PART 4: FUTURE SAR-COV-2 AND ASSOCIATED COVID-19 MODELS ARE ONLY AS GOOD AS THE DATA THAT GOES INTO THEM
BY PROF DOUGLAS BOATENG
Without a doubt forecasting and scenario planning are an essential component of any pandemic management strategy. Obtaining data related to disease spread, infection rates, recovery rates, etc. can help leaders to make informed and science-based decisions.
Unfortunately, many of the inputs for scenario planning are outdated and sometimes do not exist. For example, the following information are difficult to obtain in most African countries:
- Population density
- Hospital bed counts
- Cultural dynamics
- “True” spend on public health
- Weather and humidity, etc.
Based on modelling from the Imperial College London, if African nations did not impose a lockdown and carried on with business as usual (i.e. no social distancing, attend mass funerals, parties, churches, no wearing of FPM, no hand washing, etc.), the continent could see over three hundred thousand (300,000) COVID-19 deaths in 2020. UNECA postulated that fatalities in Africa could exceed three (3) million. The WHO has predicted a possible one hundred and ninety (190,000) deaths in Africa from approximately two hundred and fifty (250) million infections. These very contradictory projections clearly indicate a problem with the underlying assumptions in each of these “predictory” models.
Whilst these dire projections gave African governments a wakeup call and prompted important decisions relating to lockdown requirements, going forward there is a need to base future predictions on local data.
Hypothetical scenarios presented by leading experts continue to be revised as countries collect more data to refine their forecasting models. The problem emerges when such forecasting is not based on regional or local demographics and statistics. In addition to this, outputs can portray the wrong picture if the underlying assumptions are incorrect.
What is becoming increasingly clear as we begin to navigate the realities of life with COVID-19 is that forecasting and planning needs to be increasingly adapted to local data. In addition to this, global or first world policies surrounding COVID-19 cannot be uniformly adopted by developing nations without taking the realities of the local context and resources into consideration.
In conclusion, the accuracy of any scenario model depends on the quality of the data that can be speedily gathered. If the implicit assumptions that underlie the modelling are outdated, then the output is not likely to meet expectations either.
Lessons learned from this pandemic are that nations must invest in the necessary data collection infrastructure and related human capital to be able to collect store and analyse data for policy decision making and public education.