Whether the use of online medical records (OMR) can contribute to personal medical decision-making has attracted scholars’ interest in recent years. Working on 1,547 respondents from the Health Information National Trends Survey (HINTS) data in the United States, this study estimates the causal effects of OMR on medical decision-making by triangulating the findings of several traditional econometrics and advanced machine learning methods. Our work reveals two interesting results, including causality and interpretability. Regarding causality, our findings confirm that the use of OMR has a positive causal impact on medical decisions, as the average treatment effects provided by all 11 estimation techniques range from 0.1 to 0.2 and are statistically significant at 1%. We also conduct several refute tests and verify the robustness of the sign and magnitude of our causal results. Regarding interpretability, the SHAP plots highlight top five confounders in this estimation: perceived usefulness of OMR, online healthcare provider appointments, online medicine/vitamin purchases, online healthcare tracking costs, and household income. The findings of this study imply the potential of employing several advanced machine learning methods in causal studies in healthcare management, especially in terms of transparency, interpretation, and error analysis.