Some Thoughts on COVID19 in Sri Lanka – Statistical Analysis | Daily News

Some Thoughts on COVID19 in Sri Lanka – Statistical Analysis

COVID19 is a new viral illness that can affect respiratory system of a person. It has now spread in 211 countries (ihttps://virusncov.com) in different intensities. The World Health Organization (WHO) has declared the COVID19 as an outbreak with a global health emergency in January 2020. The experts at Johns Hopkins Medicine, Baltimore, USA claim that Coronavirus was totally unprecedented. The normal life of Sri Lankans has been highly affected by this outbreak since middle of March although we do not have a higher intensity like in other countries. Based on available data in different countries, it is very unlikely that COVID19 spread in Sri Lanka much faster. Thus we need to provide evidence-based scientific conclusions to the Ministers and the higher officials in order to make their decisions more rational.

Comparison among Selected Countries

I have noticed that the daily trend of new COVID19 cases in Sri Lanka compared to the daily trend of new cases in Italy is inconclusive. Because I have noted that there is no significant correlation (r = 0.06) between the two trends. Furthermore, the basic autocorrelation analysis suggests that the number of cases in Sri Lanka seems to be happening randomly, while in Italy there is a significant dependence structure.

Thus, in my opinion, it is not reasonable to say that we need to wait for some more days to see the spread of the disease based on the trends between the two countries. This would mislead the decision makers for some extent.

I believe the following statistics and conditional probabilities (Table 1) can be used to estimate the pattern in the future. I used the data extracted from the world COVID19 website: https://virusncov.com for 6 countries including Italy and India.

Due to random nature it is not easy to model new cases of COVID19 in Sri Lanka. However, based on the temporal pattern a crude estimation for new cases is possible on daily basis and the predicted value for 7th April is 7±4.

Estimating RO Based on Case Count Approach

Based on the local daily data available on the websites from 13 March to 6 April, I was able to compute the number of Freely Moving Infected (FMI) cases and Currently Infected (CI) cases on daily basis assuming that each FMI case roams for 5 days. The total number of FMI cases during the above period = 759

Total number of CI cases during the above period = 129

However, as we have not yet tested all the possible cases and doctors are claiming that all the positive cases so far not detected, the total number of CI cases was assumed as three times of the actually infected cases during that period. Thus a proxy value for R0 can be taken as, (3*129)/759 = 387/759 = 0.51[95% CI=(0.47, 0.55)]

If the actual value was used the R0 is 0.17 [95% CI = 0.14 and 0.20]. All these results confirm that the R0 value for COVID19 in Sri Lanka is less than 1, provided people follow the health instructions. Assuming the expected number of COVID19 cases reported on 07 of April is 387, I computed the expected number of COVID19 cases in Sri Lanka by 5 days intervals till 08th of May. This is only for the use of policy makers, but I did not use the actual values due to obvious reasons explained above. 

Based on the results in Table 1, the following conclusions can be drawn for the attention of policy-makers and doctors who are engaged in the fight against COVID19.

•The number of total cases by far is extremely low in Sri Lanka compared to other countries. In almost all countries, the percentages of active cases are more than 70% with an exception in South Korea.

•Of those active cases, the percentage of critical cases is less than 5% with an exceptional situation in Sweden. In Sri Lanka it is 4%. This is indeed a good indication to say that we do not have a higher number of active cases at present.

•That is, if we test another 10000 suspected individuals, we would expect only 96 critical cases which Sri Lankan health system can manage.

•It should also be noted that the above conditional probabilities can be used as good estimators for long term predictions as those findings were based on more number of days.

•However, of the closed cases, the percentages of either dead or recovered vary among the countries.

•The recovery rate has been the highest in South Korea (96%) followed by Sri Lanka (88%). That shows the good work of the Sri Lankan medical staff and other supportive staff.

•These results very clearly indicate that COVID19 does not spread as fast as in other countries, provided that people are abide by the instructions issued by the health authorities.


There is 1 Comment

Add new comment