Now that the notion of eradicating coronavirus from Sri Lanka hasn’t quite materialized, let’s not contemplate the basis of that objective. It is time that the Government and the officials adopt the scientific approach of disease control and prevention.

The objective of writing this blog is to bring some important facts to the attention of decision-makers. If they have already done so, I beg their pardon.

In order to keep this blog strictly objective, I will try to back my statistics with relevant studies and reports as much as possible.

For starters, the following is an attractive info-graphic that encapsulates the situation thus far

Control Vs Eradication

Eradication of a disease caused by a virus that can spread through human contact is certainly a tall task. In fact, there seem to be only 2 viral diseases that are fully eradicated to date.

See this article by the American Society of Microbiology

Apart from these, many countries have eradicated Polio through a series of sustained global initiatives that spanned several decades.

Disease control, on the other hand, has been achieved many times by many countries and regions. Some of them were raging pandemics such as the SARS outbreak in 2002 and H1N1 in 2009.

Past Virus Outbreaks provides some interesting facts on this

Define “Control” for Sri Lanka

One of the main questions that we need to answer as a nation is how we define the ‘controlled status’ for COVID19. Many countries define controlled status as a status where the number of infections is at a level that can be handled by the health system without undue stress.

Hence, understanding the capacity of a healthcare system is a very good starting point. The government-funded universal health care system in Sri Lanka is touted to be one of the best in Asia. Following infographic is from the Annual Health Bulletin 2017

While the number of beds to population is comparable to some of the more developed countries, critical care capacity appears to be comparatively low in Sri Lanka: the number of ICU beds is a widely used measure of critical care capacity. It is just shy of 500 beds according to the following report published in 2017.

Dedicated Capacity for COVID19 Treatment

Just like any other planning process, the best place to start is probably at the very end. We can work it back from thereon.

For simplicity, I’m going to assume some numbers which I think are reasonable. However, these numbers are mere assumptions for calculations

Maximum number of hospital beds that can potentially be dedicated to COVID19 treatment: 2000

It is my opinion that 2000 (=80000*2.5%) hospital beds can be a realistic target given the historical occupancy rates in hospitals in Sri Lanka. However, the entire health care system will have to be effectively utilized to achieve this target given the fact that the occupancy rate is not uniformly distributed amongst all hospital types

Critical Care Bottleneck in Sri Lanka?

Based on some key data that are emerging from countries such as China, Italy, and the USA the Hospitalizations to Critical Care ratio can be assumed at around 20%. This article summarizes data related to critical care and refers to original publications.

There is clearly an acute shortage of critical care facilities at the moment if we work back from the above data and assumptions. It stems from the fact that if 2000 patients are treated in hospitals at a given point of time, around 400 of them should probably require critical care, and there are only 500 ICU beds in total!

Therefore, it is imperative that authorities plan to increase the critical care capacity as soon as possible. Having said that, we need to be realistic about our ability to increase critical care capacity at short notice. Life supporting equipment needs to be vigorously tested before usage for obvious reasons.

There is another important dimension that also needs to be discussed with respect to the same topic.

There is a debate over the effectiveness of critical care for COVID19. Mortality rates as high as 50% are reported worldwide for patients who underwent critical care according to this article published by a highly reputed medical journal

Given the acute shortage of critical care facilities in Sri Lanka and due to the aforementioned concerns, we must formulate a critical care policy for COVID19 patients in Sri Lanka. This may involve the unpleasant topic of deprioritizing patients who will have extremely poor survival probability due to age and other pre-existing conditions.

However, in order to continue the discussion, I have assumed;

Maximum number of dedicated ICU beds for potential COVID19 treatment: 100 (=500*20%)

Calculating Backwards from the Critical Care bottleneck in SL

Without any attempts of prioritization, 100 ICU beds mean the ability to treat 500 patients at any given point of time. However, if we implement a scheme of prioritization, in a way such that it will only have a marginal impact on mortality (patients who have low survival rate are the ones who are deprioritized) we can choose to treat more patients, say 1000

Readers may notice that my above-mentioned assumption of 2000 hospital beds for the potential treatment of COVID19 is not restrictive in this case.

At this point we have defined the “Controlled Status”, at least from a strictly medical framework, to be 1000 hospitalized patients at a given point of time

Manageable Hospitalizations to Manageable Infection Rate

Based on studies conducted in countries that experienced major outbreaks such as China, Italy, Spain, the UK, and the USA, the average hospitalization period for COVID19 patients is about 14 days. The discharge guidelines also are set around this date.

I’ll refer to this article which calculates the average hospitalization period to be around 13 days from a sample patient cohort from China.

This leads us to a critical number. If we can effectively treat 1000 patients at a time and the average hospitalization period is about 14 days, we can admit about 70 (=1000/14) patients each day.

And this number leads to another extremely critical number in the planning process. That is, “The number of new infections that can be effectively handled”

We can derive this based on the hospitalization rate which is widely accepted at around 20%. Refer to this article amongst many

Hence, the number of new infections that can be effectively handled will be around 350 (70*100/20) per day!

The following figure summarizes all the back-calculation we just did!

Controlling COVID19 Pandemic in Sri Lanka

Now that we have a good definition of the ‘Controlled State’, albeit a set of assumptions made above, we can talk about measures that can be taken to control the spread. In other words, we can talk about how to keep the daily infections below 350

Unfortunately, the statistic that is constantly tracked and monitored in Sri Lanka, (just like any other part of the world) is not that reliable in this context. This statistic obviously being the number of reported cases. The number of reported cases is obviously not as same as the number of actual infections on a given day. I will not dwell too much on this as many readers are well aware of the drawbacks associated with reported cases.

For completeness, let me just mention that it is highly dependent on the amount of testing conducted. Also, it is a lagging indicator due to the symptom onset delay.

Widespread testing combined with the statistical estimation technique called ‘Nowcasting’ (as in fore-casting) the number of daily infections can be estimated. Nowcasting is very important in this context since there is a significant gap between someone getting infected and reported/detected. Working with lagging indicators can be disastrous in the epidemiology context.

Nevertheless, widespread testing has become a huge challenge for Sri Lanka as well as many other countries. There is an ongoing debate in Sri Lanka about the amount of adequate testing.

What is the Adequate Level of Testing for Sri Lanka?

Widespread and aggressive testing obviously helps the process of contact tracing. But can we determine the adequate amount of testing?

Let us consider a group of 1000 people. Assume 10 of them (1%) are infected. What is the chance of detecting at least 1 of the infected (out of 10) if we conduct 10 tests? It turns out the probability is quite low, a mere 9.6%. At 100 tests it reaches 65%. However, if there were 20 infected as opposed to 10, we will detect at least one of them almost certainly when we conduct 100 tests.

The following figure shows how this probability increase as we increase the number of tests.

This gives us a possible framework for testing. First, we need to identify the areas (population segments) of around 1000 ppl. We should only carry out testing if we suspect there is a possibility of at least 1% (=10 infections) infection rate. In other words, even if we have the slightest of suspicion (around 1%) we test. If not, we do not test that segment of 1000.

Given the data we have, it is highly unlikely that Sri Lanka has a 1% infection as we speak. Simply because 200,000 (1% out of 20 million population) infections would have meant significantly more hospitalizations and deaths than we have experienced so far.

Therefore, this elimination criteria should eliminate a very high % of the population. Let us assume it is around 95%. The remaining 5% (=1 million) would need 100,000 tests initially. That is 100 tests for each segment of 1000 in phase 1. After phase 1, we should only consider segments that had at least 1 positive result.

According to some recent news reports Sri Lanka has 100+ (Unfortunately I’m not able to verify this information) PCR testing equipment. Hence, we should be able to carry out these 100,000 tests in less than 10 days!

Testing, Lockdown and Quarantine Framework for Sri Lanka

The following framework could require about 100,000 tests in Phase 1 and would jump-start the economy in most of the areas of the country

Why Do We Have to Go Through All This Trouble?

Like any other country, Sri Lanka too cannot sustain a prolonged lockdown. The socio-economic impact can easily dismantle the very fabric of Sri Lankan society. The cost of lockdown can be as much as Rs 25,000 million (25,000 *1,000,000) per day! (Assuming 50% of the GDP is lost each day due to limited economic activity). A lot of things the society takes for granted, such as the government pay, pensions, etc. will be under significant financial stress if we don’t plan well and plan now!

CEO and Chief Data Scientist at Linear Squared.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store