9 November 2022: Expert Group on Statistics - A

The rest of the day had been given over to the postponed appearance of representatives of the Expert Group on Statistics. A friend who lectures in the Mathematics Department of a University recognised most of the names of those on the Expert Group as being outstanding in each of their respective fields. Their latter appearance at the Inquiry was reportedly down to so many of them being required to provide information to the Government in relation to the Covid-19 situation. So, with such venerable panel members, what was the probability that viewers and listeners would understand everything being said by these masters of number crunching? Look out for the Chi-squared analysis to follow.

Interestingly, the first panel member made his oath to God. What were the chances of a statistician believer in God existing existing?

Counsel took the opportunity to clarify a matter on numbers which were part of a document released by the Inquiry. There had been media reports about the numbers related to children which could have led to confusion or inaccuracy of interpretation.

The appearing experts’ backgrounds gave a short overview of their career development. These included infectious diseases, bio-statistical work, Covid statistics, hospital-specific child death and Shipman investigations, among other topics. The other panel members had a similarly relevant range of important and useful experiences in their CVs.

From the outset, one of the panel members wanted to explain that while the focus of their work was all about numbers, they recognised that each number represents a person who has died or been badly affected by infected blood. The panel members were always aware of the human experiences behind all the data. The conclusions were all based on data, but not all of it was complete or assuredly accurate. It was also noted that particularly for older time periods, there was no electronically available data. There was also the factor of people changing the definitions that lead to data being recorded differently. Some things were not recorded at all. There are also differences in data collection processes, for example, between England and Scotland. It was stated that things are easier these days, but for some aspects of their report some conclusions had to be qualified. Where possible, the group were able to comment on the confidence or uncertainty of results by giving intervals of confidence or subjective statements based on expert judgement summed up as a low to high range.

It was explained that the original task sent to the panel by the Inquiry covered a wide range of requests in the form of questions to be answered. However, due to the Covid-19 factor and the draw made on the panel members’ time to respond to that, some of the areas were not able to be addressed in the ways initially sought. (So, it can been seen how Covid and the response to it has continued to have negative impacts on vulnerable people. Will the missing analyses ever be carried out and made public, or will that information forever be lost to the black hole of Covid consequences?)

The Executive Summary of the panel report included the basic numbers of deaths from HIV and HCV arising from contaminated blood product use. It was assumed that everyone who had HIV was also infected with HCV. For transfusion patients, of the 79 HIV infected people, 85% had died. It was stated there were 26,800 transfusion patients infected with HIV. Over 21,000 had died.

For bleeding disorder patients, the three sources were firstly the Macfarlane Trust, secondly the UK Haemophilia Centre Doctors Organisation (UKHCDO) which came from the National Haemophilia Database (NHD), and thirdly the UK Health Security Agency (UKHSA). The NHD does not cover individual patient care, so that was a factor. As far as the panel knows, there was no collaboration between the three sources to compare data. UKHSA broke the infections down to three categories; haemophilia, other blood products, and blood products (undetermined). These were not considered to be fully distinct from each other. The route of infection was largely based on self-reporting by patients, and this was sometimes unknown or was prone to mis-reporting due to people being unwilling to admit to, say, illicit drug use. Other limitations were highlighted, such as the infections by the Macfarlane Trust only including direct NHS infections and not partners or children infected by association. It was noted that the UKHSA included the largest single number as being in the category of “haemophilia + undetermined” due to so many records not being available.

The three sources for HIV/bleeding disorder deaths come up with roughly the same figures. It was taken by the panel that the figure of around 1,250 people with bleeding disorders have been infected with HIV, with three-quarters having died. There was lower confidence in the data from death certificates, in part because the raw data was not collected together into one set, and in part due to the resistance or uncertainly of recording a new virus as being a cause of death. There was a comparison between HIV deaths and Covid-19 deaths in this regard. Things tend to improve after a diagnostic test becomes available. There can also be a tendency for families not to have certain causes of death listed, although doctors are meant to record deaths accurately. It was highlighted that the Inquiry had received examples of incomplete or inaccurate death certificates. There was also the possibility of a doctor seeking not to list a condition, even if the family had not requested it be missed out. The panel saw the fact of there being discrepancies in the numbers between known cases versus smaller number of cases after death as a factor to consider.

The larger stream of patients, according to their individual records, who had bleeding disorders and were infected with HCV highlighted the differences between what was recorded on individual records compared to the macro figures held by the UKHCDO. It was mentioned that many people will have been HCV infected before being HIV infected. The panel was able to see how people must have been tested and diagnosed and a period of time passing before them being told. The timings between infecting, testing, diagnosing, and telling had to be taken into consideration. Double-counting risks were lessened by knowing that the UKHCDO database automatically shifted someone to HCV infected once it was recorded that they had received factor concentrates.

The Skipton Fund was another source of data, but it was not too helpful since it did not record peoples’ demographic information to their viral status. The UKHCDO had advised the panel that it was reasonable to assume that all HIV infected people were also HCV infected.