Wednesday, May 27, 2020

Comparitive Snapshot of COVID19 in USA, Italy and India


  This note covers three issues, the effect of testing on the number of COVID19 cases, the spread of infected cases and the fatality rate among the infected. It uses three very diverse countries, India, USA and India to illustrate where India stands at this point of time, from the perspective of stabilization of the situation with respect to the disease.
Recall that the index W0 measures whether the SARS Corona Virus 2 problem is increasing if W0 >1 and decreasing if W0  < 1 and over if W0 is 0.  The good news is that in Italy, one of the countryies most severely affected by COVID19, the virus seems almost completely under control with W0 at 0.02 (figure 1a). USA has achieved significant success in slowing the spread, but with an index stabilized at 0.1 the spread is not over (figures 1b). India is however, still at stuck at a W0 of 3.5-3.7, and some way to go till it reaches a safe level of index of 0.1 like USA or completes the process like Italy.
 
Table 1a: Spread of COVID19 in Italy measured by W0 index
 

 Table 1b: Spread of COVID19 in USA, measured by W0 index
 

 
Table 1c: Spread of COVID19 in India measured by W0 index
 
 
  One of the questions that has excited worldwide debate, is to what extent testing(cvt) increases the number of Corona virus identified cases(cvd). Our analysis suggests that this happens in the initial phase of testing, with the number of cases increasing rapidly with number of tests, but after some time the curve flattens out and there is little relationship between testing and number of cases. 

Figure 2a: Effect of testing on detected cases: USA
 

 Figure 2b: Effect of Testing on Detected Cases: Italy
 
 
USA and Italy have reached the latter phase, but India is still in the first phase (figure 2a,2b & 2c). The turning point came at a log (cvt) value of around 14 in USA and 13 in Italy. In India the turning point is still not visible  at 14.8(figure 2c). This difference likely due to the difference in population size, with US population 5 times Italy’s and India’s population more than 4 times USA.

Figure 2c: Effect of Testing on Detected Cases: India
 

 
    I estimate the following model for the effect of testing (cvt) on registered coronal virus cases(cvc).
Log(cvc) = a + b log(cvt) = c t, where t is time variable which starts on the day the first case is registered. The results are as follows:
 
Table: Effect of Testing on registered cases
 
 Note Except for &. All other coefficients are significant at 1%
 
This shows that the effect of testing on registered case is lowest in USA (0.74) and highest in Italy(1.7). India (1.6) is lower than Italy, but more than double that of USA.

     The fatality rates provide the best news for India, relative to the other two countries. The fatality rate is still at an extremely high rate of 14.3% and on a mildly increasing trend (figure 3, right hand scale). The fatality rate in USA has bottomed out at around 6% in USA (figure 3, left hand scale). The fatality rate in India is about half that of the USA and on a very gradually declining trend (figure 3, LHS). It should be noted that even if there is variation in the accurate identification of number of COVID19 cases in different countries this would not affect the fatality rate of identified/registered cases. Further even in poor countries less hospital facilities relative to population size, serious cases are highly likely to make their way to hospitals, so the fatality rate could also be biased in upper direction. I therefore assume no bias.
 
Figure 3: Fatality Rate (%) in Italy, USA and India
 

 
Note: This work is part of a joint project with Surjit Bhalla, Executive Director, IMF, and the result of jointly collected data (Johns Hopkins and Oxford univresity data bases) and many areas of joint analysis and discussion. Any views expressed in this note should not, however, be attributed to him.