centres are not proportionate to the respective populations of the regions. Dr Gillon thought that the total number of 28 underestimated by a considerable margin the residual number of undetected TT HCV cases.^[140] Historically, the SNBTS tended to see reports from haematologists, for instance, who were dealing regularly with patients requiring multiple transfusions. There was fairly close and constant interaction with them. Similarly, in units which used a lot of blood, like cardiac surgery, renal units and other units with whom the SNBTS had clinical links, clinicians were more alert to the possibility of transfusion-transmission than a surgeon at a district general hospital who was much less likely to think of the possibility.^[141]

3.121 Actual dates of transmission have not been established in these cases. For all but one patient, probable dates lie between 1979 and March 1991. One patient received multiple transfusions between 1970 and 1983 and identification of the infective donation was impossible.^[142]

3.122 As regards this group of known infected patients, one cannot equate the number of cases of transfusion-transmitted Hepatitis C reported by clinicians, with the prevalence of transmission. As already noted, only about 20% of individuals who contract Hepatitis C will experience an initial acute period of illness (eg jaundice) and may be aware that they have contracted a disease. The corollary is that the vast majority of patients who contract Hepatitis C, and their clinicians, will be unaware at the time that infection with the virus has occurred. While a patient may, many years or decades later, develop symptoms and be tested for Hepatitis C, only in some cases will the patient and treating clinician attribute that to a prior transfusion and only in a minority of these cases will this be reported to the SNBTS. Many more cases will remain undetected or unreported for a number of reasons, including the death of many patients before symptoms from hepatitis developed, the patient or clinician not attributing infection to a transfusion occurring many years previously or the clinician not reporting the matter to the SNBTS. While 28 cases in this category were accepted by the SNBTS as being 'known' cases of infection, Dr Gillon stated: 'I think there is no doubt that a total of 28 is way off what is the reservoir of such cases in the population. We don't know by how much of course.'^[143]

Summary of SNBTS known or recorded cases

3.123 On SNBTS evidence, the total number of patients infected by transfusion and identified in these various exercises was therefore 220 (133 patients identified by the look-back exercise plus 59 donors who were thought to have been infected by a prior transfusion, plus 28 patients reported to the SNBTS by clinicians). Dr Gillon thought that the total numbers identified by his studies were probably minimum numbers for infections in Scotland.^[144]

3.124 Dr Gillon acknowledged that there was a theoretical possibility of overlap among the three groups of patients, but it was unlikely to have occurred. He knew of no HCV positive donor in the SNBTS group who was subsequently identified as a patient in the targeted look-back. The documentation and procedures used in the targeted look-back differed from the other groups. Patients from the third group were individually researched. There was a risk that a patient identified within the third group might subsequently have been identified by the targeted look-back without a connection having been made. The anonymising of data recorded by the HPA, and the numbers of cases recorded by the HPA, made it impossible to rule out the possibility, but the risk was unlikely and the numerical impact of such an event would be small.^[145]

3.125 It would not be appropriate to conclude on the basis of the SNBTS donor data that only 59 HCV positive blood donors had been infected by blood transfusion. In particular, donors who stopped donating prior to the introduction of Hepatitis C screening in September 1991 would not be included. Similarly, neither the 133 patients identified by the look-back exercise nor the 28 reported by clinicians provides a reliable guide to the total who may have become infected. Nor is it possible to develop from these numbers an estimate of the total number of patients who may have been infected by transfusion. None of the individual figures provides a basis for inference of that total. Individually, the cohorts are too small to support a more general conclusion. In total they represent an aggregate only of small numbers in specific groups: they are not random samples from which a more general picture can be inferred.

Renal and other units

3.126 Following the development of tests for HCV antibodies, certain patients were screened for HCV on a regular basis by local virology laboratories. The patient groups included people with haemophilia, bone-marrow transplant recipients, and renal dialysis patients. Some positive data were derived from this work. Transfusion transmitted HCV was confirmed from the patients' treatment records.

3.127 Renal dialysis patients considered to be known positives for TT HCV were identified separately. The renal unit in the west of Scotland had monitored the patients because of the well-known risk of hepatitis transmission in renal units.^[146] Initial interest had been in Hepatitis B, which had a high prevalence in patients on chronic dialysis, for reasons that were poorly understood at the time. In addition, patients with chronic renal failure are usually anaemic, and those with severe anaemia as part of their condition require regular transfusions. Therefore, many of them would have had numerous transfusions and most of those would probably have predated the time when the sample archive was started in the west of Scotland in 1986. The patients in this group were reported by the renal units in the west of Scotland as possibly having been infected with transfusion-transmitted Hepatitis C when they started routine testing for HCV antibody.^[147] These patients may have been infected through blood transfusion, but the large number of transfusions administered, and substantial risk of nosocomial infection from other sources, made it impossible to say if and when they had been infected by transfusion-transmission.^[148]

3.128 In addition, a very small number of patients had, prior to the national look-back exercise, been identified by the West of Scotland Blood Transfusion Service as HCV positive. These patients numbered 18 and were primarily people with leukaemia who had received multiple transfusions over many years, leading to bone marrow transplantation. The dates of the transfusions were between 1982 and April 1991; identification of the infective donations was considered impossible because of the number of transfusions involved.

3.129 Dr Gillon said:

In such a case, really the only way you can find if there is a donor who transmitted would be to have archive samples which you can go back and test. We have certainly in Edinburgh done occasional look-backs where we have tested between 100 and 200 samples, which is a very big exercise. Occasionally it does provide results. But some of these patients will have had hundreds of individual units of transfusion and many of them will have had these transfusions before the archive samples started, which would make it impossible.^[149]

Conclusion on the SNBTS data

3.130 On the basis of these SNBTS data, it is possible to conclude only that 238 known patients (220 + 18) were infected by transfusion-transmitted Hepatitis C, but that they are very likely to be parts only of larger cohorts of patients.

Health Protection Scotland data

3.131 As noted above, HPS took over data collection from SNBTS. The changes introduced by the Public Health etc. (Scotland) Act 2008 with effect from 1 January 2010, appear to give regulatory effect to the practice of HPS in obtaining information directly from HCV diagnostic laboratories. Professor Goldberg said, however, that HPS 'got better information by going direct to the hepatitis testing laboratories' on a voluntary basis, and that the 'voluntary approach' enabled HPS 'to get the information that we actually needed'.^[150]

3.132 The SNBTS data were passed to HPS' National Microbiology Reference Unit.^[151] The remit of HPS is primarily related to preventing future infection and disease. As noted above, HPS established its Hepatitis C diagnostic database in 1996. By then blood transfusion and blood products were not serious issues in the context of preventing further transmission of HCV (since by that time blood donations were screened for HCV and blood products were subjected to processes that were effective in inactivating any Hepatitis C virus in the product). Currently 1000 to 1500 people are thought to become infected with Hepatitis C in Scotland every year, and behaviour exposing people to risk is the focus of much of the energies of HPS, but this does not relate to blood transfusion.^[152]

3.133 As already noted at paragraph 3.20, information is acquired from Hepatitis C testing laboratories in Scotland. The various request forms document demographic information and identifying information. In deriving the information they need for surveillance and epidemiological purposes, HPS then includes other pieces of information that are obtained through surveys of various population groups and also clinical information obtained through a national clinical database.^[153]

3.134 The data available to HPS and recorded in the HCV database indicated that 304 individuals may have been infected with HCV by blood transfusion. In these cases, transfusion could only be regarded as a possible and not a definite or confirmed route of acquisition.^[154] In arriving at the figure of 304 cases, 51 individuals were excluded because they recorded intravenous drug use as well as having received blood transfusions. Between 80% and 90% of drug injectors at most centres were infected with Hepatitis C and it is thought that the 51 individuals were much more likely to have been infected by that means than by blood transfusion.^[155] A further 49 had also been previously excluded because the information provided indicated very strongly that they received their infected transfusions abroad.^[156] That did not conclusively prove that the remaining 304 were transfused in Scotland, only that they might have been transfused in this country.^[157]

3.135 The data are acceptable as accurate reflections of the information available, given the criteria applied, but they do not define the scope of the problem of transfusion-transmitted HCV infection. Professor Goldberg said:

I think that, because there is so much uncertainty about these cases ... what we are doing here is just taking some information that has been recorded on a request form. We did not seek additional information. We didn't clarify whether indeed the information provided was accurate. So that was why the word "possible" was provided here.

If I had to put money on it, I would say that less than 50 per cent of the 304 contracted their HCV through blood transfusion.^[158]

3.136 That was his personal assessment, but he was not sure that there were sufficient data to make a judgement, and his ultimate position was that 'don't know' was probably the correct answer.^[159] It is appropriate to accept that view of the position. The identification of 304 individuals as possibly infected by transfusion-transmission already involves a number of assumptions. Breaking the number down further, without a rational basis, seems problematic.

3.137 However, while acknowledging that the lack of confirmation associated with blood transfusion was a weakness in the system, Professor Goldberg stated that he was 'pretty confident' that the information that Scotland had on Hepatitis C was 'as good as, if not better than, anywhere else in the world'. He added:

I think there are weaknesses in our information base with respect to blood transfusion and blood factor, but in general our information about Hepatitis C is pretty good. If you want to compare our diagnostic information, ie numbers of people known to be infected, with the information available in England, then we are in a far superior position in terms of the completeness of our data and indeed its accuracy.^[160]

Conclusion on the HPS data

3.138 Like the SNBTS data, the HPS data relate to patients who had been identified as HCV positive and for whom it was likely that infection had been transfusion transmitted. The data do not reflect directly the possible or probable level of infection from transfusion in the whole population of patients exposed to risk from infected blood or components.

3.139 Even if all 304 individuals on the HPS database acquired Hepatitis C as a result of blood transfusion, that is highly unlikely to reflect accurately the number of transfusion-transmitted infections in Scotland, largely for the reasons discussed above. In particular, patients will not appear on the HPS Hepatitis C database if they had died (for example from age-related conditions or from the underlying medical condition which necessitated transfusion) prior to the database being set up in 1996. Nor will those patients who contracted Hepatitis C but have never, or have not yet, presented to their GP for examination or treatment. Nor will those patients who have presented but for whom, for whatever reason, their Hepatitis C infection has not been diagnosed or, if diagnosed, has not been attributed to a prior transfusion.

The Skipton Fund

3.140 The Skipton Fund was established on 25 March 2004 by the Department of Health, to make ex gratia payments to people who were infected with Hepatitis C through treatment with NHS blood or blood products prior to September 1991.^[161] Payments are made in two stages, the first on proof of infection and the second on proof of progression to serious liver disease. Those living applicants who qualify for a stage 2 payment qualify also for regular payments thereafter. In 2011 the Caxton Foundation was set up to provide charitable benefits to certain individuals who have received blood, blood products or tissue from the NHS and in consequence have been infected with HCV, and to individuals who have had infection transmitted by such a person (the 'primary beneficiaries') and to the partners, carers, children and dependants of primary beneficiaries living or who died before 29 August 2003.

3.141 As at 25 February 2011, there had been 636 Scottish applications to the Skipton Fund for stage 1 payments, of which 134 patients had also applied for and received stage 2 payments, as shown in Table 3.10 below.^[162]

Table 3.10: Scottish applications to the Skipton Fund at 25 February 2011

Patients with:	Stage 1	Stage 2
Haemophilia	210	33
Haemophilia with HIV	21	9
Non-bleeding disorder	405	92
Total	636	134

3.142 At least 24 of the total of 636 had died by 25 February 2011. Fifteen of the stage 1 payments and 15 of the stage 2 payments went to an estate. At that stage there were 405 non-haemophiliac recipients of blood or blood components in Scotland who had been alive or were survived by qualifying beneficiaries as at August 2003 and who received payments from the Skipton Fund. The figure was unlikely to represent the actual number of patients who contracted HCV as a result of treatment with NHS blood or blood products. A minimum of 12 years, in many cases 20 years or more, will have elapsed between the time of infection and the commencement date of the Fund in 2003. Mr Nick Fish, the Scheme Administrator, provided further information on 1 March 2012.^[163] By then Skipton had made 670 stage 1 payments in Scotland (34 more than shown in Table 3.10 above). 162 of the individuals, or their estates, had also received stage 2 payments. It is unlikely that there will have been a material increase in the sizes of the haemophilia groups, given the level of monitoring in their case, and it seems reasonable to proceed on the basis that the additional 34 cases could be added to the post-transfusion group which had increased to 439.

3.143 The accuracy of the Skipton data depends on the accuracy of the medical certificates submitted and the rigour of the assessment procedure carried out.

3.144 Professor Goldberg commented on the Skipton Fund survival data, in particular as a basis for back-calculating plausible estimates of the number of individuals who must have been infected in the first place. He noted:

[I] asked the Inquiry for details confirming that the individuals considered eligible for Skipton funding had, indeed, acquired their infections through blood transfusion in Scotland but such information was unavailable. It is my understanding that, for some individuals, such information exists but for others, perhaps the majority, judgements were based 'on the balance of probability'. Balance of probability is insufficient information for us to make a judgement as to how likely an individual was infected as a consequence of a blood transfusion, without having access to additional information.^[164]

He and his colleagues therefore considered the Skipton data to be insufficiently robust for their purposes.

3.145 As discussed above, It is clear that HPS' own data and analyses lack the level of certainty that would enable one to say that the reference cohort of infected individuals had 'indeed' acquired their infections through blood transfusion. The Inquiry cannot avoid expressing a view on the ground that certain facts can only be established on a less demanding test. The question is whether the Skipton data can be used, along with other sources (including HPS' evidence), in arriving at a conclusion about the likely level of transfusion-transmitted HCV infection.

3.146 The Inquiry has identified one case in which a claim was admitted where there is no evidence that the deceased patient acquired his Hepatitis C infection from blood or blood products. There may be more. One cannot proceed on the view that the Skipton data are unquestionably accurate. The information available indicates that applications are decided by the Fund on a balance of probabilities test and if refused, applicants can appeal to an independent Appeal Panel.^[165] It appears reasonable to accept the figure of 405 (or 439, if the assumption in paragraph 3.142 is valid) as sufficiently accurate to be taken into account in an overall assessment of the level of transfusion-transmitted infection.

3.147 However, as previously discussed at paragraph 3.92, some patients may have cleared the virus, some may have died without infection with Hepatitis C being diagnosed, and some may survive without signs or symptoms of infection having been noted or brought to the attention of those keeping the relevant records. In the case of the Skipton Fund, it cannot be assumed that every person entitled to make a claim did so. In the result, both the number of those who died while infected, and the number of individuals who remain infected, but who have not been registered by the Skipton Fund, are unknown.

3.148 That number may be significant, and it will be appropriate to return to the Skipton data to consider whether estimates can be made on the basis of the recorded data.

Summary of recorded data

3.149 In summary, known or recorded cases of transfusion-transmitted Hepatitis C infection represent a minimum number of patients who were infected with Hepatitis C as a result of transfusion and do not reflect the actual number of patients who are likely to have contracted the disease in that way.

3.150 A wider exercise was required to ascertain whether the available data could, with appropriate analysis and expert advice, provide answers.

Statistical modelling

3.151 As is illustrated by the earlier discussion of data for actual cases of transmission of infection, attempts by epidemiologists retrospectively to form views on the incidence and prevalence of an infection may be hampered by many factors. The general problems discussed in paragraphs 3.89-3.93 arise also in the context of statistical modelling.

3.152 With respect to HCV infection, of the two main groups under consideration in this section, the haemophilia population is much more homogeneous. Patients tend to be much younger at the date of infection than transfusion patients. On the whole, haemophilia patients are much more rigorously followed up. In their case it is far easier to get reasonable, if not perfect, estimates of numbers infected with HCV than in the very much larger and more heterogeneous population of surgical and medical patients who have received transfusions in the course of treatment of their primary condition.

3.153 In addition to these general difficulties, HCV is not a virulent, rapidly progressing disease resulting in severe early clinical manifestations or death. Infection may be diagnosed only after long periods of time since the putative transmission. The retrospective study of transfusion-transmitted HCV from 1991 onwards has been affected by all of these factors. The unreliability of the data on actual cases of infection, and the difficulty in forming general conclusions on the basis of those data, are not unexpected.

3.154 For these reasons, in the absence of reliable recorded data, statistical modelling techniques were adopted by experts for calculating the numbers of transfusion patients possibly exposed to risk, and the numbers of patients infected. Statistical modelling inevitably depends on a number of assumptions in developing general inferences from the data collected. Dr Gillon said that it was difficult to find good data about prevalence in the donor population, to which to relate reported data on patients.^[166] Projections were qualified by the accuracy or inaccuracy of the data available and of the assumptions used to derive them. Transforming the raw data into acceptable numbers proved difficult. All of these difficulties have been readily recognised by the investigators in the following exercises.

Dr Soldan, Public Health Laboratory Service (PHLS)

3.155 In 2002 Dr Kate Soldan, an epidemiologist based at the Communicable Disease Surveillance Centre, PHLS, and a large team of collaborators, estimated the number of individuals who might have been infected with HCV through blood components administered in England.^[167] Dr Soldan later carried out a similar exercise relating to the position in Scotland. The results of her English research were to be published shortly after she carried out her work in relation to Scotland.^[168]

3.156 Dr Soldan constructed a model of the path followed by blood components (red cells, platelets, fresh frozen plasma and cryoprecipitate), which had been prepared from donations collected by eight blood centres in England, and had been traced, from donation to recipient, during the national Hepatitis C look-back programme. The eight centres handled 80% of all blood components entering the programme in England. The number of components, 9222, handled by these centres provided the base data used. Some components were followed through every stage, from delivery to application, and some fell out of the process. The proportion of the components that remained 'observed' at each successive stage in the development of the model was calculated. It was assumed that components that did not complete the observed path would probably have followed the path of those that did, and so they were re-entered into the model at the point at which they fell out of the process. The probability that components with an unidentified fate would follow the observed route was calculated at each stage and used to predict the probable fate of those re-entered components.

3.157 The history of the 9222 components is shown below in Table 3.11.

Table 3.11: Dr Soldan's observed path data

Total Components	9222	100%
of which:
Components not traced	2119	23%
Components traced	7103	77%
Components traced	7103	100%
of which:
Components transfused (4586/7103)	4586	64.6%
Components transfused	4586	100%
of which:
Components with no known recipient (154/4586)	154	3.36%
Components with identified recipients (4432/4586)	4432	96.64%
Number of identified recipients	4424	100%
of which:
Recipients known to have died (2711/4424)	2711	61%
Recipients assumed to be alive (1713/4424)	1713	39%
Recipients assumed to be alive	1713	100%
of which:
Recipients not tested (651/1713)	651	38%
Recipients tested (1062/1713)	1062	62%
Recipients testing positive for HCV (539/1062)	539	50.75%

3.158 In addition to the 1062 tested recipients of components, from the 80% of centres which handled the 9222 components, test results were available (through the look-back exercise) for 271 other individuals. These patients had received components from one or other of the centres in England, providing the remaining 20% of blood. 50.75% of that number (the percentage testing positive in the observed path: 539 out of 1062) provided 138 additional positive tests, giving an overall total of 677 positive tests from 1333 recipients tested.^[169]

3.159 Applying the observed probabilities to components that fell off the process prior to recipient testing and were re-entered, added 3373 HCV infections:

• 946 where the fate of the component was not traced.
• 107 known to have been transfused but with no recipient identified.
• 1870 known to have been transfused to recipients who had died by the end of 1995.
• 450 who declined testing.

Of these, 1870 (55%) were known to be dead and another 645 (19%) were expected to have died by the end of 1995.^[170]

3.160 Dr Soldan next estimated the probable number of infections from donations that did not enter the look-back study. The number of donations collected between 1 January 1980 and 1 September 1991 was 25,864,035. An infectivity rate of 0.066% was assumed on the basis of observed data from the first four months of donor testing after 1 September 1991. At that rate, it was estimated that 17,086 anti-HCV positive donations would have resulted. At 1.6 components per donation (reported to be an observed datum in England), 26,647 components from infected donors resulted arithmetically.^[171] That number necessarily included all or some of the components that entered the look-back programme, and an adjustment was made to deduct those. Of the 26,647 components, it was assumed that 9756 entered look-back in the material period and that 16,890 did not. That would have predicted an extra 10,905 transfused recipients (16,890 x 64.5%), of whom 6034 were infected (based on the observation that 55% of look-back components resulted in infection). However, that was thought unreliable. It was estimated that 73% of the 9756 components entering the look-back programme were anti-HCV positive and accordingly the estimated number of additional anti-HCV positive components not entering look-back was adjusted to 19,525 (26,647 - (9756 x 0.73)). On the consistent assumption that 64.5% of the components traced were transfused, that was calculated to equate to 12,606 recipients of whom 75% (9455) would be infected, on the basis that only HCV RNA-PCR positive donors transmitted the infection.^[172]

3.161 The aggregate, 13,505 (4050 + 9455), was the number of recipients of blood components estimated to have been infected in England with HCV during the decade prior to the start of anti-HCV testing. Over 8300 (61%) of these patients were either known or expected to have died by the end of 1995.^[173] Dr Soldan and colleagues noted that: 'There were, by necessity, many assumptions and extrapolations used, and the results are not therefore expected to be exact.'^[174] They stated:

We may have underestimated or overestimated the infections transmitted from 1 January 1980 to 1 September 1991 by using the prevalence of infection at the start of testing without accounting for selective removal of infected donors during the 1980s, or accumulation of prevalence over time. This uncertainty, and others, prohibited including earlier years. If the prevalence of anti-HCV amongst blood donors during the 1970s was assumed to be the same as at the end of 1991, inclusion of the 1970s data would generate approximately 10,000 extra HCV-infected blood recipients.^[175]

3.162 It has not been possible to verify the assumptions made.

3.163 Professor Goldberg commented:

[W]hat she has done here is she has got some pretty solid data ... and then she has extrapolated the findings to those components for which she doesn't have solid data, and I think that's a reasonable thing to do.^[176]

3.164 Professor Goldberg's evidence is discussed further at paragraph 3.176 below. For the purposes of this Report, it is sufficient to proceed on the basis that the statistical method as applied in England, while open to question in some respects, was not fundamentally undermined by Professor Goldberg, and its results were not seriously disputed on the evidence before the Inquiry.

3.165 As noted below, the results were used by the Department of Health (DoH), apparently without the need to explain adjustments. On the basis of Dr Soldan's modelling, the number of HCV infections as a result of blood transfusion in England between 1970 and 1991 can therefore be assumed to have been about 23,500.

Department of Health

3.166 On 10 January 2011 the DoH published a review of the support available to individuals infected with Hepatitis C and/or HIV by NHS-supplied blood transfusions or blood products, and the dependants of those infected. Annex 2 of the review contained an estimate of the number of Hepatitis C-infected individuals in the UK, who were infected over the period 1970-1991. It was estimated that 28,043 individuals had been infected with Hepatitis C as a result of blood transfusion in the UK during that period.^[177] The source for that figure was stated to be the 2002 paper by Soldan and others on the numbers infected in England, 'corrected to UK'.^[178]

3.167 The population data for the United Kingdom over the period between 1971 and 1991, derived from census returns, are set out in Table 3.12.

Table 3.12: Population of the UK and constituent countries, 1971-91 (thousands)

	1971	1981	1991
England	46,411.7	46,820.8	47,875.0
Wales	2740.3	2813.5	2873.0
Scotland	5235.6	5180.2	5083.3
Northern Ireland	1540.4	1543.0	1607.3
Total	55,928.0	56,357.5	57,438.7

3.168 England had 83% of the total infected population consistently over the period, Wales had 5% and Northern Ireland 2.8%. The Scottish percentage fell from 9.36% to 8.95% and then to 8.85%. For present purposes, it is sufficient to treat Northern Ireland as having 3% and Scotland 9% of the total UK population.

3.169 The period covered by the DoH estimate was approximately twice as long as the primary period discussed by Dr Soldan in detail (1970-91 as against 1980-91). As already noted, in their published paper the Soldan group commented that if the prevalence of anti-HCV donors during the 1970s was assumed to be the same as at the end of 1991, inclusion of the 1970s data would generate approximately 10,000 extra recipients of HCV-infected blood. Dr Soldan's 2002 paper therefore suggested that there may have been 23,505 HCV infections as a result of blood transfusion in England between 1970 and 1991.^[179]

3.170 If the United Kingdom blood donor population were assumed to have had the same rate of HCV infection on average throughout the period, the DoH total would be allocated as follows: England 23,276; Wales 1402; Scotland 2524; and Northern Ireland 841. The figure for England is a close approximation of Dr Soldan's estimate of 23,500.

3.171 Without information about the methodology used by the DoH to adjust for the United Kingdom as a whole, it is not necessarily the case that Dr Soldan's conclusions were accepted nor, if adjusted, to what extent. The most one can establish is that Dr Soldan's conclusions were not challenged by DoH.

3.172 It seems appropriate to assume that other figures would be reasonable approximations of the data available to the DoH, at least as a starting point. However, the mechanisms by which the DoH values were arrived at are not disclosed, and it is not impossible that they were simply the arithmetical result of extrapolating from Soldan on the basis of population data.

Dr Soldan's Scottish estimates

3.173 In 2002 Lord Ross' Expert Group^[180] asked Dr Soldan to provide an estimate of the number of patients in Scotland likely to have contracted Hepatitis C as a result of transfusion.^[181] Dr Soldan's estimate for Scotland followed the same general methodology as she had developed for the English project.^[182] The same statistical model was used and she applied some of the same assumptions, for example in relation to the probability of a blood component being transfused. She also factored in the information available from the Scottish HCV look-back exercise and hence the higher prevalence of HCV among blood donors in Scotland in the first six months of HCV screening (0.088%) when compared with the prevalence among blood donors in England for the first four months' period studied there (0.066%).^[183] The exercise was customised by using the numbers of components entering the look-back programme resulting in an identified recipient. It identified recipients who had died, recipients who declined testing, and those who were tested and found to be HCV positive. The exercise was completed using Scottish data and the parameters of the English exercise. Dr Soldan's brief report does not refer specifically to the assumed yield of components per donation. The Scottish datum would have been information that would have had to be provided by the SNBTS. Dr McClelland's letter commenting on Dr Soldan's estimate at the time referred only to 'data from the Scottish HCV look-back programme and also the higher prevalence of HCV in the Scottish donor population' as information provided to customise the exercise.

3.174 The detailed calculations cannot be replicated in full. However, the results for Scotland, as reflected in Lord Ross' report, were as shown in Table 3.13 below.

Table 3.13: Dr Soldan's Scottish estimates

Number of individuals probably infected with HCV by blood transfusion 1980 to 1991
	Total	Not known to be dead by 1995	Known/ expected to be dead
Identified by HCV look-back programme in Scotland	106	106	0
Received components that entered look-back but did not receive testing in that programme	1243	628	615
Received components issued between 1 January 1980 and 31 August 1991 that did not enter the look-back programme in Scotland	2149	878	1271
Total	3498	1612	1886

3.175 As noted above, the number identified and tested positive through the national look-back became 133 (Table 3.8), and an adjustment would be required for that change if for no other.^[184] Dr Soldan's other data cannot be reconciled with the results of the look-back study. In this exercise the known infections accounted for a mere 3% of the total projected. Dr Soldan commented that many assumptions were used to generate her estimates, some of uncertain validity, and she had serious reservations about the application of her findings to individual cases.

Health Protection Scotland

3.176 To assist the Inquiry, Professor David Goldberg and his colleague at Health Protection Scotland (HPS), Dr Christian Schnier, undertook a separate modelling exercise, in collaboration with the SNBTS, to estimate the number of people infected with Hepatitis C as a consequence of blood transfusion in Scotland during the period 1970-91 and the number who were alive as at June 2011.^[185] Professor Goldberg explained the differences between his model and Dr Soldan's model as follows:

It should be understood that the model used by Schnier and Goldberg in 2011 for the Penrose inquiry was very different to that used by [Dr Soldan] ... in 2002. Further, the [Goldberg/Schnier] model was built following considerable consultation with Dr McClelland and Dr Gillon of SNBTS. The principal differences between the models is that the Soldan one used "lookback data" to inform its estimates and did not factor in any variation in HCV antibody prevalence among blood donors during 1980-31 August 1991. The Schnier and Goldberg model did not use lookback data to inform its estimates but did use estimates of HCV antibody prevalence among blood donors for each year during 1970-1991 ...

Schnier and Goldberg, in consultation with McClelland and Gillon, made the assumption that the size of the HCV infected IDU [injecting drug users] population in Scotland was directly proportional to the HCV infected donor population.

In addition ... Schnier and Goldberg used a factor to account for the introduction of blood donor deferral in 1984 - a factor not used by Soldan et al.

Two other factors used by Soldan were modified for Scottish purposes; instead of the number of units generated from one blood donation being 1.6 (Soldan), the Scottish estimate, based on local data and expert opinion, was deemed to be 1.25; the proportion of units transfused was estimated to be 56% and not 66% (Soldan).^[186]

3.177 Using their statistical model, Professor Goldberg and Dr Schnier provided an initial estimate of the number of individuals in Scotland infected with Hepatitis C as a result of transfusion between 1970 and 1991.^[187] That was followed up by a fuller document in the form of an academic paper.^[188] In their paper they set out the various assumptions upon which their model was based.^[189] The probability of a transfused blood component being infected with HCV was influenced by factors such as the size of Scotland's Hepatitis C infected population in any particular year and the effectiveness of the SNBTS' deferral policy during the period 1984-91. Direct evidence of the size of the HCV-infected population was not available, either in the Scottish population as a whole or in the population of Scottish blood donor.

3.178 The available observed data comprised:

• The numbers of blood donations made during the period 1975-91.
• The prevalence of HCV antibody among individuals donating blood during September 1991 - February 1992.

3.179 Assumptions were required for:

• The number of blood donations made by blood donors during 1970-74.
• The average number of blood components generated by a blood donation.
• The probability of a blood component being transfused.
• The probability of a transfused blood component being infected with HCV.

3.180 The acceptability of the assumptions is fundamental to the exercise. The World Health Organization's Global Burden of Hepatitis C Working Group in 2004 (referred to in the Goldberg/Schnier paper) emphasised that precise estimation of the incidence of HCV infection was not possible, given available data. Professor Goldberg noted that it was possible only to estimate the numbers of individuals who had acquired infection through blood transfusion (or any other route). A combination of observed data and assumptions based on observation and expert opinion was employed to develop estimates.

3.181 As in England, the first firm information on the prevalence of HCV antibody in blood donors in Scotland came with the collection of data on anti-HCV among blood donors during the initial months of HCV screening. Prevalence in Scotland over a period of six months was found to be 0.088%. For present purposes, an important parameter was the proportion of antibody-positive persons who were RNA virus-positive; that is infectious. The first material assumption made by Professor Goldberg and his colleague was that the appropriate proportion was 75% (assumption ii in the report). The other 25% of HCV antibody positive donors would have cleared the infection and would have been HCV RNA-negative, with the result that the prevalence of HCV-positive infectious donors in the donor population in Scotland in 1991 was estimated to be 0.066% (0.088, less 25%).^[190] As with other parameters, uncertainty was reflected in their statistical model, which assumed values ranging from 50% to 85%.

3.182 The basis for Professor Goldberg's assumption was the WHO advice provided by the Global Burden of Hepatitis C Working Group. It had among its key areas of study the natural history of HCV infection, including 'healthy individuals', morbidity and mortality. Other estimates have been made, and a clearance rate of 20% has been referred to.^[191] There cannot be a single 'correct' value. Professor Goldberg's adoption of the WHO rate was justifiable in the absence of hard data suggesting a need for variation to account for local factors. It was within, though at the lower end of, the most up-to-date estimates reported by Dr Hay in the UKHCDO updated report of April 2012.^[192]

3.183 From the prevalence of 0.066% estimated for 1991, the next necessary step was to estimate the prevalence of infection in the donor population over the study period, beginning in 1970. There were no relevant observed data for that population. Prior to anti-HCV testing there were no available means of measurement. In developing their estimate of the number of people who acquired HCV infection as a consequence of blood transfusion in Scotland during the period 1970 to 1991, Professor Goldberg and Dr Schnier adopted the results of a study by Hutchinson, Bird and Goldberg, carried out by statistical modelling, which estimated HCV prevalence in injecting drug users (IDUs) between 1970 and 1991, and applied the trend brought out to work back from 0.066% to values for each year (assumption iv).^[193]

3.184 The Hutchinson study was based on local data from Glasgow for the prevalent number of IDUs per calendar year from 1960 to 2000. The data were adjusted for assumptions about age-related use, mortality from non-HCV causes and other factors, to bring out a pattern of the prevalence of HCV infection among current IDUs. The model was then adapted to the whole of Scotland based on available epidemiological data by adjusting for a number of key parameters. The paper was aimed at providing quantitative estimates of the current and future burden of HCV disease in the IDU population in planning a public health response to treatment needs and preventive measures. The characteristics of IDUs affecting HCV disease progression were discussed, and are reflected in the paper.

3.185 Goldberg and Schnier assumed that the change in HCV prevalence among the blood donor population was proportional to the change in the estimated number of HCV-infected IDUs.^[194] This was an important assumption. Injecting drug use continues to be both directly and indirectly the principal driver of Scotland's HCV epidemic. There is a direct impact from a donation by an infected injecting, or former injecting, drug user. An indirect impact resulted from a donor who had never used intravenous drugs but acquired HCV infection from an injecting drug user, sexually or through contact with needles or instruments in a healthcare/tattoo/barbershop setting. The Goldberg/Schnier assumption of proportionality reflected the view that approximately 90% of all individuals diagnosed with HCV in Scotland, for whom risk factor information was available, had injected drugs, and that 'an appreciable proportion' of the remainder acquired infection indirectly, by contact with an infected IDU.^[195]

3.186 In addition, the authors noted that in 1984 the SNBTS introduced a deferral policy aimed at reducing the number of donors who were at higher risk of transmitting blood-borne virus infection (ie the leaflets and questionnaires introduced from 1984 to try to exclude donors at a higher risk of transmitting HIV), and they assumed that the deferral policy reduced HCV prevalence in the blood donor population by a constant factor of 66% from 1984 onwards.^[196]

3.187 The progressive development from the 1970s and early 1980s of methods of identifying and deterring high risk donors from giving blood are discussed elsewhere in this Report at Chapter 26, Donor Selection - Higher Risk Donors, and Chapter 28, Donor Selection - AIDS. At this stage it is appropriate to note only that these steps were taken, and are likely to have reduced the incidence of infection in donations. There is a question related to the quantification of the immediate impact of the policy, and its sustained effect over the period to 1991, to which it will be necessary to return. However, some estimate of the impact of the deferral policy was necessary.

3.188 Professor Goldberg and his colleagues set out in table 1 of their report^[197] an estimate of the prevalence of HCV-positive infectious donors in Scotland as a whole before and after the introduction of donor deferral: that is from 1970 to 1983 and from 1984 to 1991 respectively. The table reflected in the first place the Hutchinson data and the assumption that the change in prevalence among donors was proportional to the estimated change in prevalence among IDUs (paragraph 3.183) and in the second place the assumption that the deferral policy reduced prevalence in the donor population by 66% from 1984. The prevalence in the donor population for each year between 1970 and 1991 which these assumptions generate has been plotted in Figure 3.5.

Figure 3.5: Effect of Goldberg/Schnier assumptions relating to HCV Prevalence in blood donor population in Scotland

3.189 The rapid build-up of prevalence to 0.067% in 1983, reflects directly the assumed correlation of the growth in the number of IDUs and in HCV prevalence in the donor population. From 1984 to 1991, the assumption was that the donor deferral policy consistently reduced the number of infected donations. Professor Goldberg was asked whether a degree of gradation should have been assumed, to allow time for the policy to take effect. He explained that alternatives had been tested, and that it had been concluded that, although the rate selected was intuitive, it was thought optimal.

3.190 Professor Goldberg's first and seventh assumptions (as listed in the report) were the same as Dr Soldan's: that every recipient of a contaminated unit subsequently developed HCV infection, subject to 'discount' for those who cleared the virus or were never to progress to chronic liver disease, and that the risk of an individual receiving two or more infected units was negligible. In the stochastic model some of the assumptions were modified to express uncertainty.

3.191 In their stochastic exercise, Professor Goldberg and Dr Schnier estimated that between 1970 and August 1991, a median number of 1533 individuals were infected with Hepatitis C as a result of blood transfusion (the median estimate of 1533 was derived from a lower estimate of 1198 and an upper estimate of 1963).^[198] An estimated breakdown per year is provided in table 3 of the paper, which indicates that an estimated median number of 232 individuals were infected with Hepatitis C between 1970 and 1979 and 1026 individuals were infected between 1980 and 1989.^[199] In the year prior to HCV screening of donors being introduced (ie 1990), it was estimated that 155 individuals were infected with HCV as a result of transfusion.^[200]

3.192 Based on the assumptions set out in their paper, the authors also estimated that a median number of 296 individuals who were infected with Hepatitis C as a result of transfusion were alive in 2011, of whom 222 were RNA positive (on the assumption that 25% of the 296 cleared the virus and were no longer RNA positive).^[201]

3.193 At least some of the assumptions made by Professor Goldberg and Dr Schnier cannot be verified, and the validity of the figures produced by their statistical model, like the figures produced by other statistical models, is necessarily qualified. If different assumptions and input data are used, different figures are generated.

Evidence relating to the statistical models

3.194 In discussing Dr Soldan's exercise, Professor Goldberg said that there was a potential bias in the exercise, in particular in applying the same factors to deceased recipients and living recipients who had not been tested, as applied to those who followed the observed path to the end.^[202] The difficulty is clearer in the case of the deceased recipients. They were not available for testing. The assumption that the proportion of positive tests derived from testing the living could be applied directly to those who had died is obviously not valid, especially where, as here, the proportion of individuals known or assumed to have died is so high. As Professor Goldberg observed, there must have been reasons why 38% of those assumed to have been alive were not tested.^[203]

3.195 Professor Goldberg said:

But the thing is that the extrapolations appear to be based on the [observed path] and the question is: can you extrapolate? Because there may well be biases in the system which mean that the numbers infected, or the expected numbers infected, may be an underestimate or an overestimate.^[204]

3.196 Commenting principally on Dr Soldan's material, Professor Goldberg said:

I'm not convinced this is the only way to estimate the size of the infected population. I think there are other ways of doing it ... if you use a combination of approaches, then you do reduce uncertainty. But that all takes time and much, of course, is dependent on the information that's available to you. So for Scotland we have information generated through the look-back, but we also have other information about the size of the infected population, ie Scottish population, during the 1980s but also during the 1970s as well. So I would expect to use these data.^[205]

And:

[I]f you just use one method, you are opening yourself up a little. I mean, you know, this is actually a very good piece of work undertaken by Kate, who probably knows more about this field than anybody else in the UK. But it does have its limitations and I think we just have to acknowledge these limitations ....

By and large, when you are doing this sort of work, if you use maybe two or three methods - but much depends of course on the information you have available to you - I think your confidence in your final outcome is very much greater because if you have considerable differences in your results, you can get an average or you can take what's regarded as the best or whatever.^[206]

3.197 These observations appear to be as relevant to the Goldberg/Schnier model as to Dr Soldan's exercise. Professor Goldberg and Dr Schnier were asked by the Inquiry to re-run their model, omitting the assumption that the donor deferral policy introduced in 1984 reduced the HCV prevalence in the blood donor population constantly by 66%. In response to that request Professor Goldberg advised:

We repeated this modelling exercise to ascertain the impact of (i) a deferral policy with no effect and (ii) a deferral policy, the effect of which increased incrementally between 1983 and 1991. For (i) the number infected until 1991 was estimated to be 1110 (90% credibility interval: 876 to 1413), the number alive as at 2011 was 230 (178 to 294) and the number alive as at 2011 and chronically infected was 173 (128 to 225). For (ii) the estimated number of those infected was 2212 (1657 to 2853), the number alive as at 2011 was 438 (325 to 566) and the number alive and chronically infected as at 2011 was 326 (241 to 435).^[207]

3.198 Professor Goldberg and Dr Schnier were also asked to re-run their model, this time omitting the assumption that the HCV prevalence in the donor population between 1970 and 1990 was directly proportional to the estimated prevalence of HCV in the injecting drug use population. They did so by adopting an approach similar to that of Dr Soldan, and assumed a flat rate based on 1991 data. Professor Goldberg replied:

We modified the model, as requested, using a constant HCV prevalence (0.09% antibody positivity); this generated an estimated infected number of 6784 (5027 to 8776), with the number having survived until 2011 being 1050 (789 to 1364) and the number having survived as at 2011 and being chronically infected, 788 (569 to 1044).^[208]

At this stage it is sufficient to note that the changes in assumptions reflected in this paragraph and in paragraph 3.205 below generate very large variations in outcome.

HCV infections as a result of blood transfusion between 1970 and 1991: discussion and conclusions on statistical models

3.199 As Professor Goldberg accepted, neither of the statistical analyses carried out provided a single, wholly acceptable source of evidence. A sensitivity analysis was not carried out to test the assumptions made in varying conditions.^[209] It will be necessary to look at the evidence in the round to determine whether one can make an estimate of the incidence of infection in NHS patients using recorded data and acceptable assumptions.

3.200 There is one point common to the exercises: the use of relatively low levels of known and objective data of HCV transmission. In Dr Soldan's English paper the known data (derived from the HCV look-back programme in England) accounted for 5% only (677/13,505) of the total infections estimated to have been transmitted by transfusion between 1 January 1980 and 1 September 1991. That is a low proportion of known instances of transmission relative to the estimated total. The percentage was higher in the case of patients who survived to 1995, at 13%. The median age of these patients in 1995 was 55 years, significantly lower than the median age of the identified recipients in the group of 3373 added by extrapolation.^[210] On any approach, the projections were very heavily dependent on the validity and accuracy of the assumptions made, and that affects the level of confidence in the result. Dr Soldan's paper recognised that.^[211] In her Scottish exercise, the known data accounted for 3% of the total infections estimated, which was an even lower proportion.

3.201 Dr Soldan's application of a constant infectivity rate, derived from the first four months after testing began on 1 September 1991 in England and the initial six months of testing in Scotland, to all donations for the period 1 January 1980 to 1 September 1991, appears to raise a problematic issue.^[212] It was one of the features considered by Professor Goldberg to be significant in comparison with his own approach. It assumed that this component of the risk of transmission of infection remained the same throughout the 12 years of the period of study. That is not consistent with a picture in which HCV infection rates were increasing as a result of the practices of intravenous drug users, for example, or, more generally, with the impression that HCV infection in the general population appears to have grown significantly over the 1970s and 1980s.

3.202 The increase in HCV prevalence in the general population was one of two relevant countervailing trends over that period. The other was the increasing proportion of potential new and returning blood donors, who were dissuaded from donating blood from the commencement of self-deferral policies onwards, by increasing efforts to discourage 'high risk' individuals from presenting at donor sessions to give blood. The deferral of potential blood donors who were at higher risk of blood-borne virus infection, implemented generally in 1984 in the AIDS period, is likely to have had some incidental impact on the risk of transmission of HCV. For that reason alone, the HCV infectivity rate prior to 1984 might have been higher than in the period from 1984 onwards.

3.203 Another element in the exercise that may have been important was the yield of components per donation assumed in the calculation. The yield in England, based on measurement there, was 1.6 components per donation. In the later Goldberg/Schnier exercise the Scottish yield applied was 1.25 components per donation. There were two possible reasons that may have contributed to the difference. Blood collection in Scotland was plasma-driven in the 1980s and an excess of red cells was inevitable. Dr Gillon speculated that there would have been a higher rate of discard in Scotland, reducing the number of components transfused.^[213] That is a reasonable speculation. In addition, excess Scottish red cell production was 'exported' to England, reducing the proportion of components produced that were actually used in Scotland.

3.204 Dr Soldan did not deal with these issues. It cannot be over-emphasised that it is accepted that Dr Soldan's work was a ground-breaking attempt to provide an estimate of the extent of the problem of post-transfusion HCV infection. The issues that may arise from her original work do not have to be resolved so far as they relate solely to England. That is not within the scope of the terms of reference of the Inquiry. Some, however, have an indirect bearing on the figures brought out for Scotland.

3.205 Statistical modelling is a complex exercise and may be influenced by the interaction of other values. It has to be emphasised that while it is relatively easy to draw negative conclusions on the assumptions made, that does not assist in providing a reliable estimate. Simple arithmetical adjustment for a single altered factor in a statistical analysis would be unsatisfactory. One cannot substitute one value for any given factor for another and express a different conclusion arithmetically. For example, if Dr Soldan did use the component yield observed in England (1.6 units per donation) in her Scottish exercise, that would have been one-third higher than the Goldberg/Schnier figure. The effect is likely to have been large. Without knowing the sensitivity of the model to a variation of that kind, the consequence in terms of the final number could not be estimated broadly.

3.206 Professor Goldberg considered that the extrapolation of the data from the observed path to the other components was a reasonable approach, though not the only approach. As already commented (paragraph 3.164), Dr Soldan's approach was not seriously disputed in the evidence before the Inquiry, and it is appropriate to take the results of her calculations into account.

The Department of Health estimate and Dr Soldan's estimate for Scotland

3.207 While the methodology for the DoH estimate of the total number of TT HCV infections in the United Kingdom between 1970 and 1991 is not clearly expressed, it seems likely that it is an extrapolation of Dr Soldan's figure for England, to take account of the whole UK population. On that basis, as noted above at paragraph 3.170, it is possible to back-calculate the DoH estimate and to suggest that the estimate for Scotland would be about 9% of 28,043, ie 2524.

3.208 This figure for Scotland is significantly different from the figure brought out in Dr Soldan's original Scottish exercise. Her value of 3498 related to components issued from 1 January 1980 to 31 August 1991; it was already higher than the number deduced from the UK values for a much longer period. In relation to England, the assumption of 10,000 additional infections for the previous decade amounted to 74% of the estimate for the later period. Since that involved application of the same general assumptions for both periods, one might expect an equivalent approach to estimating figures for Scotland to have produced an additional number of about 2600, making the Scottish number about 6100 for the whole period 1970-1991. Adopting a flat HCV donor prevalence rate of 0.09%, and applying his model, Professor Goldberg later calculated a median value of 6784 for this period. The difference between 6100 and 6784 is not material given the margins of error implicit in the exercise. On either approach, the total for England and Scotland would already exceed the DoH total for the United Kingdom, and leave no room for numbers of infections in Wales and Northern Ireland.

3.209 It is not possible in the circumstances to accept Dr Soldan's estimate for Scotland as a sufficient and acceptable basis for discussing the numbers put at risk in Scotland. Dr Soldan was not available to assist the Inquiry by reassessing the position.^[214] Her estimate remains one result of statistical modelling. Without detailed examination, it cannot be dismissed. On the other hand, the DoH figure of about 2500 for Scotland, while not supported by detailed description of methodology, has the benefit of proportionality, and has at least superficial support from that fact.

The Goldberg/Schnier model

3.210 As already indicated, there were two critical assumptions in the Goldberg/Schnier model that required particular examination: the adoption of the rate of growth of HCV infection in injecting drug users as the rate of change of HCV prevalence both in the population as a whole and in the donor population, and the impact of the SNBTS blood donor exclusion policy in and after 1984.

3.211 The sensitivity of the Goldberg/Schnier model to variations in the prevalence of HCV infection in the donor population was demonstrated by the re-run of the model excluding the assumption that HCV prevalence in the donor population was directly proportional to the estimated prevalence in the IDU population. Irrespective of the numbers brought out, the assumption was significant.

3.212 It appears to the Inquiry that proportionality implied both that:

• There was a direct relationship between the pattern of HCV prevalence in the general population of Scotland and the growth in the numbers of injecting drug users.
• There was a direct relationship between the pattern of HCV prevalence in the general population and in the blood donor population over the relevant period.

3.213 Overall, the evidence before the Inquiry showed that the growth in numbers of HCV-positive IDUs came to be the major factor to influence the prevalence of HCV infection in the Scottish population as a whole. The issue for the Inquiry was whether the data supported the assumption of direct proportionality of relationship between the IDU population and the general population of Scotland between 1970 and 1990 (assumption iv). That was explained:

The rationale of assuming proportionality is that it is estimated that 90% of HCV infected individuals in Scotland acquired Infection directly through injecting drug use and that an appreciable proportion of the remainder will have acquired infection indirectly as a consequence of injecting drug use (eg being born to an infected IDU or having unprotected sex with an infected IDU) (Hutchinson et al., 2006)^[215]

3.214 The paper by Hutchinson and others published in 2006^[216] analysed information on diagnosed HCV infection at December 2004, (18,571 individuals less 11% who had died, 16,500 net), and estimated the likely pattern of infection in 33,500 antibody positive individuals who had not been diagnosed. Allowance was made for 25% natural clearance. In each case 88% of chronic HCV-infected people were estimated to be IDUs and 12% were non-IDU's; 26.6% of whom were blood/blood factor recipients.

3.215 The 2006 paper carried forward by a year a 2005 study by Hutchinson, Roy and others^[217] of information to December 2003 which had, in turn, drawn on material from a 2004 paper by Professor Goldberg's group at the Scottish Centre for Infection and Environmental Health (SCIEH).^[218] The 2004 paper noted that at 31 December 2003 a total of 18,109 individuals were known to be HCV antibody positive. 61% of those were known to have injected drugs at some time. Those individuals represented 90.5% of the total for whom risk information was available (67% of the total of 18,109).

3.216 The SCIEH paper contained a detailed analysis of data for each NHS Board, and for Scotland as a whole, aggregating data to 1995 and thereafter presenting data annually. For the period to 1995 'blood factor'^[219] accounted for 269 reported cases of HCV antibody-positive individuals, and IDUs accounted for 1029, a ratio of just under 1:4. For 1995 the ratio was 29:648: 1:22. Over the remainder of the period reported the ratio of 'blood factor' antibody positive recipients to IDUs continued to fall.

3.217 It would have been inappropriate to subject these papers to further, more detailed critical analysis. It did not appear that the additional time and expense involved in instructing further expert opinion was justified, since it was clear that a further or different set of assumptions would be unlikely to result in a definitive conclusion. It is not inappropriate to note that the data underpinning the opinions expressed have at no time been comprehensive. Perhaps more significantly in terms of long term trends, inferences have been drawn from reference periods including the highest growth in HCV positive IDUs when, by sheer force of numbers, they were clearly the dominant factor affecting HCV prevalence in the general population. As the SCIEH report notes, for blood factor concentrate recipients there were no new infections after the introduction of effective heat treatment of factor concentrates in the mid-1980s.^[220] From 1991 screening for antibodies to HCV prevented new infections from transfusion generally. Mortality among transfusion recipients was at all times skewed by the significant proportion of NHS patients who died within a short time of the procedure that had required transfusion.

3.218 The Inquiry cannot exclude the direct linear numerical relationship between the increase in HCV among IDUs and the increase in the donor population over the period 1970 to 1991. It seemed appropriate, however, to seek further explanation of its basis. In relation to the principal issue raised with him, Professor Goldberg explained:

You contest that the change in HCV prevalence among blood donors over the two decades in question would not have mirrored the change in the estimated number of prevalent infections among injecting drug users over this period. While we cannot prove this was indeed the case, our view is that injecting drug use has been the principal driver of HCV infection in Scotland, even going back to the 1960s. As we pointed out, it was not just the direct effect but also the indirect one. What we mean by this is that people who did not inject drugs but acquired their infection through, for example, tattooing, are likely to have become infected with a virus which was originally circulating among individuals who injected drugs ....^[221]

It seems appropriate to be cautious about accepting total equivalence, as is implicit in the exercise.

3.219 The next step seems superficially difficult, involving the assimilation of the prevalence of infection in the general population and in the blood donor population. It appears generally to have been accepted that there were always significant differences between the prevalence of HCV infection in the blood donor and general populations, and that these had become more pronounced since the early 1980s. The donor population has always been selected by age and general health factors and, at least in parts of Scotland, by questioning about injecting drug use and observation of signs of injecting. In Professor Goldberg's view, the prevalence of HCV in the general population may have influenced the prevalence in the donor population to a greater extent in the past than in more recent times.^[222] In general, however, a policy of deferral on the basis of intravenous drug use, despite the limitations inherent in the practices of the SNBTS donor teams, must have had some impact on the risk of IDUs both being accepted as donors, and giving blood that entered the pool of blood available for clinical use.

3.220 There were, and are, demographic distinctions between the general population and the blood donor population. It is not at all clear that the cohort of HCV antibody-positive individuals coming to the notice of doctors as being IDUs was ever a major component of the blood donor population (and it is to be borne in mind that between 80 and 90% of that population are return donors). Some return donors at 1991 will have begun donating decades earlier, in some cases before injecting became a major issue.

3.221 Assumption (iii) in the Goldberg/Schnier model was:

Deferral policy, introduced by SNBTS in 1984, was assumed to have reduced the HCV prevalence in the donor population constantly by 66%. This assumption was based on limited local data and expert opinion.^[223]

3.222 The 'local data and expert opinion' referred to information and advice provided by Dr Gillon and Dr McClelland. The paper explained:

In 1984, SNBTS introduced a deferral policy to reduce the number of donors with a higher risk of having blood born (sic) virus infections; therefore the prevalence of HCV-positive donors in the donor population during 1984 to 1991 was lower than it would have been, if the deferral policy had not been in operation. It was assumed that the deferral policy reduced HCV-prevalence in the donor population by 66%.^[224]

3.223 The need for an adjustment to reflect the general point made is clear and is accepted: the deferral policy must have had some impact, whenever it was implemented. The year 1984 was perhaps too early in the evolution of SNBTS management systems to find that there was a 'policy' introduced by SNBTS: the era of regional autonomy had not yet passed, and there were variations in the dates when practice changed across Scotland. Further, as already noted, the policy of Edinburgh and south east Scotland, which became the basis of the general model, still reflected an erroneous view of the duration of risk presented by a carrier of HCV RNA. The question was whether 66% was an acceptable estimate for the impact of the policy change from 1984 to 1991.

3.224 The Goldberg/Schnier number of 6784 recipients of HCV-contaminated units, was based on a flat rate prevalence of 0.09%. Even at that level, the contaminated units remain a tiny percentage of the total number of units transfused, at 0.16%. The Goldberg/Schnier data showed that 4,344,795 units were transfused over the period 1970 to 1991. In terms of the numbers of donors presenting, the percentage is even lower. Over the beginning of the period when voluntary deferral was in practice, SNBTS data on donors and deferrals indicate the difficulty in drawing inferences. The pattern was as shown in Table 3.14.^[225]

Table 3.14: Donors and donors deferred 1981-82 to 1987-88

SNBTS

	Total attendances		New donor attendances
	Attended	% deferred	Attended	% deferred
1981-82	322,304	9.30	51,729	16.5
1982-83	328,086	9.10	50,241	14.5
1983-84	336,802	8.10	54,046	12.5
1984-85	338,278	9.00	50,004	9.5
1985-86	341,307	8.59	48,858	15.8
1986-87	340,890	8.78	48,350	13.5
1987-88	323,837	9.10	39,580	13.6

3.225 In relation to total attendances, the fall in the number of donors influenced by a change in deferral policy is problematic. These were individuals whose blood had been accepted, perhaps often, in the past or who had been deferred temporarily and had returned. In 1984-85 approximately 30,500 return donors were deferred. These were, by definition, individuals who had not been deterred by the policy from attending to offer blood. None who attended and were deferred on grounds of evidence of intravenous drug use could be included with those persuaded by the leaflets and other material to self-defer. The number deterred by the policy and not attending at all must be speculative. There was a significant fall in new donors presenting at centres in 1985-86, and there was a relatively high incidence of deferrals in that cohort in that year alone compared to subsequent years. If it were possible to draw any inference from these data, it would be that the reducing numbers of new donors might have reflected the effectiveness of the policy more than would the numbers of return donors. The year on year cumulative reduction in new donors attending between 1983-84 and 1987-88 was 27%. The increase in deferrals in the new donor group after 1984-85 might suggest that other factors were affecting the situation.

3.226 There is a danger in using the figures set out in Table 3.14 directly. The numbers of infected donors, returning and new, would be small relative to the totals in each group, and the relative percentage movements might have been affected by quite small changes in numbers. For the period September to December 1991, there were 33 infected donations from new donors. The number for 1992 was 55. They were, of necessity, included in the numbers who were not dissuaded from giving blood. One cannot tell what number would have attended to give blood but for the impact of the policy change, although it is not obvious that it would be large.

3.227 Dr McClelland and Dr Gillon were asked to reconsider this area. From their responses, it appears that, with very limited relevant factual data, they attempted to make a conservative estimate.^[226] Dr Gillon's explanation tends to underline the lack of numerical underpinning for the estimate:

We knew that the donor selection procedures introduced by November 1984, including as they did a signed declaration by each donor that he or she was not in the defined risk categories, led to the exclusion of a steady number of such donors throughout the period leading up to HCV screening in 1991. I received a written confidential report on every such donor, and therefore had first hand, if somewhat impressionistic knowledge of the apparent effectiveness of the procedures. In order to try to give this a numerical basis, we tried to extrapolate backwards as best we could from the reports on population prevalences published in the early 1990s and subsequently, in order to estimate what proportion of potentially infected donors we were managing to exclude ....

After commenting on HIV among intravenous drug users in Scotland, the discussion continued:

[W]e knew that the initial period of screening for HCV in 1991/2 produced a donor prevalence of 0.09%. Balogun et al (2002) estimated that the population prevalence in England and Wales peaked in 1986, at just over 1%. By 2005 the HPA (Hepatitis C in the UK. 2011 Report) estimated a prevalence in adults of 0.67%.

We interpreted these data as evidence that the entirety of donor selection policies and procedures, including publicity and donor education, reduced the risk of an infectious unit entering the blood supply by at least an order of magnitude ... We agreed in discussion that it would be appropriate to be very conservative in the final assumption, but we acknowledge that there are residual reservations about the assumptions on population prevalences. It was therefore correct to state that Assumption 3 derived in part from our expert professional opinions.

3.228 The Goldberg/Schnier report noted that 149 of 180,000 donors tested in the study period from September 1991 to February 1992 tested HCV antibody-positive. The adjusted value of 0.066% was the prevalence among those who presented at donor sessions, a cohort within the general population of potential donors but not co-extensive with that population. It was a difficult exercise to retrospectively apply that prevalence on the basis of limited data on the prevalence overall, whether in the IDU population or in the general population.

3.229 Neither Dr Gillon nor Dr McClelland offered any computational support for the percentage applied in the Goldberg/Schnier exercise, nor any source data that were relevant to the estimate. It remains heavily dependent on limited data, their local knowledge and their professional judgement. Their local knowledge and their professional judgement are undoubtedly wide-ranging, and generally reliable, but cannot be assumed to be infallible.

3.230 So far, this assumption has been discussed in terms of Scotland as a whole, as it was presented, and by reference to the time-frame selected. There are grounds for concern that detailed aspects of the assumption may not stand examination. The introduction of a deferral policy has to be set in context of other steps taken to reduce the risk of transmission of hepatitis, and the differing chronologies that emerged across Scotland. As discussed in Chapter 26, Donor Selection - Higher Risk Donors, collection of blood in penal institutions had been phased out by 1984. Edinburgh and South East Scotland BTS last collected blood from prisons in 1981; with the exception of Glasgow and south west Scotland, all other regions last collected in prisons in 1983; Glasgow and south west Scotland made their final (and much reduced) collection in 1984.^[227] Though a small percentage of total collections while the practice of prison sessions continued, they carried a disproportionately high level of risk of transmission and, if the policy was valid, its implementation must have affected the prevalence of HCV infection in the donor population before the revised deferral policy came into effect.

3.231 There were policy differences underlying the differing dates of phasing out of prison collections. There were similar policy differences relating to minimising the risk of transmission by donor selection and self-deferral. In Edinburgh and south east Scotland, Dr McClelland drafted a leaflet in 1983 promoting self-deferral.^[228] It was tabled at a meeting of the SNBTS Coordinating Group on 24 May 1983.^[229] At that meeting Dr Mitchell commented that he had introduced to the donor health questionnaire a question inviting those who were worried about AIDS to consult the doctor at the session. The Glasgow leaflet did not include a question on whether the donor had ever injected drugs, nor did a donor leaflet seemingly in use in England and Wales in 1983.^[230] Dr Urbaniak, Director in the Aberdeen Centre, had decided after consideration not to do anything locally.^[231] Dr McClelland's leaflet attracted adverse criticism from the Scottish Homosexual Rights Group.^[232] Following constructive dialogue, a revised form of the leaflet was prepared. The revised form was distributed on 15 June 1983.^[233] The position in other regions remained as before. There was now a second significant, though relatively short-lived, difference in policy between Edinburgh and south east Scotland and the rest of the country that had the potential to affect the prevalence of infection in the donor population.

3.232 A leaflet for UK-wide distribution was ready for use by 1 September 1983.^[234] It seems likely that it was available in Glasgow and the west of Scotland thereafter.^[235] However, the means of distribution and the terms of the leaflet continued to attract debate. The tortuous process of revision of the text, and of the steps to be taken for distribution, saw major developments in February, June, August and November of 1984.^[236]

3.233 On 9 February 1984 at a meeting on the infectious hazards of blood products at the National Institute for Biological Standards and Control (NIBSC), attended by Drs Cash and McClelland, it was reported that:

The policies adopted in Scotland to minimise the risk of transmission of infection were explained. The three main strategies were 1) avoidance of high risk communities (such as prisons, known homosexual areas, etc); 2) detection of clinical abnormalities by examination and careful questioning; 3) exclusion of the high risk donor, or his blood ....^[237]

3.234 Against this background the selection of a date, or period, when deferral policy had a significant impact on the risk of transmission of NANB hepatitis in Scotland is problematic, as is the rigour with which deferral may initially have been carried out. There was no uniformity of policy or practice between the two major transfusion regions. The cessation of prison donation sessions had already altered the distribution of risk by excluding one potentially major contributor to the total pattern of risk, but so recently that its impact on total risk cannot have been known.

3.235 Dr Hay emphasised that it was difficult to know the extent to which donor self-exclusion reduced the number of donors presenting a risk of transmission of infection, because HIV testing began shortly after the self-exclusion programme started. He proceeded on the basis that HIV testing and donor self-exclusion taken together reduced the number of high-risk donors giving blood.^[238] He reported:

It is difficult to quantify the effect of donor self-exclusion and HIV testing on the risk of non-A, non-B hepatitis because there were no reliable diagnostic criteria for non-A, non-B hepatitis in the early 1980s. The prevalence of chronic non-A, non-B hepatitis also varied geographically. The condition appears to have been commoner in the USA than in Northern Europe. Contemporary studies suggest that the prevalence of non-A, non-B hepatitis in Northern European blood donors was approximately 0.4-1.0% in the early 1980s. In contrast, Contreras reported a much lower rate of infectivity of 0.085% per donor unit amongst 387 UK patients transfused an average of 3 units of blood each in 1987 and tested using an hepatitis C antibody ELISA. This suggests an approximately tenfold reduction in the risk of post-transfusion hepatitis C, in the UK during the course of the 1980s, following the introduction of donor self-exclusion and HIV testing ....^[239]

3.236 He explained further in oral testimony. The risk groups for Hepatitis C and for HIV were similar and steps to exclude one virus would have an effect on attempts to exclude the other. While this would not make much difference to the risk from the concentrate derived from plasma in a large donor pool, the risk of transmission from single donor units such as cryoprecipitate or red cells was considerably reduced.^[240]

3.237 Other assumptions in the Goldberg/Schnier model included the yield of components per donation (1.25) which was a value provided by Dr Gillon and agreed by Dr McClelland.^[241] The value was derived from data generated by the SNBTS look-back. SNBTS national statistics for the period 1981-82 to 1993-94 suggest that the yield of 1.25 was reasonably accurate. As calculated by the Inquiry, units placed at issue per donor,^[242] including total cryoprecipitate, varied over the period but ranged between 1.18 and 1.31. The yield adopted was therefore an appropriate value for the purpose of estimating the number of recipients put at risk.

3.238 Similarly, the assumption (number 6 in the Goldberg/Schnier list) that 56% of the donated blood was assumed to have been transfused,^[243] with all units having the same probability of being transfused, appears to have been conservative when compared with SNBTS national statistics for the period 1981-82 to 1993-94. Dr Gillon re-examined this assumption at the request of the Inquiry and concluded that the Scottish percentage should be 66%.^[244] That is still less than indicated by available SNBTS national statistics, but Dr Gillon's conclusion is accepted that the variation he found was not statistically significant, and it appears that little turns on this factor.

3.239 Assumption 8 (a flat profile of the age distribution of recipients of blood transfusion justifying the application of up-to-date data over the whole period),^[245] seems consistent with the assumption that the donor population, including IDUs, was homogeneous, but like that general assumption is questionable on the ground of lack of evidence. Return donors seem likely, as a matter of general impression, to have spanned a wider age range than new donors, including injecting drug users. The donor's age at first recruitment does not change, but increases so long as they returns to give blood. However, there are not enough hard data to enable comment on the extent of any differences.

3.240 It is necessary to be cautious with even these assumptions. Dr Gillon observed that:

[T]he inherent unreliability of this set of assumptions, encapsulating as they do significant numbers of components which could not be traced, would make statistical calculations meaningless.^[246]

3.241 It would be inappropriate to reject statistical exercises generally on that basis. On any view they make a valuable contribution to the overall picture and assist in arriving at an opinion. This observation underlines the need for caution. Having regard to the difficulties with the assumptions in all of the models discussed above, it would be inappropriate to accept the estimates made as conclusive of the extent of transmission of HCV by transfusion.

General considerations

3.242 On the evidence available, both expert approaches appear to have followed valid principles of statistical modelling, but each depended on the application of unverifiable and possibly erroneous assumptions, as they freely acknowledged. It is necessary to treat any estimates of prevalence with considerable caution. While it is inevitable that opinions reflecting calculation have to be expressed numerically, it is particularly necessary to avoid being misled by spurious mathematical precision.

3.243 At the end of the day, all that can be done is to set out a range of estimates which vary depending on the different assumptions made. Professor Goldberg's calculation using a flat rate projection of 0.09% prevalence was worthwhile as indicating a possible maximum number of 6784. It demonstrated the response of his model to the revised assumption, but it was not an assumption for which he offered support. On that basis, the estimate produced by Professor Goldberg's statistical model is that a median number of 1533 patients in Scotland may have been infected with Hepatitis C as a result of blood transfusion between 1970 and the introduction of donor screening in September 1991. Dr Soldan's figure remains 3498 for the period 1980-91.

3.244 Of the two approaches to the problem in Scotland, Professor Goldberg's approach has the advantage of using more relevant hard data, but remains heavily dependent on assumptions which cannot be objectively verified. Dr Soldan uses less local data, and again her assumptions cannot be verified objectively. Since Dr Soldan was not available as a witness, it was not possible to explore the qualifications expressed in her report.^[247] The estimate of 2524 derived from the DoH report falls between the two. As already noted, the methodology adopted in arriving at the figures in the report is not disclosed.

3.245 So far as Term of Reference 4 is concerned, it is clear (a) that recorded data cannot provide an accurate measurement of the extent of HCV infection in Scotland over the reference period, or of the risk of post-transfusion transmission; and (b) that such statistical models as have been developed to date remain, to a greater or lesser extent, of questionable validity as indicators of total exposure to risk and of transmission. The fundamental problem is that hard data are lacking and assumptions have to substitute for measurement over too wide a range of factors.

Skipton and other data

3.246 Professor Goldberg thought the Skipton data to be 'insufficiently robust'^[248] and the Inquiry has identified one example in which it is unreliable. However, in the context of so much general uncertainty about other approaches, it has to be considered whether it is for present purposes of assistance in developing an estimate of numbers.

3.247 Of the total number of stage 1 payments made by the fund (670) 10 were paid to the estates of people who had died before 2003. Six stage 2 payments were made to a further six estates.^[249] Since the 670 included haemophilia patients co-infected with HIV, who had a relatively high mortality, it can reasonably be inferred that there were very few if any payments made to non-bleeding disorder beneficiaries who had died before 2003, and that the number is immaterial for estimating purposes. In any event, on this approach, to assume that all of the additional individuals are transfusion-associated cases is the more conservative approach. Since 2003, there have been deaths among those receiving stage 2 payments: of the 162 stage 2 payments, 37 were known to have died and there were up to 54 possible deaths among the total recipients, with 71 continuing to receive regular payments. The fate of the 54 is unknown, but Mr Fish (the administrator of the Skipton Fund) assumes that many will have died, on the basis that they have ceased to come forward to claim their regular payments. There are no survivorship data for those receiving stage 1 payments only.

3.248 The number of post-transfusion HCV-infected individuals, estimated on the basis of Skipton data to 1 March 2012 at 439 (paragraph 3.142 above), must have included representatives of individuals who had died. Of the total qualifying recipients (670) at least 24 infected individuals had died. If all of those were assumed to be post-transfusion recipients (the most conservative assumption) there would have been 415 individuals in that group alive in 2012. This number can be contrasted with the finding in the 2011 Goldberg/Schnier model of an estimated median 222 HCV RNA-positive post-transfusion recipients (paragraph 3.192).

3.249 The Scottish look-back found that 536 (60%) of the 880 recipients of infected components had died by 1995. Since a balance of recipients had not been identified, and Skipton's reference date of 2003 was later, by which time more patients would have died, it can conservatively be estimated from the Skipton non-haemophiliac recipient numbers and using similar mortality data, that a minimum of 1100 patients would have been infected. That is still considerably lower than the DoH figure and the statistically derived figures.

3.250 The look-back exercise, as devised and implemented, was itself incapable of tracing more recipients than could be associated with the infected donations received after the introduction of screening. A more widely based approach is required to ascertain whether there was support for an estimate based on the DoH exercise.

Look-back study in Denmark

3.251 A Danish post-transfusion look-back exercise in 2009 by Just and others, traced the outcome for 960 patients (1018 less 58 untraced after 1996) who prior to 1991 had been exposed to donations from 150 HCV-positive donors.^[250] By 1996, 730 had died. Of the 288 alive in 1996, 58 had to be excluded from the study because no personal identification had been supplied and they were untraced. Of the remaining 230, it was found that the pattern of infection was as set out in Table 3.15.

Table 3.15: Danish look-back data

HCV exposure in 230 Danish patients
Groups
Infected individuals:	124
of whom:
Patients with cirrhosis		23
Patients who had died			51
Uninfected individuals:	43
of whom:
Patients with cirrhosis		3
Patients who had died			10
Unknown infection status:	63
of whom:
Patients with cirrhosis		3
Patients who had died			46
Total	230	29	107

3.252 The number of infected individuals alive in 2009 was 121 (two who left the country during follow-up were excluded) which represented 12.6% of the original cohort of 960 known to have been exposed to blood from HCV-positive donors prior to 1991.

3.253 If that percentage were applied to the Skipton number of 439, it would indicate a starting cohort of 3484. However, the reference periods are not consistent. Skipton counted patients identified after 2003. Between 2003 and 2009 some of the Danish cohort would have died. There was an observed mortality rate of 4.91%. The deceased members of the cohort over the interval have to be added back to obtain a comparable number at the end of 2003. Assuming five years decrement at 4.91%, the comparable figure for Scotland is 2709, based on the 2003 Skipton numbers and the extrapolated Danish survival data.

3.254 There are, as always, issues over the reliability of the approach. In relation to the haemophilia population, Danish experience of HTLV-III exposure was compared with experience in Glasgow in a study by Melbye, Froebel and others that is discussed in Chapter 10, Knowledge of the Geographical Spread and Prevalence of HIV/AIDS 2.^[251] In that context, there was a clear distinction in therapy regimes that distinguished the two populations of Denmark and Glasgow. But no other distinctions were noted. Denmark had a well-developed health service and the Danish look-back, which took place at a time after the introduction of screening was very similar to the Scottish exercise. The most significant difference was that the Danish Health Information systems were apparently better organised than those in the UK. Comparison with Denmark provides a further figure, based on a comparable population, that may be used as a control: 2700 would be the number of persons assumed to have been infected in Scotland on this basis, and that is not dissimilar to the results of the DoH exercise (2524 individuals).

Conclusions on transfusion-transmitted HCV infection

3.255 None of the exercises described is capable of producing a firm and reliable estimate of the likely number of NHS patients, transfused in Scotland, who became chronically infected with HCV as a result of receiving infected blood or blood components.

3.256 Excluding the extremes, a wide range of values remain as indications of the possible incidence of infection. Only a rough and speculative estimate is possible, and the balance suggests a number of around 2500. From Skipton Fund data, extrapolated to 2012, perhaps about 400 of these individuals remained alive in 2012, the last year for which an estimate can be made using available data. A further unknown number of still 'silently' infected individuals probably exists, 23 years after screening began. This number cannot be ascertained reliably. Any attempt to estimate it would be affected by the fundamental lack of hard data reflected in this chapter as a whole. It might be speculated that the number would not be large. But that would involve a dangerous step into unknown territory.

3.257 Given that throughout most of the earlier years of the reference period it would not have been possible to have measured directly those contracting Hepatitis C from blood transfusion, it would never have been possible through contemporaneous records to arrive at a precise and final figure for those infected, and in particular for those infected and still alive. Estimates supported by statistical analysis would inevitably have been required. Now, even with the best support from expert epidemiologists, it is impossible to use the data available to provide a very satisfactory and reliable number.

3.258 The work carried out by and for the Inquiry suggests that further epidemiological investigation would not produce a more reliable estimate. Nothing can be done now to improve the contemporaneous records, or to provide hard data indicating objectively the scale of the problem. The reporting system, such as it was, was never designed and operated so as to be likely to be effective, even if it was enforced when means of identifying infected individuals became available. If the Scottish Government is persuaded that, for health policy and strategy, or budgeting or other reasons, it is necessary to develop a more accurate figure, it may be that further research and further expert opinion might eventually converge. That cannot, however, be recommended by this Inquiry given the extensive investigations already carried out.

Hepatitis C as a cause of death among transfused patients

3.259 Dr Gillon gave evidence that of the 133 patients identified by the HCV look-back exercise who contracted Hepatitis C as a result of transfusion, 49 patients were known to be alive as at January 2011.^[252] Information on the recorded causes of death was available for 53 out of the 54 patients known to have died.^[253] Dr Gillon's interpretation of data was that Hepatitis C was the cause of death or had made a material contribution to the death of eight of the 53 patients for whom causes of death are known.^[254]

3.260 Professor Goldberg provided evidence that of the 304 individuals recorded by HPS as having contracted Hepatitis C as a result of blood transfusion, 85 individuals were known to be dead as at December 2009 and 219 individuals were not known to be dead as at that date.^[255] Of the 85 individuals known to have died, 18 had a primary liver-related cause of death recorded on their death certificate, 13 had a secondary liver-related cause of death recorded (including viral hepatitis, liver cancer, alcoholic and non-alcoholic liver disease) and 54 had other causes of death recorded.^[256] Professor Goldberg was asked whether a death certificate was a reliable guide to whether an individual had died from Hepatitis C; he replied:

It's not a reliable guide and that's one of the reasons why we use the Hepatitis C diagnosis database in association with the death register, to identify individuals with Hepatitis C who have died, rather than going straight to the death certification register and just relying on the source of that information. It just is completely unreliable in that respect.^[257]

3.261 While the above figures are of some assistance, they require to be treated with caution given that they are based largely on the information contained in each deceased's death certificate which, in turn, depends on the thoroughness and accuracy of the investigation, recording and reporting of each death.

3.262 The mortality figures spoken to in evidence, both for patients with post-transfusion Hepatitis C and for haemophiliac patients, correlate reasonably well with well-modelled studies from outwith the United Kingdom.^[258]

Patients with bleeding disorders infected with HIV as a result of treatment with blood products

3.263. The third group of patients to be considered is referred to in paragraph 3.2, item iii, and consists of patients with bleeding disorders, primarily haemophilia, who were exposed to infection with HIV through their treatment with blood products.

3.264 The Preliminary Report set out the information available to the Inquiry prior to the oral hearings. As with other data, the information has been significantly corrected and updated at and after those hearings.

Evidence of Dr Hay

3.265 When Dr Hay first gave oral evidence to the Inquiry on 18 March 2011, it was already understood that the data provided earlier were inaccurate. Dr Hay noted that in retrospect there were not many people thought to have been infected with HIV in 1980: the majority were infected from 1981 through to 1983. The details in the Preliminary Report referred to the date of reporting when a sample had tested positive, not to the date of infection.^[259] As a result, the tabulated data were skewed to later dates than the dates of actual transmission, and gave an inaccurate impression of emerging risk.

3.266 Further, at that stage Dr Hay's data did not necessarily reflect the numbers of patients who were infected with HIV in Scotland: they showed at best how many patients managed in Scotland were infected with the virus. In a number of cases so managed, HIV infection was first reported by a centre outside Scotland.^[260] There were 1382 registered UK patients who were known at May 2010 to have been infected with HIV. Of those, 72 were registered at Scottish centres. Some of the patients registered at English centres had received treatment in Scotland and might have been infected there.^[261] Where a patient was managed at more than one centre, identification of the centre at which he was infected was problematic and could involve duplication.^[262] There was epidemiological evidence that heavy users of Factor VIII were more likely to contract HIV infection than those using smaller amounts, but that did not assist in determining total numbers likely to have been infected.^[263]

3.267 Before the oral hearings, UKHCDO data had been examined by the organisation and by the Scottish Directors in an attempt to determine how many of the HIV-infected patients registered as managed in Scotland may have been infected at Scottish centres. When Dr Hay first gave evidence, the attempt to reconcile the data from all sources had not been concluded. He thought that without going through each individual's records, it might be impossible to determine exactly where they were when they contracted HIV, and even if that were done, if there were no archived samples it might never be known for certain.^[264] At this stage (in 2010) the Scottish Haemophilia Directors were working on UKHCDO data that had yet to be corrected. There was no satisfactory reconciliation.

3.268 In an effort to resolve the uncertainties and discrepancies in these data, the UKHCDO, led by Dr Hay, and the Scottish Haemophilia Directors, notably Dr Tait (Glasgow) and Professor Ludlam (Edinburgh), conducted a series of data checking and data reconciliation exercises for HIV in parallel with those already described for HCV.

Reconciliation of UKHCDO and Scottish data

3.269 It was necessary for there to be a review of all the data available to try to establish reliable numbers. The Directors of the Scottish Haemophilia centres and the UKHCDO examined the data recorded for each patient on the UKHCDO list, to try to determine whether the patient was likely to have become infected with HIV as a result of treatment in Scotland and, if so, at which Scottish centre.

3.270 A number of issues arose in the comprehensive review that followed of all data available to Scottish Haemophilia centres. There remained the basic problem of double counting of patients. Each centre, operating in isolation, had returned data to the UKHCDO database at Oxford (the National Haemophilia Database) for each patient treated, whether or not the patient was otherwise registered at the centre. For example, a patient registered in Edinburgh who visited Inverness and there received necessary treatment, would be recorded as a separate patient receiving treatment in each centre.^[265] For earlier periods, data were returned manually and later entered into the national computer system. It appeared that there had been some transcription errors and other misunderstandings. The Scottish Haemophilia Directors each prepared a methodology to attempt resolution of the problems with the data.^[266]

3.271 The starting point was the tabulated patient-specific data provided by the UKHCDO relating to patients treated in Scotland who were known to have been HIV positive. The data included the dates of the last negative HIV and first positive HIV tests, and the products with which the patient had been treated year by year, including treatment outwith Scotland. In some cases a cause of death was recorded.

3.272 The UKHCDO data were reviewed at each centre, and some patients were removed from the lists on the basis that they had been HIV positive when they first arrived at a Scottish centre.^[267] Professor Ludlam described in some detail the procedure in Edinburgh and south east Scotland. In the Edinburgh Centre alone, identification of patients known to have been positive on first attendance at the centre, served to remove six patients from the original 29 reported by the UKHCDO. Discussions then took place to exclude duplication as between Scottish centres, and in the case of patients treated at more than one centre, to agree at which centre it was most likely that the patient contracted HIV. In cases where there was little or no information available as to the likely date of infection, a judgement required to be made as to the centre at which the patient was likely to have been infected. On review of local records it was found that UKHCDO data included all patients known to the individual centres. Local data were added where available.

3.273 Professor Ludlam thought it unlikely that there were haemophilia patients in treatment with concentrates who were not registered at a centre. Patients who received treatment outwith a haemophilia centre for a bleeding problem were quickly referred to such a centre, following local laboratory tests.^[268] There was a possibility that patients listed by the UKHCDO as English or Welsh patients might have been infected in Scotland if they had been treated in Scotland before moving to England or Wales in the early 1980s. But no such patient was known, and differences in treatment regimes would in any event make it difficult to form a view.^[269]

3.274 After undertaking that exercise the Scottish Haemophilia Directors gave the following evidence as to the number of haemophilia patients infected at their respective centres.

Royal Infirmary of Edinburgh

3.275 Professor Ludlam stated that 23 haemophilia patients were infected with HIV as a result of treatment at the Edinburgh centre.^[270] All 23 patients had severe Haemophilia A.^[271] Eighteen of the 23 patients had received treatment with material from a single batch of SNBTS product (PFC Factor VIII) during the relevant period, and comprised the 'Edinburgh Cohort'.^[272] Five patients had received treatment with other SNBTS and commercial products.^[273]

3.276 The adjusted list of patients thought to have been infected in Edinburgh was submitted to HPS for any additional information from their register of HIV-infected individuals, including information from death certificates.

3.277 Routine blood samples from the 1970s were regularly stored by Edinburgh virologists as part of a study of Hepatitis B infection and its transmission in haemophilia.^[274] It was therefore possible in many cases to decide retrospectively when the patient had seroconverted to HIV. It was this information that identified as new patients those who were already seropositive when they had arrived in Scotland. This procedure was adopted at all Scottish centres.

3.278 It was originally thought that all members of the Edinburgh Cohort were infected between March and May 1984 after exposure to a single common batch of Scottish Factor VIII concentrate. Following phylogenetic analysis, there are now thought to have been at least two batches of contaminated Factor VIII responsible for infections in the group. It appears that two or three HIV-infected donors, who were not intravenous drug users or heterosexual males, contributed to the plasma pools.^[275]

3.279 One individual (E22 in the table below) is known to have been infected by Armour Factorate between 16 March and 1 December 1981 (the dates of the last sample testing negative and the first testing positive). One (E16) had samples that tested positive in 1983. One (E19) tested positive in November 1986, having tested negative in January 1985. Two (E17 and E20) were under 16 at the time of their first positive sample.^[276] The overall picture is set out in Table 3.16.

Table 3.16: Royal Infirmary of Edinburgh HIV infections

	Last negative	First positive	Treatment	Year assigned	Source
E1	16.08.1984	19.11.1984	PFC	1984	PFC
E2	05.04.1984	06.10.1984	PFC	1984	PFC
E3	13.09.1983	13.06.1984	PFC	1984	PFC
E4	13.09.1984	30.10.1984	PFC	1984	PFC
E5	21.06.1982	18.10.1984	PFC	1983-84	PFC
E6	26.10.1983	08.05.1984	PFC	1984	PFC
E7	18.01.1984	26.06.1984	PFC	1984	PFC
E8	29.05.1984	22.11.1984	PFC	1984	PFC
E9	16.04.1984	20.07.1984	PFC	1984	PFC
E10	27.03.1984	29.05.1984	PFC	1984	PFC
E11	02.03.1983	22.08.1984	PFC	1984	PFC
E12	01.02.1984	29.05.1984	PFC	1984	PFC
E13	09.04.1984	29.05.1984	PFC	1984	PFC
E14	26.06.1984	10.08.1984	PFC	1984	PFC
E15	14.12.1983	17.04.1984	PFC	1984	PFC
E16	05.08.1982	15.09.1983	PFC	1983	PFC
E17	29.03.1984	24.05.1984	PFC	1984	PFC
E18	28.11.1983	20.06.1984	PFC	1984	PFC
E19	01.01.1985	17.11. 1986	PFC	1985	PFC
E20	17.04.1984	23.04.1984	Mixed	1984	PFC
E21	15.05.1984	11.10.1984	PFC	1984	PFC
E22	16.03.1981	01.12.1981	Mixed	1981	Armour
E23	06.08.1984	2.10.1984	PFC	1984	PFC

3.280 Four of the 23 patients infected with HIV at the Edinburgh centre were still alive at the time of the oral hearings in 2011 and 19 had died. Of the 19 patients who had died, 14 had either died of HIV/AIDS or HIV/AIDS was a factor contributing to their death.^[277] While that information was compiled by Professor Ludlam from information provided by HPS (based on the patient's death certificate) and the UKHCDO database, the latter contained 'very little information' as to the cause of death and more assistance was derived from the information in the death certificates.^[278] Professor Ludlam explained that of the five patients whose deaths did not appear to have been caused or aggravated by HIV/AIDS, three deaths were due to a major catastrophic haemorrhage and that one patient clearly had a condition which was not related to HIV/AIDS at all.^[279] Professor Ludlam had the benefit of having been the patients' treating clinician and was able to make an informed judgement on each patient's likely cause of death, rather than having to rely solely on the information recorded in each death certificate.^[280]

Glasgow Royal Infirmary

3.281 Dr Campbell Tait, Director of the Haemophilia Centre at Glasgow Royal Infirmary (GRI), gave evidence about the results of the exercise in Glasgow.^[281] In the final count, 12 patients contracted HIV infection while attending the GRI Haemophilia Centre.^[282] Reconciled data are set out in Table 3.17.

Table 3.17: Glasgow Royal Infirmary HIV infections^[283]

	Last negative	First positive	Treatment	Year assigned	Source
G1	01.01.1982	15.09.1986	Mixed	1982-1986	Imported/PFC
G2	Unknown	15.05.1984	Mixed	?-1984	PFC
G3	15.12.1984	15.11.1985	Mixed	1985	Imported
G4	15.05.1982	15.01.1984	Mixed	1982-1983	Imported/PFC
G5	15.07.1982	15.12.1983	Mixed	1982-1983	PFC
G6	15.05.1981	15.09.1982	Mixed	1981-1982	Imported/PFC
G7	Unknown	15.11.1982	PFC	?-1982	PFC
G8	01.01.1982	15.02.1984	Mixed	1982-1984	PFC
G9	15.10.1984	15.10.1985	Mixed	1984-1985	PFC
G10	Unknown	15.11.1985	DEFIX	?-1985	PFC
G11	15.10.1985	15.07.1986	DEFIX	1985-1986	PFC
G12	Unknown	15.04.1981	Mixed	?-1981	Imported/PFC

3.282 Ten of the 12 patients had Haemophilia A (eight severe and two moderate), and two had Haemophilia B (patients G10 and G11). Most of the patients received both commercial products and SNBTS products and it is not always possible to be confident of a robust allocation of imported or PFC products as being the cause of HIV infection. The cause of infection in three cases (G3, G6 and G12) appears clearly to have been the use of imported Factor VIII. Three patients received only PFC Factor VIII between their last negative and first positive HIV test and it appears that these three patients were very probably infected by an SNBTS product.^[284] Retrospective testing of stored data, where available, showed that of the 12 patients infected with HIV, one patient seroconverted in 1981-82, two in 1982-83, one between 1982 and 1984, one between 1982-1986, one in 1984-85, one in 1985-86, one during 1985 and, for four patients whose date of last negative HIV test is not known, the first HIV positive tests were, respectively, April 1981, November 1982, May 1984 and November 1985.^[285] Ten of the 12 patients infected with HIV were dead by 2011, five from an HIV/AIDS-related cause, one possibly from an HIV/AIDS-related cause, two patients' cause of death appears to have been liver cirrhosis and one patient's death was probably the result of a haemorrhage.^[286] Dr Tait's view on the cause of death was based solely on the death certificate codes provided by HPS for each patient.^[287] Dr Tait started at the GRI in 1999 and, unlike Professor Ludlam in relation to the Edinburgh patients, had probably not met any of the patients who had died.

3.283 It is highly likely that the trend of infection from the use of PFC Factor VIII in the GRI was similar to the trend in Edinburgh, namely that these infections were later in time than infections from commercial product. On the evidence available to the Inquiry, heat treatment of Factor VIII from December 1984 eliminated transmission of HIV by SNBTS factor concentrates processed after that date. The distribution of the infections over the period cannot be more precisely estimated.

Royal Hospital for Sick Children, Yorkhill

3.284 Dr Elizabeth Chalmers, Director of the Haemophilia Centre at the Royal Hospital for Sick Children (RHSC), Yorkhill, Glasgow, provided evidence that 21 children were infected with HIV as a result of their treatment at RHSC.^[288] All 21 children had Haemophilia A (19 had severe haemophilia and two had moderate haemophilia). Imported Factor VIII concentrates were used extensively at the RHSC until 1982,^[289] and infections have been identified at earlier periods than elsewhere.

3.285 All 21 children received both SNBTS product (PFC Factor VIII and cryoprecipitate) and commercial product, in particular 'Factorate' produced by Armour. For 12 of the 21 children, the dates of the last negative and first positive HIV tests are known. Two of the 12 children seroconverted between January 1980 and January 1981, one child seroconverted in 1981, three children seroconverted in 1981-82, four children seroconverted in 1982-83, one child seroconverted between 1981-83 and one child seroconverted between 1982-84.^[290] For nine of the 21 children, the date of the last negative test for HIV is not known. The date of the first sample from these nine children to test positive for HIV ranged from September 1982 to May 1985.

3.286 It appears likely that about five of these 21 children were infected before the first cases of AIDS in haemophilia patients were described in the USA, and that up to 16 were infected before the first case of AIDS in a haemophilia patient was described in the UK.

3.287 The 21 children were infected with HIV relatively early in the HIV crisis, probably as a result of the widespread use by the RHSC of the commercial product 'Factorate'. Eight of the 21 children infected with HIV have died. Of those eight, five deaths were due to HIV/AIDS, two deaths do not appear to have been due to HIV/AIDS and the cause of one death is unknown.

3.288 The full data available for Yorkhill is as set out in Table 3.18.

Table 3.18: RHSC, Yorkhill HIV infections

	Last negative sample	First positive	Treatment	Year assigned	Source
Y1	Unknown	05.04.1985	Mixed	?-1985	Imported/PFC
Y2	01.01.1980	01.01.1981	Mixed	1980	Imported
Y3	15.04.1981	15.05.1982	Armour	1981-1982	Imported
Y4	Unknown	15.02.1983	Mixed	1982-1983	Imported
Y5	15.06.1981	15.10.1981	Armour	1981	Imported
Y6	Unknown	15.12.1982	Armour	1981-1982	Imported
Y7	Unknown	15.05.1985	Mixed	?	Imported/PFC
Y8	Unknown	15.02.1983	Armour	?-1983	Imported
Y9	29.01.1982	10.03.1983	Mixed	1982-1983	Imported
Y10	Unknown	15.01.1985	Mixed	?-1984	Imported/PFC
Y11	01.01.1981	19.03.1982	Armour	1981-1982	Imported
Y12	Unknown	01.01.1985	Armour	?-1984	Imported
Y13	15.01.1980	01.01.1981	Armour	1980	Imported
Y14	Unknown	29.11.1982	Mixed	?-1982	Imported
Y15	21.05.1982	08.08.1984	PFC	1982-1984	PFC
Y16	08.04.1981	11.03.1983	Mixed	1981-1983	Imported/PFC
Y17	Unknown	01.09.1982	Armour	?-1982	Imported
Y18	24.02.1982	06.04.1983	Mixed	1982-1983	Imported/PFC
Y19	17.11.1982	06.04.1983	Mixed	1982-1983	Imported/PFC
Y20	15.05.1981	15.11.1982	Mixed	1981-1982	Imported
Y21	01.01.1982	01.01.1983	Mixed	1982	Imported

Aberdeen Royal Infirmary, Foresterhill

3.289 Dr Henry Watson, Director of the Aberdeen Haemophilia Centre, provided written evidence that three patients were considered to have acquired HIV as a result of treatment at Aberdeen on the basis that these patients had been treated only at the Aberdeen centre or had received minimal treatment elsewhere.^[291] All three patients had Haemophilia A (two with severe haemophilia and one with moderate haemophilia).^[292] One patient had been treated only with SNBTS product (PFC Factor VIII and cryoprecipitate) during the relevant period.^[293] This patient's last negative test for HIV was 14 July 1983 and he first tested positive for HIV on 18 December 1984.^[294] One patient received commercial product in small amounts in 1978 and 1979 (Baxter 'Hemofil') but otherwise received treatment only with SNBTS product (PFC FVIII and cryoprecipitate).^[295] This patient's last negative test for HIV is not known and the patient first tested positive for HIV in January 1985. One patient regularly received commercial product and from time to time received PFC product.^[296] Again, this patient's last negative test for HIV is not known and the patient first tested positive for HIV on 1 January 1985. Two of these patients were known to have died. If these patients were infected by Scottish products (which was clearly the position in the first and third cases), all had PFC Factor VIII in 1983 and 1984. Patient A3 is likely to have been infected by the batch implicated in the infection of the Edinburgh Cohort.^[297]

Ninewells Hospital, Dundee

3.290 Dr Ron Kerr, Director of the Dundee Haemophilia Centre, advised that no haemophilia patients were considered to have been infected with HIV as a result of treatment at Dundee.^[298]

Raigmore Hospital, Inverness

3.291 Dr Christopher Lush, Director of the Inverness Haemophilia Centre, advised that no haemophilia patients were considered to have been infected with HIV as a result of treatment at Inverness.^[299]

Results of Scottish Haemophilia Centre Directors' review of UKHCDO data

3.292 According to the Directors' evidence, the revised total number of haemophilia patients who contracted HIV while receiving treatment and care at Scottish centres was therefore 59.

Table 3.19: Total HIV infections

	Total	Dead	AIDS deaths*	Alive
Royal Infirmary of Edinburgh	23	19	14	4
Glasgow Royal Infirmary	12	10	6	2
Glasgow Yorkhill Hospital	21	8	5	13
Aberdeen Royal Infirmary	3	2	1	1
Total	59	39	26	20

*This column lists deaths attributable or possibly attributable to AIDS

UKHCDO updated report

3.293 Following the oral hearings, the UKHCDO undertook further work on its data and provided an updated statistics report.^[300] That report listed 73 patients with bleeding disorders first reported to the UKHCDO by Scottish haemophilia centres as testing positive for HIV.^[301] The report noted:

Scotland accounts for about 10% of the UK haemophilia population but only about 5% of the patients with bleeding disorders infected with HIV. 1383 patients with bleeding disorders are known to have been infected with HIV in the UK and were reported to NHD, of whom 73 were first reported to NHD by Scottish Centres.

3.294 Of the 73 patients who tested positive for HIV, 29 were reported from the Edinburgh centre, 24 from Glasgow Royal Infirmary, 11 from the RHSC, Glasgow, seven from Aberdeen, two from Inverness and none from Dundee.^[302]

3.295 The UKHCDO then analysed all of the data available on the National Haemophilia Database (NHD) to determine (a) which patients in the list of 73 were probably not infected in Scotland; and (b) which patients first reported to NHD by centres outwith Scotland were probably infected in Scotland. Five patients were probably infected outwith Scotland. The position of one additional patient was unclear: he was a severe Haemophilia B patient who had been treated extensively in Scotland and England in the nine months within which he seroconverted, with a variety of NHS and commercial products. Detailed analysis of the records of patients in the second group identified one patient who was probably infected in Scotland and two more who might have been infected either in Scotland or in England, where the data indicate treatment in both countries at critical periods but could not identify a single place where infection probably occurred.

3.296 According to the UKHCDO, at that stage the range of patients probably infected in Scotland, taking account of the uncertain cases, was therefore 68-70.^[303]

HPS HIV Diagnosis Database

3.297 Professor Goldberg gave evidence that following the development of an HIV antibody test in 1985, HPS (then known as the Communicable Diseases (Scotland) Unit) established an HIV Diagnosis Database for Scotland.^[304] The database holds data on all individuals diagnosed HIV antibody positive in Scotland and individuals previously diagnosed outside Scotland who came to reside in Scotland. Since 1989 a standardised National HIV Test Request Form has been used by laboratories; the form is made available to clinicians in clinical settings, allowing relevant information to be recorded by the clinician at the time of blood sampling.

3.298 In a letter dated 23 March 2011 from HPS to the NHS Scotland Central Legal Office, it was stated that, 'As at 31 December 2010, HPS had recorded 46 deaths among 76 haemophiliac cases who are presumed to have been infected via the receipt of contaminated blood products in Scotland'.^[305]

Further investigation

3.299 Discrepancies in the numbers of patients considered likely to have become infected with HIV as a result of treatment in Scotland for blood coagulation disorders, prompted a request to Dr Hay and Professor Ludlam to explain the differences and, if possible, to reconcile the data. They carried out a further exercise and on 27 February 2013 produced a short report that dealt comprehensively with the main questions posed.^[306] Professor Lowe and Dr Tait (Glasgow) and Dr Watson (Aberdeen), along with staff of the UK National Haemophilia Database (NHD) were involved in the collaboration.

3.300 Using patient data held locally, including details not recorded in the NHD, and the NHD data for treatments at centres other than those at which the patient was registered, enabled an accurate estimate to be made of the time each patient contracted HIV and the centre which administered the treatment responsible for the infection. In the course of the work of reviewing the records, a Scottish patient was identified who had not been registered or treated at any Scottish Haemophilia Centre but had probably been infected by treatment at a non-specialist Scottish Centre. It was concluded that this patient should be added to the Scottish Haemophilia Directors' previous total of 59, giving a new total of 60.

3.301 All 59 members of the Scottish Haemophilia Directors' initial group were included in the NHD data. The total number derived from preliminary review of the NHD as probably or likely to have been infected in Scotland was 74, one more than published in the report of April 2012. On detailed analysis of the treatment records and HIV blood test results of the patients, it was found that three patients were probably infected in England. Ten patients who had arrived in Scotland from abroad, and were registered directly at Scottish Haemophilia Centres, were shown to have been HIV positive when they arrived. One further patient from abroad was found to be anti-HIV negative on arrival in Scotland, but HIV positive shortly thereafter. On review of his treatment records it was concluded that he had been infected before arrival, but was in the 'window period' following infection, before antibodies were detectable, 'incubating' the disease on arrival. That was thought to fit better with the known incubation period of HIV than the alternative, which would have required infection immediately on arrival in Scotland, and an unusually short incubation time.

3.302 The result was that NHD data were fully reconciled with the SNBTS Haemophilia Directors' data. Sixty patients with bleeding disorders were infected with HIV in Scotland by treatment in Scotland.

3.303 The results from the collaborative investigation carried out by the Scottish Haemophilia Directors and Dr Hay with the cooperation of the UKHCDO are accepted as accurate. Health Protection Scotland's data are an accurate reflection of the information collected. However, the patients notified to HPS were those with a positive HIV test who were classified on the request form submitted to the virology laboratory in Scotland as either having haemophilia or having been a recipient of a clotting factor concentrate. HPS does not have access to information about where patients were likely to have acquired HIV infection. As a result, some data are likely to relate to individuals infected outwith Scotland who subsequently were tested in Scotland. The need to adjust the UKHCDO data for this factor demonstrates that it may be sufficiently significant in itself to undermine the reliability of the HPS numbers for present purposes. Overall however, the investigations carried out by Dr Hay, Professor Ludlam and others have now been sufficiently specific and detailed to make the outcome preferable to inferences drawn from HPS data.

3.304 Not all deaths due to AIDS were registered as such, and reporting of the condition while patients were alive may also have been less than complete because of the stigma associated with HIV infection and the AIDS complex of diseases. The detailed work undertaken in the course of the collaborative exercise underlines the difficulty in arriving at any conclusive view on the actual prevalence of HIV infection, even in a closely monitored cohort such as haemophilia patients, given the lack of precise information on some individual patients' cases. Although the figure of 60 may not be absolutely accurate it is unlikely that more precise figures will ever be established than those available from the collaborative exercise.

Source of infection - SNBTS or commercial product

3.305 Dr Bruce Cuthbertson gave evidence that the SNBTS considered it likely that in total 25 patients were infected as a result of treatment with SNBTS products (comprising 19 patients in Edinburgh, four patients at Glasgow Royal Infirmary and two patients in Aberdeen).^[307]

3.306 While in many cases it is difficult to identify the particular product that infected a patient with HIV, it appears in general that commercial blood products carried a greater risk of transmission of HIV than products produced by the SNBTS from plasma from Scottish blood donors.

3.307 Dr Cuthbertson said:

For those who were known to have received both SNBTS and commercial products, the assignment of probability [as to the source of infection] will partly be determined by date of first detection of antibody to HIV and the date of the last negative HIV test result. For those patients who received both products within a time period consistent with HIV infection, it is more likely that the source of infection was the commercial product. This supposition is based on SNBTS current understanding of the frequency of HIV transmission by commercial and NHS products at that time, and is further supported by the much lower HIV infection rate in Scotland when compared with those haemophilia populations which were treated exclusively with commercial FVIII of US manufacture .... This difference is found in many publications, including that from Moffat, Bloom and Mortimer, 1985 (Lancet 1, p935)^[308] who found a significant difference ... between the infection rate in patients treated with US FVIII concentrate and those treated with UK concentrate. Similarly, Kroner et al, 1994 (J. Acquired Immune Deficiency Syndromes, 7:279-286) reported that more than 90% of moderate and severe haemophiliacs treated with American product seroconverted to HIV.^[309]

3.308 While it does appear that there is support for the assumption relied on in allocating contentious cases, it has to be observed that the Scottish product was infectious in some cases, and, as the UKHCDO April 2012 report shows, allocation can still be problematic. Dr Cuthbertson's evidence is accepted, but with a note of caution: so far as it is based on assumption it is not wholly based on hard data and cannot be accepted as absolutely accurate.

3.309 The updated UKHCDO statistics report notes that Scotland accounts for about 10% of the UK haemophilia population but only about 5% of patients with bleeding disorders in the UK infected with HIV.^[310] In the updated report Dr Hay observed:

The relatively low proportion of Scottish patients infected with HIV reflects the fact that Scotland was more self-sufficient in blood and blood products from PFC than England during the period of risk. BPL fractionated most of England's requirement for factor IX but only approximately 40% of England's requirement for factor VIII. England was therefore far more dependent on imported factor VIII concentrates than Scotland. These imported products were largely manufactured from US-sourced plasma and plasma obtained from various other countries, including Africa. HIV spread earlier through the US and African donor population than the UK donor population and so patients using commercial products had a much higher risk of contracting HIV, especially early in the HIV epidemic in 1980, 1981 and 1982. Where English patients were maintained on a single brand of concentrate during the period of risk the risk of contracting HIV was much higher (approaching 100% in some centres) in those patients using US-sourced concentrates.^[311]

3.310 Newcastle and Liverpool were centres in that last category. Referring mainly to severely affected patients, Dr Hay explained:

In some centres, including the one that I worked in as a junior doctor, people were kept on one brand of concentrate as long as possible. It was felt that that might minimise their chance of contracting non-A non-B hepatitis. In fact, that proved to be completely false. It didn't really matter which concentrate they got from that perspective but it did provide us with evidence that those that just used English Factor VIII had approximately half the incidence of HIV observed in the group treated with commercial concentrates. I think that that was largely because HIV spread earlier into the US donor population than into the UK donor population. Of course, there may have been differences in the donor population between Scotland and England.^[312]

3.311 That there was a greater risk of HIV from commercial concentrates appears to be illustrated by the respective incidence of HIV infection in the Scottish centres.

3.312 From the figures supplied by the Scottish Haemophilia Directors, all but two of the 59 patients with haemophilia identified by them as infected with HIV as a result of treatment at a Scottish centre, had Haemophilia A. Approximately five per cent of patients with Haemophilia A registered at GRI were infected with HIV as a result of their treatment at that centre (10/211).^[313] Similarly, approximately 5% of patients with Haemophilia A registered at Aberdeen were infected with HIV as a result of their treatment at that centre (3/60). No haemophilia patients were considered to have been infected with HIV as a result of their treatment at the Dundee or Inverness centres. In the early 1980s all of these centres used primarily SNBTS rather than commercial products.^[314]

3.313 The Edinburgh Centre used mainly SNBTS products in the early 1980s. While there was a greater incidence of HIV at the Edinburgh Centre compared with the other adult centres in Scotland (approximately 13% (23/172) of patients with Haemophilia A registered at Edinburgh in 1985 were infected with HIV as a result of their treatment at that centre) it appears, as discussed above, that 18 of the 23 patients infected at Edinburgh were infected by a small number of batches of SNBTS/PFC Factor VIII. Infection by these batches of concentrate made up about 38% of the total number of Scottish infections.

3.314 In contrast, the greatest user of commercial concentrate in the Scottish centres in the early 1980s was RHSC Yorkhill, where almost 30% of the children with Haemophilia A registered at that centre in 1985 were infected with HIV (21/73).

3.315 In conclusion, in Scotland, as was the case elsewhere, commercial concentrates produced by US companies appear to have carried the greatest risk of transmitting HIV.

HIV - Secondary transmission

3.316 As noted in this Report at Chapter 8, Knowledge of HIV/Aids Now, paragraph 8.24, secondary transmission of HIV is a real risk, whether from mother to baby or as a result of sexual contact, though reported numbers are small. Professor Ludlam advised that he was not aware of any partners of HIV positive haemophilia patients who became infected with HIV.^[315] Dr Tait advised that to the recollection of staff in Edinburgh and Glasgow, no partners of haemophilia patients were HIV positive, albeit there were limited data in that regard.^[316] Dr Henry Watson, at Aberdeen, stated that there were no accurate data relating to HIV infection in the partners of infected patients.^[317] Beyond noting that some patients who contracted HIV as a result of infected blood or blood products may have inadvertently infected others, it is not possible for the Inquiry to make precise findings or estimates in that regard.

HIV as cause of death

3.317 As noted at paragraph 3.302 above, 60 patients with bleeding disorders were infected with HIV in Scotland by treatment in Scotland. Examination of the histories of 59 of those patients, reported in Table 3.19 above, showed that 26 had died of AIDS-related causes. The remaining patient was also recorded as having died of AIDS.^[318] Thus, of the 60 patients infected by treatment in Scotland, 27 died of AIDS-related causes.

3.318 This figure does not represent the total number of deaths from AIDS-related causes among patients in Scotland with bleeding disorders. Figures from the UKHCDO of 48 deaths^[319] and from HPS of 46 deaths^[320] include cases where the patient's HIV infection was acquired from treatment outwith Scotland.

3.319 The recent data provided by the UKHCDO do not include direct data on the patients registered with Scottish haemophilia centres who died of HIV. The data do, however, disclose that of 158 patients with Haemophilia A treated at Scottish Haemophilia Centres and who died between 1969 and 2011, 48 patients died of HIV (ie just over 30%).

3.320 In addition, as noted at paragraph 3.298 above, HPS noted that, as at 31 December 2010, they had recorded 46 deaths among 76 people with haemophilia assumed to have been infected through blood products in Scotland.

3.321 The pattern of deaths among HIV-infected patients shown in Table 3.19 above illustrates the relative mortality risks among patients who were young or old at the time of infection. Patients at RHSC, Glasgow have a significantly better survival record historically and, having survived into the era of effective treatment, a much better prognosis and life expectancy than patients who were older at the date of infection.

Patients infected with HIV as a result of blood transfusion

3.322 The fourth group of patients to be considered is referred to in paragraph 3.2, item iv, and consists of patients who were exposed to infection with HIV through transfusion with blood or blood components.

3.323 An HIV look-back exercise started in 1985 as a UK-wide initiative of the Health Protection Agency.^[321] The Blood Transfusion Services in England, Wales and Scotland agreed to participate when assured that the data would be sufficiently anonymised for no patient to be identified through it, and that the data would remain available to clinicians throughout the United Kingdom for research.^[322]

3.324 From the date of formal commencement of routine testing (15 October 1985) the look-back exercise traced the fate of all blood components from donations made during the preceding five years, by donors who were after that date found to be anti-HIV positive on returning to make a repeat donation. From a starting point of 39 anti-HIV positive donors with a history of previous donation, this process identified 10 anti-HIV positive patients in Scotland. In addition the SNBTS had information from the sporadic reports of clinicians, about patients with HIV infection where the sole risk factor was transfusion.

3.325 The data for infected donors in Scotland are shown in bar form in Figure 3.6.

Figure 3.6: HIV-infected blood donations: Scotland

3.326 The 39 anti-HIV positive return donors represented the numbers presenting at donor sessions between October 1985 and the end of 1997. Fourteen return donors were found to be HIV-positive between 1998 and 2009. The look-back exercise targeted the period when the highest concentrations of infected return donors presented and were tested, but it was not comprehensive. Return donors who continued to appear and test positive after the reference period for the look-back study were not included. While the possibility of additional return donors since 2009 appears remote, the inclusion of any who did appear might have led to identification of additional patients who received infected components. It is not possible, however, to deduce from these data the numbers of possible return donations after 2009.

3.327 A further difficulty that arose with the look-back exercise was that comprehensive archive samples were not available for the whole period of risk of transmission of HIV. Systematic storage of archive samples began in Edinburgh in mid-1984 and in Glasgow in 1986.^[323] In the circumstances, reliable information on numbers of infected donations was available from about 1985.

3.328 There was a supplementary source of data in returns from virology laboratories to HPS. As already noted at paragraph 3.297, in 1985 HPS (formerly SCIEH) put in place a reporting system throughout Scotland. The system used a single unified referral form for clinicians' use in requesting virology laboratories to carry out an HIV test. SCIEH automatically received a copy of the referral form for every test that proved positive. From that time, every transfusion patient testing positive for HIV at a Scottish virology laboratory was reported.^[324] But the data were anonymised, limiting the scope for further investigation.^[325] In 1985, the virus had only been present in the general population, and therefore by inference in the donor population, for about five years. When the test was introduced, the number of potentially dangerous previous donations from any individual donor was limited to that period.^[326] So far as available, the data were reflected in the final figures presented by the SNBTS/HPS.

3.329 As a result of concerns about the published information, all available data were reviewed for the Inquiry with a view to developing a more reliable picture of the information collected. The evidence was presented by Dr Gillon on behalf of the SNBTS and HPS. In total 18 patients are known to have contracted HIV infection as a result of blood transfusion in Scotland.^[327] Ten of these patients were identified through the targeted look-back exercise by the SNBTS following the introduction of HIV screening in October 1985. Eight patients were reported by their clinicians, independently of the look-back exercise, as possible transfusion-transmitted infections. In four of these eight cases it was possible to identify an HIV positive blood donor as the likely source of infection. Of the remaining four, a circumstantial connection was established in three cases, but it was not possible to establish a date of transfusion. Each of the three patients had received more than one transfusion, transfusion had occurred before testing for HIV had begun, and there were no archived samples. But the patient had no other likely source of infection.^[328] The remaining patient was reported to SCIEH by clinical staff at the Western Infirmary, Glasgow, but was not known to the SNBTS.^[329]

3.330 After further research, Dr Gillon provided a table showing information on probable dates of infection for all 18 patients.^[330]

3.331 In chronological order, the dates of transfusion of HIV-infected blood or blood components in these patients are shown in Table 3.20. These data represent the best information available to the SNBTS.

Table 3.20: Transfusion-transmitted HIV infection: dates of transfusion

Patient	Date of transfusion	Annual total
1	13 August 1983
7	31 December 1983
2	1983 or 1984	1983:2/3
6	11 January 1984
8	January 1984
4	March 1984
3	April 1984
14	25 April 1984
13	'early' 1984
12	Before June 1984
10	9 June 1984
15	September 1984
17	30 October 1984
5	November 1984
9	1984	1984:12/13
16	September 1985	1985:1
11	August 1986
18	August 1986	1986:2

3.332 For two patients, infective transfusion in August 1986, well after HIV screening of donors had been introduced, tends to confirm that the early anti-HIV tests were not 100% reliable in detecting infected donors.^[331] In particular, as antibody tests, they could not detect infection during the window period following infection and before antibodies were produced.

3.333 Eight patients were infected in Lothian, six in Greater Glasgow, three in Tayside and one in Lanarkshire health board areas. The distribution by Regional Transfusion Service was:

• Edinburgh and south east: eight
• Glasgow and west of Scotland: seven
• Dundee and Tayside: three.^[332]

Neither of the other two regions, Highlands and Grampian, had any cases of transfusion-transmitted HIV/AIDS.^[333]

3.334 The most critical years for confirmed infection were 1983 and 1984, a factor that will become significant when the date of introduction of testing is considered. It will also be significant that the first nine months of 1985 saw one case only of transfusion-transmitted HIV infection.

3.335 Fifteen of the 18 patients were known to have died as at 31 December 2010. The causes of death were not known, but Dr Gillon commented that, based on HPS data, many of these patients will have died of AIDS.^[334]

3.336 The figure of 18 patients known to have contracted HIV as a result of blood transfusion in Scotland is probably an underestimate. The actual incidence of HIV in Scottish donors between the assumed arrival of the HIV virus in the Scottish blood supply (around 1982) and the commencement of HIV screening of donors (October 1985) is not known. On general grounds, a significant (though not necessarily large) number of individuals infected with HIV by transfusion probably died of the underlying illness for which they were treated, or from other unrelated causes (such as the diseases associated with advancing age), before AIDS-related symptoms appeared. Similarly, an inherent problem of look-back is that some donors may have made HIV-infected donations before October 1985 but not presented to make further donations after the commencement of the exercise. The possible number of patients infected by such donors is likely to have been small, yet it remains unknown.

3.337 On the other hand, it is unlikely that there were more than a very few post-transfusion AIDS or AIDS-related deaths which remained unidentified among those who survived the typical asymptomatic post infection period. AIDS and the AIDS-related complex present as very serious, usually fatal, illnesses once they develop, and it is highly likely that almost every Scottish AIDS case will have had a cause attributed to it: eight of the 18 known cases were reported by clinicians.

3.338 The Preliminary Report included, at appendix 4, a table from the SCIEH (now HPS), showing the cumulative total, of AIDS cases as at 30 September 1999, AIDS deaths and HIV-infected persons registered or reported in Scotland. The table indicated that there were 33 recipients of blood or blood products recorded as being HIV positive, 18 male and 15 female. For the purposes of the table, 'blood product' does not appear to include blood products prescribed to haemophilia patients (e.g. clotting factor concentrate), as haemophilia patients are shown separately in the table. While not all of these 33 individuals necessarily received the infective blood transfusion in Scotland,^[335] the record does perhaps suggest that the figure of 18 patients discussed above may be a little low, albeit the true figure will not be as high as 33. The HPS figure included patients who might have been infected outside Scotland.

3.339 It is not possible to be more precise on the evidence available.

Co-infection with HIV and Hepatitis C

3.340 It is not known whether any of the 18 patients known to have been infected with HIV as a result of blood transfusion in Scotland were also infected with Hepatitis C.^[336]

3.341 It is likely, however, that all haemophilia patients who contracted HIV as a result of treatment with blood products made from large pools of plasma also contracted Hepatitis C. Dr Tait stated:

As we know, unfortunately some of the HIV patients would have died before Hepatitis C testing became available but I think it's a fairly reasonable assumption that virtually all HIV positive patients would also have been Hepatitis C positive.^[337]

Summary of conclusions

• Generally, contemporaneous records, whether of patients exposed to risk or of the incidence of infection, were either not maintained at all or were or have become over time, incomplete and unreliable.
• In the case of each of HIV/AIDS and HCV, there was a period when the agent of transmission was not identified, and no reliable diagnostic tests were available, and when effective recording of comprehensive data on the incidence of disease would therefore have been impossible.
• The Public Health (Infectious Diseases) (Scotland) Regulations 1975 did not achieve the reporting of viral hepatitis, pre-dated any requirement for the reporting of HIV and in any event were never enforced.
• Throughout the period when transmission of the index infections by blood, blood components and blood products exposed NHS patients to risk, there were no regulatory or other administrative systems capable of recording and monitoring the numbers of NHS patients in Scotland treated with blood and blood products, or the numbers exposed to risk of infection with the Hepatitis C virus and HIV, or the numbers contracting either or both such infections as a consequence of such treatment.
• The Blood Safety and Quality Regulations 2005 and the Public Health etc. (Scotland) Act 2008 have provided a statutory framework for improved reporting and recording of infection, and should be effective provided they are enforced and are supported by appropriate procedural rules.
• So far as patients with blood disorders are concerned, the National Haemophilia Database maintained by the UKHCDO and the records of individual haemophilia centres, as now adjusted, have provided the best evidence. However, due to anonymisation and reliance on voluntary compliance with requests for information and reporting, the records may not have yielded comprehensive information relating to infection and the consequences of infection for NHS patients.
• In the case of each of the index infections, the data extracted from records of known or recorded cases of transfusion-transmitted virus infection represent minimum numbers of patients who were infected. They probably do not reflect the actual number of patients who are likely to have contracted the disease in that way.
• So far as Term of Reference 4 is concerned, it is clear that recorded data cannot provide an accurate measurement of the extent of HCV or HIV infection in Scotland over the reference period, or of the risk of transmission of infection.
• None of the statistical analyses carried out provided a single reliable source of information. A sensitivity analysis was not carried out to test the assumptions made in varying conditions. Such statistical models as have been developed to date remain of doubtful validity as indicators of total exposure to risk and of transmission. The fundamental problem is that hard data are lacking and assumptions have to substitute for measurement over too wide a range of factors.
• The work carried out by and for the Inquiry suggests that further epidemiological investigation would not produce a more reliable estimate, at least without disproportionate expense. Nothing can be done now to improve the contemporaneous records, or to provide hard data indicating objectively the scale of the problem. The reporting system, such as it was, was never designed and assembled so as to be likely to be effective even if enforced. If the Scottish Government is persuaded that, for health policy and strategy, or budgeting or other reasons, it is necessary to develop a more accurate figure, it may be that further research and further expert opinion might eventually converge. That cannot be recommended by this Inquiry.
• Excluding the extremes, a wide range of values remain as indications of the possible incidence of infection. Only a rough and speculative estimate is possible, as set out in Table 3.21.

Table 3.21: Estimated numbers of NHS patients infected and outcomes

Infection	Estimated number infected	Estimated number of deaths from liver-related causes (HCV infection) or AIDS (HIV infection)
Bleeding disorder HCV	478[338]	Unknown overall: 21 to 30 deaths reported from various studies[339]
Transfusion-transmitted HCV	Around 2,500[340]	Unknown overall: 8 of 133, per Dr Gillon; 31 of 304 per Professor Goldberg[341]
Bleeding disorder HIV	60[342]	27 AIDS or AIDS-related deaths[343]
Transfusion-transmitted HIV/AIDS	18 minimum[344]	Up to 15[345]

3.342 Whatever the exact number of cases of transfusion-transmitted HCV might be, there are individuals currently infected with HCV as a result of blood transfusion who may not have been diagnosed. This topic is taken up in Chapter 35, An Investigation into the Steps Taken to Identify the Individuals who were Infected (Look-back). The question of further investigations to identify these patients is discussed there. Any data recovered would incidentally have a bearing on the statistics. The primary focus should clearly be on identifying those who may benefit from treatment and who, despite previous attempts to identify them, remain unknown to the NHS.

---

4. Experiences of the Patients and Their Families - Witness Statements >