Covid testing: Learning from Getz

Too much noise everywhere. Understandably so. Such is the concern and panic. Quality can often be lost in the crowd with opinion and analyses overdose – which is not at all times directed in the right direction. Getz Pharma, one of Pakistan’s top pharmaceutical facility, conducted Covid-19 census testing of 100 percent of its more than 1500 employees. This got lost in the crowd and remained largely unreported in mainstream media.

Getz Pharma does not call it a study, as the objective was to make the company’s facility a Covid free zone. But at the same time, it has offered the government epidemiologists to use the rich data generated to extrapolate and project the prevalence across the country.

Some 7 percent of mostly asymptomatic healthy people tested positive. The testing provides a representative sample of any other largescale manufacturing facility “in a large urban city with similar demographics like Karachi, Lahore, Faisalabad etc.” Getz Pharma assigns a high level of statistical confidence in this conclusion, also pointing at the accuracy of test route taken which is as high as 99 percent, as compared to the oft-used PCR tests with an accuracy range of 30-64 percent.

Be that as it may, this is by far the single largest example of cluster testing in Pakistan, and if the government has not yet benefitted from the data, it should and soon. What is intriguing to know is that over 90 percent of the employees tested were completely asymptomatic – yet the incidence of positive results is remarkably close to the national incidence at over 8 percent. Minus the Zaireen, the number will come further down. Recall that most of the national testing is very selective, and based on symptoms, travel or contract history.

There may also be a lesson or two in which testing kids to use in order to get more accurate results. It is not known if the inaccuracy of the PCR tests is on the lower or upper side. But a completely random sample testing giving values so close to that of highly selective testing – does indicate there is high chance that kits used by majority at national level may have missed a few positive results.

An alternative way to look at this is the inaccuracy may well balance both sides, and that the Getz Pharma results show that the virus’ prevalence is high, regardless. This looks a plausible explanation, as the sample size is well diversified. Extrapolation will surely be done by those who do it, but even very conservative estimates, will tell that people who have, have had or will have the virus – could be in millions in the urban centers of Pakistan.

Forget testing ten times of the possible cases. No country in the world has done that. Maybe the estimated number could bring about a shift in strategy from testing focus to other areas, such as prevention, hospitalization etc.

Getz Pharma has also observed, and it apparently makes sense that the results “may be suggestive of an innate immunity in the Pakistani population”. This is a very important piece of the puzzle and could be decisive in managing affairs around reopening the economy, relaxing restrictions and so on. The death rate so far does instill hope. Watch out for more on the mortality related analysis tomorrow.

Meanwhile, Getz Pharma may well have shown a way to the other big MNCs and local companies, who employ thousands across various urban facilities. Maybe it is time to go beyond joining hands for “12000 ration bags” and give back in the form of 100 percent employ testing. Big data, around this time, is no less vital than other cash and kind contributions.