(19 April 2020) COVID19: Why diagnostics?

There is a lot of clamour about testing for COVID19. My view is that testing for testing sake is of no use. Every diagnostic test should provide an answer to a useful question being asked with a larger purpose in mind.

I thought I will use this blog to put down some useful purposes that testing can serve. (A disclaimer — there different types of diagnostic tests which answer different questions; so, please do not assume that every test does the same thing. Also different tests will have different levels of certainty with which they give results.)

Purposes for doing diagnostic tests for COVID19: 

  1. To support clinicians in their choice and sequence of therapy
    1. To identify if a patient has COVID19 to guide treatment
    2. To quantify viral load and severity
    3. To decide when to discharge the patient
  2. To guide R&D on drugs and vaccines
    1. To provide vital information and clues to R&D scientists on the progression of the disease
    2. To select candidates for clinical trials in a better manner so to come up with more meaningful results
    3. To quantitatively measure the impact of therapy
  3. To  support decisions of  public health authorities
    1. To identify those people who can spread the virus. To identify those who need to be quarantined.
    2. To identify those who can be allowed to leave quarantine
    3. To get better insights into who is more vulnerable and therefore should be protected
    4. To screen travellers and migrants at the borders for potential risks
  4. To provide important data (especially, as basis for assumptions and cross checking if the model predicts reality accurately) for mathematical models that are used to support decisions of the Government such as
    1. Imposing lockdowns
      1. When?
      2. Where?
    2. Lifting lockdown
      1. When?
      2. Where?
    3. For planning operations, budgets, medical services/supplies needed etc:
      1. How large is the real spread of the disease?
      2. What proportion of the population is already immune to the disease?
        1. Infected and recovered
        2. Residual immunity from earlier
      3. When infected, what proportion of the population will be
        1. asymptomatic,
        2. symptomatic (home isolated or quarantined)
        3. symptomatic (hospitalised)
        4. will die

Realism in testing:

  • Clearly one cannot be testing everybody very frequently
  • Testing methods have their own limitations and that should be understood by all: a) the sample has to be collected correctly so as to not miss out the virus due to poor sample collection, b) every test method has some limits of sensitivity and specificity, c) different test methods answer different questions.
  • The test methods which detect the SARS-COV-2 specific  parts of the virus RNA tell you “if the virus RNA is present in detectable concentrations inside you at that time in the location from where the sample was taken“. What are the assumptions here? a) we know exactly which gene is unique to SARS-COV-2, b) virus is present in concentrations above minimum detectable concentration, c) one has chosen the right place to sample from, d) one has collected sample properly so as to not miss the virus.
  • The test methods which detects an antibody that the body has produced in response to the presence of the SARS-COV-2 virus, tell you that “the virus was present in the body at some point and the body responded by producing antibodies to counter it“. What are the assumptions here? a) the antibody being detected is specific to the SARS-COV-2, b) the antibody is not being confused with another one looking similar, c) the antibody is present in concentrations higher than detectable limits, d) the person did not get an external infusion of antibodies or a vaccine causing production of the antibody.