Who benefits, who is left behind? Intersectional inequities in unmet health care needs before and after the 2010 Swedish choice in primary health care reform using decision trees.

In 2010, Sweden implemented the Choice in Primary Health Care reform, allowing private healthcare providers to establish themselves at self-selected locations. While intended to improve efficiency and responsiveness through increased competition, concerns were also raised about its potential to reinforce inequities in healthcare access across different population groups. This study examines intersectional inequities in unmet healthcare needs (UHCN) in relation to the reform's implementation. Using data from the Health on Equal Terms survey (2007-2014, N = 69,644), we applied a decision tree-based analytical method (Model-Based Recursive Partitioning) combined with a post-hoc selection approach based on the Area Under the Receiver Operating Characteristic Curve to identify intersectional subgroups and assess reform-related effects. Although UHCN prevalence decreased by 11% after the reform, this overall improvement did not result in a narrowing of the equity gap across intersectional groups. Those born outside the Nordic countries, intersecting with low income, low education, and middle age, remained at the highest risk of experiencing UHCN both before (PR = 6.67, p < 0.001) and after (PR = 5.14, p < 0.001) the reform's implementation, indicating that, while absolute levels of UHCN decreased, the relative positioning of disadvantaged groups remained largely unchanged. These findings illustrate that overall improvements in healthcare access following market-oriented reforms do not necessarily address pre-existing structural inequities, and suggest that complementary equity-targeted measures will be required to ensure equitable access across diverse populations in Sweden. The decision tree-based approach proved valuable for evaluating health policies by uncovering heterogeneous effects across complex, intersectional patterns of healthcare outcomes.