Choice for public hospital patients

Assessing choice for public hospital patients

MUCHE assessed Australians’ preferences for attributes associated with public hospital care for elective surgery. The study aimed to address three questions: do patients value hospital quality relative to convenience, what type of hospital quality matters the most, and how patient preferences differ under different surgical urgency levels and across patient characteristics.

A discrete choice experiment was conducted, asking 1,000 Australians to choose between two hypothetical hospitals for undertaking total hip replacement. Respondents were randomised into two versions: semi-urgent and non-urgent. Hospitals were described using distance, waiting time, GP’s opinion, other patients’ rating, health gain, rate of adverse events, and readmission rate. The first two measured “convenience” and the rest “quality”. The conditional logit and latent class logit were estimated and linearity of preference was formally tested. Willingness to travel and willingness to wait were computed to examine the trade-off between “quality” and “convenience”.

The experiment concluded that respondents are willing to trade-off between convenience and quality. Regarding relative importance, health gain was valued more than potential risks, and the GP’s opinion and other patients’ rating were valued similarly. Urgency only impacted preference for waiting time. Respondent preferences were segmented into three classes (all attributes were statistically significant at 5% in each class), from weak to strong, with gender, education level, household income, location, and past elective surgery experience predicting class membership. The test rejected linearity of preference under the conditional logit and did not reject it under the three-class logit.

This is the first study to examine the impact of urgency levels of elective surgery in this literature. It provides additional evidence for using health gain as a key quality indicator. It also demonstrates the interplay between preference heterogeneity and functional form of preference which may offer a potential solution to the problem of calculating welfare measures when the numeraire’s functional form is non-linear.