Variation in Psychiatric Hospitalisations: A Multiple-Membership Multiple-Classification Analysis.

The complexity of variation in healthcare, particularly in mental health, remains poorly understood. However, addressing this issue presents an opportunity to opti-mise the allocation of scarce healthcare resources. To explore this, we investigated the variation in psychiatric care measured as the number of psychiatric hospitalisations. We estimated multiple-membership multiple-classification models utilising Danish register data for 64,694 individuals and their healthcare providers, including 2101 general practitioners, 146 community-based care institutions, 46 hospital departments, and 98 municipalities. This approach recognised that data are not strictly hierarchical. We found that, among individuals attending a single healthcare provider, 67.4% of the total variance in the number of hospitalisations corresponds to differences between individuals, 22.6% to differences between healthcare providers' geographical location, 7.02% to differences between healthcare providers, and 3% to differences between the geographical locations of the individuals. Adding characteristics to the model ex-plained 68.5% of the variance at the healthcare provider geographical level, but almost no explanation of the variation was found on the three other levels despite the nu-merous characteristics considered. This suggests that medical practice may vary un-warrantedly between healthcare providers, indicating potential for optimisation. Streamlining medical practices, such as adhering to clinical guidelines, could lead to more efficient supply of mental health resources. In conclusion, understanding and addressing variation in psychiatric care may impact resource allocation and patient outcomes, ultimately leading to a more effective healthcare system.