Part 3: The Third Way. How Hyper-Targeted COVID Policies Can Save Lives (and Livelihoods) in Africa
What should African governments do in response to COVID-19?
Two competing narratives have emerged in the debate over policies aimed at reducing COVID transmission in sub-Saharan Africa.
Public health experts warn that an unabated epidemic could wreak havoc in sub-Saharan Africa, overwhelming already stretched health systems and challenging hard-won gains against other diseases like HIV, tuberculosis, and malaria. Epidemic models (here and here) predict millions will be infected and hundreds of thousands will die. These sobering figures have led many African countries to adopt physical distancing measures to slow transmission and buy time to prepare.
The second narrative focuses on the detrimental effect that mitigation policies have on vulnerable populations. For those living at the edge of subsistence, mobility restrictions and business closures can lead to malnutrition and lack of basic necessities. Critics of physical distancing argue that “lockdowns” would do more harm in low-resource countries than COVID-19.
These tension between slowing the epidemic and economic activity are being played out across the globe, but in sub-Saharan Africa the stakes are much higher. With millions already living in poverty, the trade-offs are stark: Lives lost to unmitigated COVID-19 must be weighed against lives lost to the effects of pushing people into abject poverty.
For African leaders, these competing narratives about COVID-19 offer an impossible Sophie’s Choice. Take coronavirus seriously, impose physical distancing policies, and see your economy tank and many people starve? Or prioritize the economy, accept COVID-19 infections and deaths as inevitable, and endure the suffering until herd immunity arrives?
We think there’s a third way that can blunt the epidemic at manageable economic and social cost: Using hyper-targeted physical distancing policies, coupled with targeted interventions to mitigate any resulting social and economic harms.
In our most recent post, we showed how physical distancing can save lives in sub-Saharan Africa by reducing the total number of infections, not just by minimizing pressure on stretched health systems. Our post also noted that policymakers can use granular, localized data on economic and social vulnerability to mitigate those consequences.
This post looks at how to do that, using existing data to model the likely spread of the epidemic and assess how vulnerable different areas are to economic and social disruption.
Let’s see what this kind of modeling looks like in Malawi:
Previously, we shared some initial analyses conducted in support of Malawi’s Ministry of Health. The Ministry is providing us with continual feedback, which we incorporate into each iteration of our epidemiological model.
We’ve now added the capability to apply policies at the district level. This means that we can predict the health effects of enforcing lockdowns or other measures in certain districts, but not others.
These charts show the predicted effects of applying restrictions in different areas across the country. The blue and brown lines reflect the effect of imposing targeted physical-distancing restrictions in different sets of major urban centers.
As we can see, sustained social distancing in a few urban centers would significantly improve nationwide outcomes:
Focusing distancing measures on the highest-risk areas would be more feasible than a nationwide lockdown and would enable policymakers to precisely target the limited resources available for social support.
A key point here is that the choice facing governments of low-resource countries is not between doing nothing and full national lockdowns. Hyper-local data allow analysts to model, and policymakers (at all levels, from national authorities down to districts) to choose, narrowly tailored measures that reduce collateral social damage.
Next, let’s bring the population’s vulnerability to the economic and social consequences of COVID-related restrictions into the analysis.
For this model, we created an economic vulnerability index to assess, at a high geographic resolution, those areas which might be most vulnerable to the effects of lockdowns, movement restrictions, or closure of essential businesses.
Our model used these inputs:
We start with poverty data, which we use to rank districts within the country:
Next, let’s bring in food security, availability of soap and water, and wasting and stunting data . (Fraym is providing the last four free of charge for COVID-19 support in Malawi.)
These data will go into our vulnerability index, but they’re also useful on their own: Policymakers can use these data to better target food relief or to provide hand-washing facilities as a public-health intervention.
It’s worth highlighting one nuance in the choice of inputs here. The poverty data is available only down to the District level — a fairly large administrative unit (there are 28 in the country). Fraym’s data on availability of water, soap, stunting, and wasting, by contrast, is available at square-kilometer resolution. Adding hygiene, stunting, and wasting as additional proxies for vulnerability allowed us to achieve more granular geographic resolution.
Putting these indicators together, we can generate an overall ranking of areas by economic vulnerability:
Now, we can align estimates of economic vulnerability with epidemiological predictions of severity: current infections and the percentage of population at high risk for severe COVID-19 (darker areas are at higher risk).
These maps can help policymakers assess where the epidemic poses the greatest risk and how vulnerable each of those areas would be to economic disruption from lockdowns or physical distancing. Policymakers can use that knowledge to conduct cost-benefit analysis and to mitigate economic consequences by targeting food aid and other relief.
And these interventions can be hyper-local: The data allows us to pinpoint specific areas within a district where we can conduct targeted testing, tracing, and physical distancing approaches and corresponding economic and social support that are contextually relevant for each community. These interventions can be developed in consultation with community leaders, using the data and their insights to develop shared approaches.
We welcome your comments. For more information about our approach click here.