Objectives:To understand the extent to which various demographic and social determinants predict mental wellbeing status and their relative hierarchy of predictive power in order to prioritize and develop population-based preventative approaches.Design:Cross-sectional analysis of survey data.Setting:Internet based survey from 32 countries across North America, Europe, Latin America, Middle East and North Africa, Sub Saharan Africa, South Asia and Australia.Participants:270,000 adults aged 18-85+ who participated in the Mental Health Million project.Primary and secondary outcome measures:We utilized 120+ demographic and social determinants to predict the aggregate mental health score of individuals (MHQ) and determine their relative predictive influence using various types of machine learning models including random forest, gradient boosting and logistic regression. The MHQ is derived from self-ratings of 47 mental health elements spanning ten disorders and provides a score that positions individuals along a spectrum from negative to positive mental health status that aligns with life impact and function criterion.Results:Classification models correctly identified 80% of those with a negative MHQ, while regression models predicted the specific MHQ score within ±15% of the position on the scale. Factors with the biggest predictive impact were young age followed by frequency of social interaction with friends, frequency of good sleep and physical exercise, and number of traumatic experiences. Age had twice the predictive power of social interaction which, in turn, was twice as important as the next four most important factors. Other predictive factors included sexual abuse, cyberbullying, and use of sleeping pills and sedatives.Conclusion:Social determinants of traumas and adversities and lifestyle can account for 60-70% of mental health challenges. However, additional factors are at play, particularly in younger age groups that are not included in this data and need further investigation.