RESUMEN
Objectives Since COVID-19 first emerged in 2019, mathematical models have been developed to predict transmission and provide insight into disease control strategies. A key research need now is for models to inform long-term vaccination policy. We aimed to review the variety of existing modelling methods, in order to identify gaps in the literature and highlight areas for future model development. Study design This study was a systematic review. Methods We searched PubMed, Embase and Scopus from 1 January 2019 to 6 February 2023 for peer-reviewed, English-language articles describing age-structured, dynamic, mathematical models of COVID-19 transmission and vaccination in high-income countries that include waning immunity or reinfection. We extracted details of the structure, features and approach of each model and combined them in a narrative synthesis. Results Of the 1109 articles screened, 47 were included. Most studies used deterministic, compartmental models set in Europe or North America that simulated a time horizon of 3.5 years or less. Common outcomes included cases, hospital utilisation and deaths. Only nine models included long COVID, costs, life-years or quality of life-related measures. Two studies explored the potential impact of new variants beyond Omicron. Conclusions This review demonstrates a need for long-term models that focus on outcome measures such as quality-adjusted life years, the population-level effects of long COVID and the cost-effectiveness of future policies -- all of which are essential considerations in the planning of long-term vaccination strategies.