I think we should show more faith in government

Photo by Tim Mossholder on Unsplash

Last Thursday, 21 May, the government announced that cases in Kenya will peak between August and September. They predict that at peak they will be announcing 200 new cases daily. Nearly two months ago, on 30th March, the government announced that it projected at least 10,000 cases of Covid-19 in Kenya by May. Its not clear whether that projection materialized given how they were doing just 600 daily tests at the time.

Across the world, governments have relied on projections to understand the spread of the coronavirus and how to respond to it. The projections are found by utilizing available data which provide a model for future scenarios. The modelling approaches of the government are not clear at this time. But from a birds eye-view, there are several reasons why projections related to the corona-virus fail to materialize. First of all, little is known about the virus, and more information continues to be uncovered as the world deals with it. For example, a recent report on the Atlantic indicates that the disease appears more transmittable by respiratory droplets and particles that we emit when we talk, sing, or cough, and less because of the door handles or surfaces we touch. Such information can have a bearing, for instance, on the infectiousness of the virus, also called the rate of reproduction. Thus, there could be factors or characteristics that models do not account for because no one knows or understands them yet. Secondly, public and personal interventions put in place like masks and social distancing, have an impact, which many projections may not account for. Because of the interventions, the virus is not distributed evenly in Kenya, its more incident in Nairobi and Mombasa, but projections probably rely on the total country population for analysis. Furthermore, the amount of data the government has collected so far may not reliably represent the total population. For instance, when they announced cases would peak in August-September, they had tested the most samples so far, 3000. Are they projecting a peak of 200 daily cases with 3000 tests? Wouldn’t peak cases be higher if they did more tests? Don’t few tests risk under-representing or missing the peak? So how much testing is enough?

The WHO recommends that countries should test “at the level of 10 negative tests to 1 positive.” the percentage of positive cases from total samples tested. Under the guideline, countries are testing correctly if the number of new cases are not more than ten per cent of total tests done. This statistic, the percentage of positive tests out of total tests, is referred to as the test positivity rate. Currently, Kenya’s positivity rate is 1 per cent as per Saturday's data, and 2 per cent from total tests done. Under the right circumstances, like proper random sampling, 2 per cent would definitively indicate incidence of the virus. So, what kind of sampling is the government doing?

The test positivity rate is more reliable to determine incidence of disease if the test data is from a random sample, which more accurately represents the overall population. It becomes less reliable as testing focuses on people who likely to have disease. At the onset of the pandemic, most countries focused their testing resources on people who were presenting with symptoms. Such tests were more likely to be positive. Currently, Kenya seems to mix different testing approaches. One approach is surveillance, where they find and test all contacts of the positive cases they identify. They also do targeted testing of hotspot communities. For instance, they have done this is in Kawangware and Old Town. Based on the recommendation by WHO, simply, a test positivity rate below ten per cent, is evidence that the government is testing randomly. But a second clue was the revelation that 70 per cent of cases the government uncovered were asymptomatic. That finding indicates that their strategy was picking people who were not necessarily presenting symptoms or severe sickness. But is the sample size large enough to represent the overall population?

Qualtrics is a website that has several survey tools including a sample size calculator. Using that calculator, for a country of 53 million, 360 samples, with a 5 percent margin of error, is ideal under a simple random sampling. For a much lower margin of error, 1 per cent, the ideal sample size would be 9602 tests/samples. Kenya’s maximum daily testing ability so far has been 3000 test samples. So should we be showing more faith in government projections? It seems so. Generally, the test positvity rate, which I deduce prevalence from, has been 1–2 per cent, but has been rising in the last week, and not necessarily because of increased testing. The highest , 28 May indicated 5.19 per cent, but the latest data (29th) saw a drop to 3.2 per cent. So is it safe to reopen?

Test Positivity in the Last 9 days

Again, from the WHO guideline on test positivity, it seems okay to deduce that the test data available is adequate to make projections and decisions on. The current test rate is 1.34 tests per 1000 people, or 1380 per one million people. In addition to the government’s 10000 cases projection, some private organizations have also made available their analysis. Rescue.co is a private emergency health service that has worked on one model to understand the spread of the virus in Kenya. Their model predicts the state of the pandemic depending on how the public applies the mitigation measures recommended to deal with it, like social distancing. According to their model, as of 5 May, under strict social distancing measures, like total lockdown, cases could peak at 416,000 in August. Under relaxed measures, Kenya could experience more than eight million cases at peak. Moreover, yesterday, researchers from various universities, under an organization called KAVI, predicted 2 million cases by February 2021. There approach was not reported. The most concerning aspect of our situation remains that the testing capacity hardly seems adequate to uncover these huge case numbers. It would seem that a confident reopening of the economy would entail a lot more testing than Kenya has.

8 day data

Since most Covid-19 cases are asymptomatic, experts recommend that authorities develop the ability to isolate positive cases, as they ease restrictions. You do not want many asymptomatic people going outside, as they could easily infect others, and cause a surge. Countries that were unable to implement an effective isolation regime early on, like Italy, US, and UK, were forced to do wholesale shutdowns. Those countries used the time in lockdown to ramp up their testing, but also some seemed to wait for cases to peak before easing the lockdown. Kenya’s measures were more relaxed, a curfew, travel bans out of counties where the virus was most prevalent, and restrictions on gathering. Unlike Italy or the US, we have not experienced a surge in cases, deaths, or rising demand on health facilities during the lockdown. The restrictions the government had put in place are probably the main reason why the virus has been controlled, so far. But are we doing so well as to stop these restrictions?

The only sure-fire way to open is when we can all access vaccines. Until then, countries have to balance restrictions and freedom depending on how they control the virus. Last Saturday, the President, in a live televised address mentioned that: “I have told my Cabinet Secretaries that we cannot stay in lockdown forever. But if we open the country by lifting the restrictions and the virus continues to spread, it will be our fault for not taking the precautions outlined.” People have a responsibility to be more careful but government should be prepared for the ramifications of the decisions they take. As far as this nosey writer is concerned, its the right call to relax some of the measures in place. But to whichever extent the government reopens, more people across the country will be exposed to the virus. In particular, rural areas, where majority of the elderly are, will now get exposed. In a press conference,CAS Mercy Mwangangi indicated that the government has prepared medical facilities and human resources across Kenya to address the pandemic. In any case, they likely have the best view of things so far, and you would imagine a responsible government would relax any measures to the extent they can mitigate rising risk.

journo + tech