Report Examines Worldwide Longitudinal Datasets To Improve Mental Health Research

A new report led by researchers at the Institute of Psychiatry, Psychology & Neuroscience (IoPPN) at King’s College London and commissioned by Wellcome explores the value of longitudinal datasets for mental health research.

Researchers identified 3,068 longitudinal datasets across 146 different countries in an extensive search to map longitudinal datasets across the world. The report, published today, contains an in-depth analysis of the identified datasets, their potential for mental health research and ways they can be enriched to improve and facilitate research. It can be used as a resource for researchers to identify longitudinal data to support transformative mental health research, particularly in depression, anxiety and psychosis.

Researchers examined the richness of data – the level of detail and complexity – to determine its value in mental health research. The report identifies several especially rich datasets for mental health research, but the team did not identify one perfect source. Instead, the report highlights that groups of datasets could complement each other and increase their value by being used together.

Working with our national and international collaborators, we identified and reviewed over 3000 longitudinal datasets from around the world. Our report highlights that there’s not one perfect source of data for studying mental health – instead researchers could analyse multiple datasets to increase research value. If researchers coordinate their approach to conducting longitudinal research and improve the discoverability of existing datasets, we can maximise on the huge financial, time, and resource investment that has been made in longitudinal research so far.

Professor Louise Arseneault, Professor of Developmental Psychology at King’s IoPPN and project leader
Researchers found that some of the richest datasets for mental health research included strong measures of depression and anxiety, mental health data across the life course, and biological and genetic data. However, they also uncovered limitations to available mental health data. Only 3% of the longitudinal datasets identified had large sample sizes (>8,000 participants at inception), detailed assessments and mental health data from participants aged 14 to 30 years. Of those, only one in 10 primarily focused on mental health.

All parts of the world contributed longitudinal datasets, with 25% originating from Africa, the Middle East and Asia. However, nearly 50% were based in Europe and the Pacific. Longitudinal datasets in low- and middle-income countries were under-represented and lacked resources (e.g., wearables, neuroimaging and the ability to collect data routinely).

The report identifies ways to enhance the value of these datasets by improving retention of minoritised and marginalised participants, expanding existing samples, improving measurement methods and collection, improving discoverability and encouraging the involvement of lived experience experts.

Rather than overburdening study participants by collecting too many data too frequently, researchers and funders can join forces to optimise the value of a single dataset by encouraging representative participant retention and recruitment, improving measurements and working with lived experience experts. This could allow rich datasets to improve our understanding of mental health without compromising participation rates and high-quality data.

Bridget Byran, PhD student at King’s IoPPN and research assistant on the project
Researchers have made the list of 3,068 identified longitudinal datasets (as of May 2023) openly available on their website to improve discoverability and facilitate future research.

The Landscaping Longitudinal Research project was commissioned by Wellcome and conducted in partnership with MQ Mental Health Research, the Open Data Institute and lived experience experts, alongside a range of national and international collaborators.