Investigating time-dependent COVID-19 pandemic mental health data: Challenges and opportunities of using panel data analysis

Amidst the widespread ramifications of the Coronavirus Disease 2019 (COVID-19) pandemic, its impact on population mental health has attracted considerable interest [1], with the potential for adverse mental health consequences widely recognised at an early stage.The spring of 2020 saw a proliferation of online surveys, many focusing on mental health and well-being [2], but unlikely to provide robust research output because of design limitations.However, among these numerous initiatives, some important and more carefully considered projects were set up such as the large, prospective panel survey reported by Feifei and colleagues in PLOS Medicine [3].This follow-up of an online sample of over 57,000 adults in England from March 2020 to April 2022 was able to evaluate associations of depressive/anxiety symptoms with both contextual factors (the "stringency" of lockdown policy and COVID-19 deaths) as well as with individual factors (such as confidence in institutions, concerns about COVID-19, and social support).Particular advantages lie in the large size of the sample, the relatively high follow-up rates, and frequently repeated measures over the time period of interest.Of course, not everyone has access to or interest in online surveys and, although steps were taken with oversampling and weighting techniques to improve representation, the people who participated in the study are unlikely to be nationally representative, as acknowledged by the authors.Within these limitations, the study does provide a good opportunity to measure population-level changes in reported symptoms of depression and anxiety over the course of the pandemic and to investigate the factors influencing these.In particular (and perhaps rather bravely, given continuing acrimonious social media debates), the authors attempt to disentangle the causal impact of the pandemic from that of the policies taken to reduce viral transmission.There are self-evident challenges in this exercise with so many events occurring over the same time periods-the direct and indirect consequences of infections, the direct and indirect consequences of "lockdown" and other policies, co-occurring changes in the national economic outlook, changes in healthcare provision and access.Although factors such as isolation, financial instability, and uncertainty about the future are obvious potential explanations for the upswing in symptoms of depression and anxiety observed during the pandemic [4], they are likely to depend substantially on individuals' personal circumstances.For example, for one person, a working-from-home policy might provide relief from social stressors and improve work-life balance, saving commute travel time and costs; for another, this may result in stressful family relationships, challenging childcare, or loneliness.Adolescents and young adults appear to have been particularly vulnerable to early-pandemic adverse mental health [5], but continued fears about infection and other health threats, as well as self-imposed isolation, may translate into a higher mental health burden for older adults in years to come.No attempt was