National mental health surveys play a critical role in determining the prevalence of mental disorders in a population and informing service planning. However, current surveys have important limitations, including the exclusion of key vulnerable groups and increasing rates of non-response. This review aims to synthesise information on excluded and undersampled groups in national mental health surveys. We conducted a targeted review of nationally representative adult mental health surveys performed between 2005 and 2019 in high-income OECD countries. Sixteen surveys met our inclusion criteria. The response rate for included surveys ranged between 36.3% and 80.0%. The most frequently excluded groups included people who were homeless, people in hospitals or health facilities and people in correctional facilities. Males and young people were the most commonly underrepresented groups among respondents. Attempts to collect data from non-responders and excluded populations were limited, but suggest that mental health status differs among some of these cohorts. The exclusion of key vulnerable groups and high rates of non-response have important implications for interpreting and using the results of national mental health surveys. Targeted supplementary surveys of excluded or hard-to-reach populations, more inclusive sampling methodologies, and strategies aimed at improving response rates should be considered to strengthen the accuracy and usefulness of survey findings.
