Marginalized two part model for analyzing multilevel semicontinuous medical costs in Iranian households

Medical costs (MCs) represent a significant burden on household finances and often lead to economic challenges. This study analyzed the data of 8,993 Iranian households from 2021 collected by the Iranian Statistical Center. Using a marginalized two-part model (MTP) with lognormal and gamma distributions, the relationship between MCs and factors such as age, gender, education, and household characteristics was examined. A two-level structure was applied to account for heterogeneity across provinces, with analyzes performed using R software. The mean annual MC was $180 with high variability (SD = $324.39). The main determinants included family size, residence area, education level, and socioeconomic status. Single households and families with more students had lower MCs. Among the models evaluated, the MTP-Lognormal model (MTP-LN) performed better than the MTP-Gamma model (MTP-G), as it provided better predictive accuracy and better reflected cost differences between province. These results highlight the socio-economic and demographic factors that influence household MCs in Iran. The MTP-LN model provides valuable insights for identifying at-risk groups and developing targeted interventions to reduce the financial burden of healthcare, especially for vulnerable populations. This study emphasizes the importance of tailored interventions to address regional inequalities and promote equitable access to healthcare.