Consensus elements for observational research on COVID-19-related long-term outcomes.

ABSTRACT

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and its long-term outcomes may be jointly caused by a wide range of clinical, social, and economic characteristics. Studies aiming to identify mechanisms for SARS-CoV-2 morbidity and mortality must measure and account for these characteristics to arrive at unbiased, accurate conclusions. We sought to inform the design, measurement, and analysis of longitudinal studies of long-term outcomes among people infected with SARS-CoV-2. We fielded a survey to an interprofessional group of clinicians and scientists to identify factors associated with SARS-CoV-2 infection and subsequent outcomes. Using an iterative process, we refined the resulting list of factors into a consensus causal diagram relating infection and 12-month mortality. Finally, we operationalized concepts from the causal diagram into minimally sufficient adjustment sets using common medical record data elements. Total 31 investigators identified 49 potential risk factors for and 72 potential consequences of SARS-CoV-2 infection. Risk factors for infection with SARS-CoV-2 were grouped into five domains demographics, physical health, mental health, personal social, and economic factors, and external social and economic factors. Consequences of coronavirus disease 2019 (COVID-19) were grouped into clinical consequences, social consequences, and economic consequences. Risk factors for SARS-CoV-2 infection were developed into a consensus directed acyclic graph for mortality that included two minimally sufficient adjustment sets. We present a collectively developed and iteratively refined list of data elements for observational research in SARS-CoV-2 infection and disease. By accounting for these elements, studies aimed at identifying causal pathways for long-term outcomes of SARS-CoV-2 infection can be made more informative.