Abstract
Scientific research on the mental health effects of disasters has primarily focused on short-term consequences. This review aims to provide a long-term perspective, examining multiple contributing factors simultaneously. We registered this review (PROSPERO 2020, CRD42020108528), and searched Medline, PsycInfo, PTSDpubs, Web of Science, and SocINDEX from inception to July 2024. After screening 33,205 titles and abstracts, we extracted epidemiological longitudinal data from 71 studies, with up to six waves and a total of 137,004 participants. Disaster type (natural/human-made), category (e.g., earthquakes, floods, terrorist attacks), mental health outcome (e.g., posttraumatic stress disorder, depression, anxiety), population type (children/adolescents, adults), disaster year, measurement month, study quality, and country income were included in a multilevel meta-analysis. Pooled average prevalence of current or recent mental health problems was 22.1% (95% CI [10.95-39.57]). When controlling for the other factors, prevalence did not differ by disaster type, category, mental health outcome, population type, year, study quality, or country income. Despite high heterogeneity, the mental health burden in exposed populations decreased from month 1 to 300 after an initial peak in the first months, followed by a second peak after approximately a decade, before declining again ( p <0.001). Mental health burdens post-disaster may be more universally distributed than previously believed. Earlier studies identifying gradual post-disaster recovery may have underestimated the long-term effects. The analysis revealed a shortage of high-quality studies with measurements beyond four years. The topic requires further investigation with consistently repeated measurements at standard intervals to confirm the presence and early predictors of both larger and smaller peaks.