The minimum sample of individuals who would need to be tested would depend on the anticipated rate of covid-19 prevalence in a region
Despite efforts to ramp up covid-19 testing, India remains among countries with the lowest testing rate. The lack of surveillance has meant that there is no credible estimate of the true number of infected individuals in India. A new article in The Lancet by S.V. Subramanian and K.S. James of the Harvard Centre for Population and Development Studies recommends that India use the National Family Health Survey (NFHS) mechanism to monitor the spread of covid-19 in the population.
In the absence of universal testing, a random sample-based framework should be used for testing, they argue. Since NFHS has a ready sampling framework and the required infrastructure, using the same for covid-19 data collection would keep operational costs low. The only major expense would be the laboratory costs for testing samples.
The minimum sample of individuals who would need to be tested would depend on the anticipated rate of covid-19 prevalence in a region. The lower the anticipated prevalence, the higher would be the minimum sample needed to reliably estimate the true prevalence.
Estimates based on such wider samples would be more reliable than those based on the testing of selected at-risk individuals: those with influenza-like symptoms, those who have had contact with infected individuals, healthcare professionals, or those with a travel history to an affected region.
The authors show that the strategy of assessing disease risk based on selective testing of at-risk individuals has not proved effective in the past. In 2002, the National Intelligence Council of the US projected 25 million HIV-positive cases for India, with a 3-4% prevalence rate in Indian adults, based on such a strategy. NFHS estimates showed much lower morbidity (2.5 million HIV-positive cases) and a prevalence rate of only 0.28% in adults. This discrepancy shows the shortcomings of extrapolating disease risk from selective testing, the researchers argue.