Prioritizing COVID-19 vaccination efforts and dose allocation within Madagascar.

BACKGROUND: While mass COVID-19 vaccination programs are underway in high-income countries, limited availability of doses has resulted in few vaccines administered in low and middle income countries (LMICs). The COVID-19 Vaccines Global Access (COVAX) is a WHO-led initiative to promote vaccine access equity to LMICs and is providing many of the doses available in these settings. However, initial doses are limited and countries, such as Madagascar, need to develop prioritization schemes to maximize the benefits of vaccination with very limited supplies. There is some consensus that dose deployment should initially target health care workers, and those who are more vulnerable including older individuals. However, questions of geographic deployment remain, in particular associated with limits around vaccine access and delivery capacity in underserved communities, for example in rural areas that may also include substantial proportions of the population. METHODS: To address these questions, we developed a mathematical model of SARS-CoV-2 transmission dynamics and simulated various vaccination allocation strategies for Madagascar. Simulated strategies were based on a number of possible geographical prioritization schemes, testing sensitivity to initial susceptibility in the population, and evaluating the potential of tests for previous infection. RESULTS: Using cumulative deaths due to COVID-19 as the main outcome of interest, our results indicate that distributing the number of vaccine doses according to the number of elderly living in the region or according to the population size results in a greater reduction of mortality compared to distributing doses based on the reported number of cases and deaths. The benefits of vaccination strategies are diminished if the burden (and thus accumulated immunity) has been greatest in the most populous regions, but the overall strategy ranking remains comparable. If rapid tests for prior immunity may be swiftly and effectively delivered, there is potential for considerable gain in mortality averted, but considering delivery limitations modulates this. CONCLUSION: At a subnational scale, our results support the strategy adopted by the COVAX initiative at a global scale.

Cross-sectional cycle threshold values reflect epidemic dynamics of COVID-19 in Madagascar.

As the national reference laboratory for febrile illness in Madagascar, we processed samples from the first epidemic wave of COVID-19, between March and September 2020. We fit generalized additive models to cycle threshold (Ct) value data from our RT-qPCR platform, demonstrating a peak in high viral load, low-Ct value infections temporally coincident with peak epidemic growth rates estimated in real time from publicly-reported incidence data and retrospectively from our own laboratory testing data across three administrative regions. We additionally demonstrate a statistically significant effect of duration of time since infection onset on Ct value, suggesting that Ct value can be used as a biomarker of the stage at which an individual is sampled in the course of an infection trajectory. As an extension, the population-level Ct distribution at a given timepoint can be used to estimate population-level epidemiological dynamics. We illustrate this concept by adopting a recently-developed, nested modeling approach, embedding a within-host viral kinetics model within a population-level Susceptible-Exposed-Infectious-Recovered (SEIR) framework, to mechanistically estimate epidemic growth rates from cross-sectional Ct distributions across three regions in Madagascar. We find that Ct-derived epidemic growth estimates slightly precede those derived from incidence data across the first epidemic wave, suggesting delays in surveillance and case reporting. Our findings indicate that public reporting of Ct values could offer an important resource for epidemiological inference in low surveillance settings, enabling forecasts of impending incidence peaks in regions with limited case reporting.

Existing human mobility data sources poorly predicted the spatial spread of SARS-CoV-2 in Madagascar.

For emerging epidemics such as the COVID-19 pandemic, quantifying travel is a key component of developing accurate predictive models of disease spread to inform public health planning. However, in many LMICs, traditional data sets on travel such as commuting surveys as well as non-traditional sources such as mobile phone data are lacking, or, where available, have only rarely been leveraged by the public health community. Evaluating the accuracy of available data to measure transmission-relevant travel may be further hampered by limited reporting of suspected and laboratory confirmed infections. Here, we leverage case data collected as part of a COVID-19 dashboard collated via daily reports from the Malagasy authorities on reported cases of SARS-CoV-2 across the 22 regions of Madagascar. We compare the order of the timing of when cases were reported with predictions from a SARS-CoV-2 metapopulation model of Madagascar informed using various measures of connectivity including a gravity model based on different measures of distance, Internal Migration Flow data, and mobile phone data. Overall, the models based on mobile phone connectivity and the gravity-based on Euclidean distance best predicted the observed spread. The ranks of the regions most remote from the capital were more difficult to predict but interestingly, regions where the mobile phone connectivity model was more accurate differed from those where the gravity model was most accurate. This suggests that there may be additional features of mobility or connectivity that were consistently underestimated using all approaches but are epidemiologically relevant. This work highlights the importance of data availability and strengthening collaboration among different institutions with access to critical data - models are only as good as the data that they use, so building towards effective data-sharing pipelines is essential.

Existing human mobility data sources poorly predicted the spatial spread of SARS-CoV-2 in Madagascar.

For emerging epidemics such as the COVID-19 pandemic, quantifying travel is a key component of developing accurate predictive models of disease spread to inform public health planning. However, in many LMICs, traditional data sets on travel such as commuting surveys as well as non-traditional sources such as mobile phone data are lacking, or, where available, have only rarely been leveraged by the public health community. Evaluating the accuracy of available data to measure transmission-relevant travel may be further hampered by limited reporting of suspected and laboratory confirmed infections. Here, we leverage case data collected as part of a COVID-19 dashboard collated via daily reports from the Malagasy authorities on reported cases of SARS-CoV-2 across the 22 regions of Madagascar. We compare the order of the timing of when cases were reported with predictions from a SARS-CoV-2 metapopulation model of Madagascar informed using various measures of connectivity including a gravity model based on different measures of distance, Internal Migration Flow data, and mobile phone data. Overall, the models based on mobile phone connectivity and the gravity-based on Euclidean distance best predicted the observed spread. The ranks of the regions most remote from the capital were more difficult to predict but interestingly, regions where the mobile phone connectivity model was more accurate differed from those where the gravity model was most accurate. This suggests that there may be additional features of mobility or connectivity that were consistently underestimated using all approaches, but are epidemiologically relevant. This work highlights the importance of data availability and strengthening collaboration among different institutions with access to critical data - models are only as good as the data that they use, so building towards effective data-sharing pipelines is essential.

Cross-sectional cycle threshold values reflect epidemic dynamics of COVID-19 in Madagascar.

As the national reference laboratory for febrile illness in Madagascar, we processed samples from the first epidemic wave of COVID-19, between March and September 2020. We fit generalized additive models to cycle threshold (C t ) value data from our RT-qPCR platform, demonstrating a peak in high viral load, low-C t value infections temporally coincident with peak epidemic growth rates estimated in real time from publicly-reported incidence data and retrospectively from our own laboratory testing data across three administrative regions. We additionally demonstrate a statistically significant effect of duration of time since infection onset on C t value, suggesting that C t value can be used as a biomarker of the stage at which an individual is sampled in the course of an infection trajectory. As an extension, the population-level C t distribution at a given timepoint can be used to estimate population-level epidemiological dynamics. We illustrate this concept by adopting a recently-developed, nested modeling approach, embedding a within-host viral kinetics model within a population-level Susceptible-Exposed-Infectious-Recovered (SEIR) framework, to mechanistically estimate epidemic growth rates from cross-sectional C t distributions across three regions in Madagascar. We find that C t -derived epidemic growth estimates slightly precede those derived from incidence data across the first epidemic wave, suggesting delays in surveillance and case reporting. Our findings indicate that public reporting of C t values could offer an important resource for epidemiological inference in low surveillance settings, enabling forecasts of impending incidence peaks in regions with limited case reporting.