Dear Editor,

THE surprising increase in COVID-19 cases raises concern of an under-detection of the virus. The level of testing in the ten administrative regions should be increased to a level where it is statistically significant as a percentage of the population in each Region. Once a better understanding of the level of the virus in society is understood, the timing of the phased reopening of the country will be more safely accomplished. Recent increased infections in the teaching community does raise concern. The consequences of over-testing is less severe than those of an undetected increase in infected persons.

On a per capita basis, our testing should be pushed to hundreds per day. Our world in data website shows that in key locations in South America, as the level of testing increased, so has the number of infections per number of tests. If held constant, an assumed rate of increase of the spread of the virus can be calculated. However, more importantly, the data provides a clearer view of the level of infection within the populace. Containment of the virus will benefit from this approach of increased testing, and will also help better prepare the Ministry of Public Health to cater for the needed level of corrective measures to be taken, in addition to determining the required supplies.

The flattening of the infection curve, provided associated containment is implemented, should follow once the level of detection testing becomes more statistically significant within each Region. The Ministry should allow for the sample size to have a confidence level of 95% ±5% of the measured value. For example, if we assumed that each sparsely populated region is at minimum 10% of the population, then 139 tests should be carried out in that region. Similar assumptions can be made along the spectrum of percentage of population.

However, it is best to err on the side of over-testing and create a minimum testing level in each region of approximately 400, with triple that amount in Region 4. This creates a factor of safety that benefits the public, and can be applied to the testing/ percentage of population spectrum. Once the infection rate within the regional sample is determined, then expansion of the associated regional testing should follow accordingly, with the frequency of testing & retesting driven by the reduction & increase in the infection rate. I am hopeful that the Ministry’s in-house statisticians will be able to add some additional insight into this matter.

Regards,
Jamil Changlee