Amitava Banerjee: Communicating risk in the time of corona

I entered the national lockdown in March six days early, when the UK had 1950 cases and 81 deaths from coronavirus. My lockdown was driven partly by concerns that we were not following lessons from Wuhan, but mainly by a question that day in my clinic from a long-standing patient with heart failure. I often said he was fitter than me, so he was confused at being at “high-risk” (“clinically vulnerable”), as per the list of conditions announced on 16 March 2020, with recommended stricter “social isolation”. 

Every affected country reported increased deaths with older age and certain conditions, such as heart disease. I wanted to communicate my patient’s individual risk to him. I wanted to investigate whether risk of death differed between conditions on the government’s list, from chronic kidney disease to diabetes, to check we were isolating the right people. Finally, I suspected unchecked infection rates would create thousands of deaths. My lockdown was with my laptop. I did not see my family for four days.  

By 20th March, our four-strong research team (Laura Pasea, Spiros Denaxas, Harry Hemingway, and me) had convinced twelve colleagues to join us in this research. Everybody contributed with skills from virology to statistics via phone calls, emails and Microsoft Teams. On 22nd March a further “extremely vulnerable” list of conditions was released, and we published our preprint after sending to the Chief Medical Officer’s office. The new list included 1.5 million individuals who should stay at home for at least 12 weeks (“shielding”), including those on chemotherapy. 

The following evening, the Prime Minister announced lockdown. We showed that 10% infection rate (likely in March/April) would lead to 30 000 and 70 000 excess deaths over one year, only preventable by reducing infection rate. With refinements, we published the research in The Lancet on 12 May, showing 20% of UK adults are “high risk”, the day after the prime minister announced the beginning of easing of lockdown. As of 18 September, there have been at least 381 614 confirmed cases and 41 705 deaths, now including colleagues, friends, and several patients. 

I am in a family of strong women. My wife, a dentist, returned to work full-time in June after “the curve flattened”, PPE supplies improved, and clear professional guidance emerged (the aerosol-generating risk is obviously among the highest in dentistry among health professions). She was home-teacher to our 6-year old daughter (the other strong madam in our house), who returned to school last month. There is another strong woman I missed during lockdown. My mother, in her sixties with diabetes, is in Yorkshire, and was mostly self-isolating. Since the prime minister’s announcement on 11 May, she wanted to know if and when she should return to teaching. She used to ask regularly if my exercise might lead me up North from Barnet (with her granddaughter of course). The relief of being able to see each other a few times, is now replaced by anxiety as case numbers rise and local lockdowns become a reality. 

Whereas we have guidance for extremely vulnerable individuals, it is confusing for people like my mother, who represent a larger proportion of the population. Over 90% of excess deaths occur in this category. Moreover, care of underlying conditions is likely to have been reduced, probably due to patients not attending hospital and health system strain. To avoid further tragedy, we must focus our efforts here, protecting them from infection and ensuring treatment of their comorbidities. 

We urgently need to communicate risk, regardless of covid-19, to patients, carers, researchers and policymakers, and that information needs to be simple, yet as specific to individuals as possible. During intensive care shifts at the Nightingale Hospital, I saw how evidence could help conversations with relatives in difficult times. The government has commissioned a specific tool to provide covid-19-specific risk information, which will help these discussions between clinicians and patients. 

“Listen to your patient, he is telling you his diagnosis,” said William Osler. My patient told me the health system diagnosis at an early stage in the pandemic, which still rings true now. Even in a health system which is universally accessible with available patient-level data for clinical practice and research, he needed more specific risk information. Which data, which risk information, and for which audience are important questions, which researchers, policymakers, and clinicians cannot answer without patients and members of the public. We need to reflect on which health data should be accessible for urgent and sometimes immediate research in public health emergencies. Data has long been labelled the “new oil”. Much more than that, it could be the “new drug” or the “new vaccine”.

Amitava Banerjee, Institute of Health Informatics, University College London, London.

Competing interests: None declared