Logo Utrecht University

Dutch Mplus Users Group

Lars Tummers

No, we won’t! Analyzing the willingness of psychologists and psychiatrists to implement Diagnosis Related Groups using Bayesian statistics

In line with a current trend in numerous countries (Austria, China, the U.S., and Germany) the Dutch government introduced Diagnosis Related groups (DRGs) into mental healthcare. DRGs are introduced to improve transparency and control costs. Instead of implementing the DRG policy, psychologists and psychiatrists started forcefully resisting it, which could seriously endanger policy effectiveness and legitimacy. However, research concerning attitudes and behavior of psychologists and psychiatrists concerning DRGs is very scarce. This article analyzes the (un)willingness of psychologists and psychiatrists to implement DRGs, including its antecedents. A representative sample of 1,307 psychologists and psychiatrists is used. A double moderated mediation model is built using new features of the software Mplus v7. Additionally, we used Bayesian statistics instead of bootstrapping to deal with the non-normal distribution of indirect and interaction effects. In addition to this, we switched to Bayesian statistics because of several statistical advantages (among else Bayesian interpretation of the parameters, see also the other contributions in this special issue). Model development and syntax are included in the appendix, which can be beneficial for future research. The results showed that psychiatrists and psychologists were on average rather unwilling to implement DRGs. Psychiatrists being most resistant. The mechanisms explaining (un)willingness were also partly dependent on the professional group. More generally, the mechanisms explaining outcomes can be dependent on the profession, something not often analyzed in health services research.