The prevalence of insomnia in different COVID-19 policy phases: Longitudinal evidence from ITA.LI – Italian Lives

Abstract

Abstract Background This study investigated changes in the prevalence of insomnia in Italy during COVID-19, starting from the first lockdown period (8 March 2020). We hypothesized that lockdown precipitated increased prevalence of insomnia symptoms relative to the pre-pandemic period; b) the gradual relaxation of containment measures – post-lockdown period (Phase 2 and Phase 3) – reduced insomnia severity, leading to a relative recovery of pre-pandemic levels; and c) we tested age-related heterogeneity in sleep responses, with an expected higher increase in insomnia in younger and middle-age groups. Methods Analyses drew on a subsample (N = 883) of respondents to ITA.LI – Italian Lives, a recently established panel study on a probability sample of individuals aged 16 + living in Italy. To estimate patterns of change in insomnia, we first fitted a random-effects ordered logistic model on the whole sample. We then added an interaction term between policy phases and the respondent age to test whether the relationship between insomnia and policy phases differed across age groups. Analyses accounted for survey non-response weights. Results The fraction of respondents reporting moderate (“somewhat” + 0.159, S.E. 0.017) or severe (“very much” + 0.142, S.E. 0.030) sleep disturbances significantly increased during Phase 1. The prevalence of insomnia followed an inverted U-shaped curve across policy phases, with further increases from baseline levels (“somewhat” + 0.168, S.E. 0.015; “very much” + 0.187, S.E. 0.030) during Phase 2, followed by a relative reduction in Phase 3, although it remained significantly higher than in the pre-pandemic period (“somewhat”, + 0.084, S.E. 0.016; “very much”, + 0.045, S.E. 0.010). There were significant age-related differences in insomnia patterns, as the discrete change from pre-pandemic levels in the probability of not suffering from insomnia was negative and significant for the younger age group (− 0.269, S.E. 0.060) and for respondents aged 35–54 (− 0.163, S.E. 0.039). Conclusion There is reason to believe that the emergency policy response to the COVID-19 crisis may have had unintended and possibly scarring effects in terms of increased prevalence of insomnia. The hardest hit were young adults and, to a lesser extent, the middle-aged; however, older respondents (55 +) remained resilient, and their insomnia trajectory bounced back to pre-pandemic levels.