Accuracy of COVID-19 Test

Are U Asking the Right Question?

Faulty test kits and failed test methods are already slowing the race to be ahead of SARS-CoV-2. Gross errors in test accuracy, upto 60% are noticed! 

By the way, What does this accuracy mean? 

With testing being the definitive constraint of the emerging pandemic COVID-19, the question needs clarity. It is also important for most of us to understand it we are going to be increasingly dependent on tests results and diagnostics to keep us a watch on our health, if not already.

 Asking just, what is the accuracy of test and getting a number (%) may actually tell you not much, rather can give you either complacency or anxiety.

What is the accuracy or effectiveness of test? We need to translate this into practical questions that helps in taking right decision. These are:

If I undergo a test, what is the chance that the test truly detects my condition. So, if I am declared positive for COVID-19, what is the chance that I am truly infected? If I am declared negative, what is the chance that I am not infected? A third factor, so important during epidemic is, how does this effectiveness change and clinicians take decision if I belong to a high prevalence group versus a low prevalence group?

These are also the questions we ask in management science, when we test a hypothesis, e.g. what is the chance that a hypothesis or assumption will deliver the bottom line results it claims it will. What is the chance that the achieved results are only because of the proposed hypothesis or assumption. And why do you claim that the problem in my organization is prevalent enough to be solved by your generic method. So much important is doing the right test to get the right answer!

Remember, in the case of emerging disease, where the science of test is still evolving, it is too much to ask for 100% effectiveness, and there lies the nervousness and controversy.

For starters, Effectiveness of test comprises of two parameters Sensitivity and Specificity.

Sensitivity is the percentage of cases detected positive from a given sample of positive population. So, if Sensitivity is 90% then 10% people go undetected. This 10% is called False Negative; if care is not taken, the false negatives can spread the disease unknowingly.

Specificity is the percentage of cases ruled negative by the test from a given sample of negative population.  So, if it is 90% then 10% of the people go declared positive even though they are negative. These 10% are called False Positive and may have to go through painful process of walking through sensitive areas of clinics or hospitals and face severe administrative hassles.

It must be noted that the two parameters are inter-related, i.e., a test with high sensitivity would have low sufficiency, and vice versa. 

SARS-CoV2 is a new virus and the whole world is still trying to understand its dynamics. So, the test systems are new and need their effectiveness to be carefully ascertained. The 10% false detection can be substantial in an emerging scenario like that of COVID-19 when the R0 factor is significant and could cause terrible errors in predicting and preparing remedial actions. It can put significant burden on infrastructure when you are still chasing the virus and not leading.

In addition to Sensitivity and Specificity, clinicians examining patients also look at a third factor, ‘Positive Predictive Value’ before they give their decision. For more details on testing effectiveness and how final recommendations are made check this link on Testing Effectiveness. This value changes dramatically depending upon prevalence of disease.  Using same test method (device) in red zones ie. high prevalence scenario and in green areas , i.e. low prevalence zone, would give different positive predictive value of the test.

Even when tests are being conducted, the speed at which COVID-19 is moving, it makes our goal to be ahead of virus really challenging. And, it is possible that several would roam undetected and yet shed virus on others (read, un-isolated pre-symptomatic and asymptomatic cases, and add to this the false negative cases). An effective screening mechanism is the key though.. that is what the whole world has been after so far...