Abstract and scope of work: Although the majority of COVID-19 cases to date have now occurred in low- and middle-income countries, there are few primary accounts of the epidemiological dynamics and clinical severity of infection in such settings. This circumstance presents an untenable blind spot as policymakers seek to plan the scale-up and lifting of interventions against COVID-19, guided in part by models which have projected burden based on SARS-CoV-2 epidemiology in high-income settings. Continuing an ongoing collaboration with the health ministries of Tamil Nadu and Andhra Pradesh in South India, we aim to clarify fundamental aspects of transmission dynamics and epidemiologic burden in this LMIC setting, and to inform response, through real-time analyses of comprehensive testing datasets from clinical settings, contact tracing, and serological surveillance. Our specific aims are:
(1) To assess age-specific infection fatality ratios on the basis of paired serological and mortality data, and compare these to estimates from high-income settings;
(2) To fit age-structured models of transmission to longitudinal age-specific seroprevalence data to understand the role of differing age groups in transmission. We will also compare results to data from prospectively-tested contacts of known positive cases to assess consistency with inferences from contact-tracing, and the completeness of such case-finding efforts.
(3) To assess changes in transmission intensity associated with changes in country-wide and state-level lockdown policies, and to use fitted models to estimate infections, cases, and deaths averted by these measures. Short-term forward projections of transmission dynamics will estimate the potential burden associated with continuation or relaxation of specific nonpharmaceutical interventions to inform decision- making.
To date, our work in collaboration with the health ministries of Andhra Pradesh and Tamil Nadu has centered on assessments of Rt based on PCR testing surveillance data, and differences in transmission across exposure settings and contact types (age, sex, relationship) based on contact tracing data (pre-print available from https://doi.org/10.1101/2020.07.14.20153643). Analyses as of early June included over 64,000 contacts of 4,000 cases, making this one of the largest SARS-CoV-2 contact tracing datasets available and the only one to date from a resource-poor setting. Extensions of the ongoing work to address serosurveillance data will alleviate biases associated with changes in diagnosis and PCR testing practices, and enable estimation of severity measures from a denominator of all infected persons.
The collaboration has to date been undertaken without funding for the PI or other investigators. Urgent Grant program funding is urgently needed to allow its continuation and expansion.