Quantifying the potential value of entomological data collection for programmatic decision-making on malaria control in sub-Saharan African settings

Abstract

Background: The availability of many tools for malaria control leads to complex decisions regarding the most cost-effective intervention package based on local epidemiology. Mosquito characteristics influence the impact of vector control, but entomological surveillance is often limited due to a lack of resources in national malaria programmes.

Methods: This study quantified the monetary value of information provided by entomological data collection for programmatic decision-making using a mathematical model of Plasmodium falciparum transmission. The 3-year impact and cost of various intervention packages was simulated in different sub-Saharan African settings, including combinations of scaling-up insecticide-treated nets (ITN), switching to next-generation ITNs, and a treatment and prevention package. The DALYs averted and their net monetary benefit were compared at different cost-effectiveness thresholds and the value of resolving uncertainty in entomological model parameters was calculated.

Results: Across transmission settings and at cost-effectiveness thresholds over US$170 per DALY averted, the most cost-effective intervention package was switching to and scaling up pyrethroid-pyrrole ITNs combined with the treatment and prevention package. The median expected value of perfect information on the entomological indicators was US$0.05 (range 0.02-0.23) and US$0.17 (range 0.09-1.43) per person at risk at thresholds of US$75 and US$1000 per DALY averted, respectively. This represented less than 2% of the net monetary benefit of implementing the most cost-effective intervention package. Value of information estimates at cost-effectiveness thresholds over US$250 were higher than current investments into entomological monitoring by the US President's Malaria Initiative.

Conclusions: These results suggest that entomological data collection should not delay implementation of interventions with demonstrated efficacy in most settings, but that sustained investments into and use of entomological surveillance are nevertheless worthwhile and have broad value to national malaria programmes.

Keywords: Entomological surveillance; Health economics; Malaria; Mathematical modelling.

Fig. 1

Schematic of entomological indicators in the model and potential data sources. Uncertainty in the four entomological indicators affects the impact of different classes of ITN and IRS relative to all other interventions in the model. Data on these indicators can be obtained through routine entomological surveillance and other sources like experimental hut trials and statistical modelling, while the setting-specific impact of vector control can also be assessed directly using (post-implementation) impact evaluations. Estimates of indoor biting behaviour also require data on human activity in addition to entomological field studies

Fig. 2

Impact and cost-effectiveness of the different intervention packages relative to the corresponding baseline setting with existing interventions, stratified by seasonality. Bars represent the median across settings (including PfPR_2–10 5% to 40% and ITN use 20% to 60%), with error bars indicating the 95% credible interval (across settings and parameter variation). Points represent the median value in each prevalence and baseline ITN use setting (12 for each intervention per seasonality panel). The “treatment and prevention” package consists of a combination of treatment scale-up, SMC and/or RTS,S vaccination depending on the setting. A DALYs averted by the different intervention packages over 3 years. In seasonal settings, intervention packages including treatment and prevention avert more DALYs than in other seasonality settings because of the inclusion of SMC (not applicable in perennial settings and assumed to be already implemented at baseline in highly seasonal settings). B Net monetary benefit (NMB) of the different intervention packages over 3 years at cost-effectiveness thresholds of US$75, US$250 and US$1000 per DALY averted. A positive NMB indicates that the intervention is cost-effective compared to existing interventions at baseline; the optimal intervention package is the one with the highest average NMB. Calculation of the NMB depends on assumptions about the value of an averted DALY, therefore NMBs are higher for a cost-effectiveness threshold of US$1000 per DALY averted than the US$250 threshold

Fig. 3

Expected value of perfect information (EVPI) by prevalence and seasonality setting for the set of interventions A without IRS and B including IRS. The EVPI is shown for cost-effectiveness thresholds of US$75, US$250 and US$1000 per DALY averted and averaged across ITN use levels in each setting

Fig. 4

Value of information on different groups of entomological indicators for the set of interventions A without IRS and B including IRS. The distribution of value of information estimates across settings is shown when eliminating uncertainty in pyrethroid resistance and ITN entomological efficacy combined (expected value of partial perfect information (EVPPI) on ITN effectiveness), in the proportion of mosquito bites taken on humans and taken indoors/in bed combined (EVPPI on vector behaviour parameters), and the expected value of perfect information (EVPI) on all four entomological indicators combined. The last column for all indicators corresponds to the EVPI values in Table 3

Fig. 5

Comparison of the estimated expected value of perfect information (EVPI) with investments into entomological monitoring by the US President’s Malaria Initiative (PMI) in 2021–2023. A Distribution of PMI entomological surveillance funding across African countries. Investments in US dollars were divided by the population at risk of malaria in the respective country, with funding bands representing quartiles. B Comparison of these investments with the estimated EVPI for the set of interventions without IRS as shown in Table 3. Countries without PMI funding were excluded from this comparison