Murky data calls into question quarantine strategy (FT Comment March 25 2020)

Opinion Coronavirus

Murky data calls into question quarantine strategy

Seasonal flus teach the virtues of calmly dealing with things as they come


MARCH 25, 2020

The writer is researching transformational medical tests and treatments and is the author of ‘Venturesome Economy’

Like sepsis, a life-threatening, uncontrolled immune response to infections, draconian efforts to contain the coronavirus outbreak may cause more harm than the infection itself. Yet the measures now paralysing the western world before many have actually died are based on heroic, unverifiable assumptions about infection and death rates.

The clamour for widespread testing is loud, and the unavailability of test kits harshly criticised. People with a range of symptoms — or even no symptoms — are overwhelming clinics with their demands for testing.

But, as with any new virus, problems of reliable testing are acute. Errors can arise from mistakes in the underlying science, flaws in the design and construction of instruments and inexperience in administering tests and interpreting the results. We cannot even reliably estimate the accuracy of the new testing procedures.

Turning promising laboratory research into practically useful tests — and developing reliable testing protocols — requires extended trial and error; the recent Theranos blood-testing scam should warn us about trusting miraculous leaps. Training clinicians to operate and interpret new tests is another huge practical problem as the history of mammography shows.

But, without reliable tests and well-maintained registries of infections, epidemiologists cannot escape classic “garbage-in, garbage-out” problems of modelling infections.

Alarmists point to arresting reports of high proportions of fatal cases. According to one published in the New England Journal of Medicine of patients in Wuhan — the epicentre of China’s coronavirus outbreak — 1.4 per cent of 1,099 “laboratory confirmed” cases had died. This is about 10 times the estimated proportion of deaths attributed to flu infections in the US. But the American flu estimates probably include lower-risk patients diagnosed without laboratory confirmation. The US Centers for Disease Control and Prevention recommend laboratory tests only for hospitalised or high-risk patients.

This means there is no basis yet for projecting that for patients of similar risk profiles, death rates from coronavirus infections will be astronomically higher than for flu patients. Indeed, we can take comfort from the Wuhan data that 98.6 per cent of patients whose condition prompted laboratory testing survived.

The wide range of estimates related to the 2009 swine flu pandemic reflect the questionable reliability of such numbers. Estimated infections ranged from 700m to 1.2bn — greater than the total estimated in 1918-19 Spanish flu infections. The range of estimated deaths from swine flu was even wider — from about 150,000 to 575,000 deaths. A later study found that swine flu resulted in no higher serious illness than yearly seasonal flus.

German Chancellor Angela Merkel says that coronavirus infections will probably infect 60-70 per cent of the population without much harm to most patients. But, she warns, even with small proportions requiring intensive care in hospitals, rapid growth of infections could overwhelm the healthcare system.

Even if self-isolation that slows the rate of infections — known as flattening the curve — is the right answer, how best to do this is murky. If infections are already widespread, the scope for reducing the growth rate is limited. But, without reliable mass testing we cannot know how many people are now infected. Predictions of a second wave of new cases are also highly speculative.

Similarly, we cannot predict what combination of “tough” measures will slow progression. Sending potentially infected college students back to their hometowns may accelerate the spread of the disease. Declarations of emergency have led hoarders to crowd into supermarkets. And, if we cannot reliably track infection rates, we cannot select and tweak interventions to maximise their effectiveness.

Recent numbers from China suggest that harsh restrictions have slowed diagnosed infections. But there is no assurance that the slowdowns will last: pandemics typically continue for longer than a year.

Quarantine, travel bans, closing bars and public performances and emptying college campuses have jeopardised the livelihoods of economically vulnerable individuals who have no financial cushion to tide them over. Financial distress in turn exacts a well-known toll on physical and mental health.

Experience with seasonal flus teaches the virtues of calmly dealing with things as they come. Death rates in the US attributed to flu infections swing widely and unpredictably — from 12,000 in good years to 61,000 in bad years. Intensive care units, whose capacity is optimised for normal demands, face acute stress in bad years. And, every new season may bring mutations as deadly as the catastrophic Spanish flu.

Lockdowns each winter could potentially turn a bad year into a good one, saving 50,000 deaths, reducing stress on intensive care units and forestalling catastrophes. Fatal car accidents would also fall, further reducing demands on emergency care.

Yet sensibly, we reject safety at all costs: we risk more deaths to live better lives. That’s a lesson worth remembering now.