9 November 2022: Expert Group on Statistics - D

Question: How many statisticians does it take to change a light bulb? Answer: 3.26, on average.

The chapter of the report covering transfusion infected HCV patients was still the topic under consideration. Some of the audience must have been apprised of this slow death by numerology and duly absented themselves. They missed out on hearing about how the “annual risk of dying” is commonly referred to by statisticians as “the hazard”. Then there was a short focus on the situation related to teaching hospitals. The previous paradigm of just giving a woman in childbirth a blood transfusion to “top them up” has decreased over time. Just as the number crunching looked like it might be becoming more relatable, the panel warned of the possible need to put a “wet towel” around the head for the next explanation. They were not wrong. The chart analysis became brain paralysis. Hazards related to dying seem to be like background radiation, always there but not too worrying until you build up reasons, such as age and illnesses, which make you more susceptible to increasingly lesser doses of radioactivity until the background become toxic in itself.

Infectious HCV has a hazard ration of 1.53 but it is seen as difficult to work out the overall effect of having chronic HCV because there are so many aspects involved. The 1.53 figure means that your risk of death is 50% greater than it would have been had you not had HCV. One in three deaths of the infected group will be attributable to HCV. The ability to come up with this figure was down, in part, to the support of the Inquiry to allow research to be carried out, including by making available data that was not available from other sources. The life expectancy factors improve for people as medical science progresses, but of course, they (we) are all getting older and that is itself a contributing factor for consideration in the mix.

It was stated that there actually was good Hepatitis C epidemiology, but this is down to the studies carried out on IV drug users. While some of this data is useful, there are significant differences such as age and co-morbidities. The gaps are around earlier research on the impact on contaminated blood infected persons. Suddenly, there came to be a rather strange section about numerically coding names in relation to notifiable diseases. This was a way of maintaining confidentiality and so raising the confidence to report. By the way, vowels don’t count and because the letters d and t sound the same, they are coded as 3, obviously.

The issue of “calibration” was the next exciting and engaging topic. Amazingly, the keyboard this writer was using stopped working, so i …..

… ly got to the last question. It was about recommendations for data collection. It was recognised how much improved things were compared to past years. The prospect of another blood-borne virus needed to be considered, so it might be useful to have stored samples and linked databases; recognising the need to due confidentiality. The NHD was a possible option to build on, with cross-referencing and improved rigour which would improve confidence in the process. Reference was made to the Goldacre review to build a better trusted research environment. Linkages needed to allow GP records to be more central. There appeared to be other possible recommendations, so the panel suggested they get together outwith the hearing session and work out a more considered response to prepare for the Inquiry. This was welcomed by the Chair.