Modeling the effects of a policy change on supplier induced demand for sixteen hospital treatments

Introduction: Changes in the budgeting system of hospitals affects the admission rates. We estimated the effect of the 2001 policy change in the budgeting system of Dutch hospitals by focusing on the ambiguity of treatment. We used different statistical models to assess the effects of the policy change on 16 conditions. Methods: The Difference-in-Difference model was used to describe the 2001 policy effect on admissions by separating the total population into two groups, where we believed ambiguity of treatments was higher, and the control group. The DID model was transformed to fit into the panel nature of our date, and to fully capture the heterogeneity between the age groups. We also conducted a robustness analysis in order to validate our assumptions. Results: The average results of the simple DID model is 29, on the aggregated model is 35 and on the mixed effects model is 28.8. Bootstrapping had only a minimal effect on the SE. Conclusion: On the modeling side, this model can answer some questions as to how ambiguity of treatment affects the hospital admissions of specific age groups, under the light of a policy change. Based on our results in general, the difference between the two groups identified in the population on the hospital admissions, before and after 2001, increased by 30 in 10000 hospital admissions. Based on our assumptions and the positive and significant result that the policy change of 2001 affected more the age group where ambiguity of treatment is higher.