Cost and Quality Differences in Elective Total Hip Arthroplasty between Swiss Hospital Departments

Objective: Even in highly regulated and advanced health care markets, treatment costs and outcome quality substantially vary across facilities. Interhospital variation — particularly when it is not attributable to different patient compositions — might indicate inefficiencies in the health care system. This paper studies the sources and extent of these variations using novel data on total hip arthroplasty (THA) in Switzerland.
Methodology: The study sample comprises 20,918 patients from 110 hospital departments who underwent THA in 2016, 2017, and 2018. Two primary dependent variables are considered: quality of the outcome is measured by a binary variable that indicates a revision within two years of the surgery, and THA case costs represent the total inpatient costs of the initial THA as reported by the hospital. Interdepartmental cost and quality differences are explored using hierarchical two-level models. The study considers patient health factors and departmental factors, namely THA case volume, hospital type, and surgical techniques, as explanatory variables.
Results: Even after controlling for the patient mix, there is a significant variance in both costs and quality between hospital departments. Patient health factors account for 35% and 60% of the variance in interdepartmental differences for THA cost and quality, respectively. Several patient health factors, especially BMI and certain comorbidities, are important drivers of these differences in costs and quality. Additionally, both patient health and departmental factors together explain 50% and 80% of the significant interdepartmental variance in costs and quality, respectively. Low case volume, the department's proportion of cemented implants, and department affiliation with university hospitals were significantly associated with higher inpatient costs.
Discussion: A relatively high share of interdepartmental variance in both quality and cost can be explained by different departmental patient mixes. The residual unexplained variance, albeit significant, is modest in international comparison. This suggests that there are no substantial unexplainable inefficiencies among hospital departments in Switzerland. The results of departmental level suggest that THA costs could be reduced by economies of scale and increasing the share of uncemented prostheses without compromising the quality of care. The high reduction of interdepartmental variance by controlling for patient mix and the fact that most but not all patient health risk factors are linked to both higher costs and lower quality underlines the importance of carefully-designed patient health risk adjustments in potential future pay for performance (P4P) programs in THA.