A simple approach to fairer hospital benchmarking using patient experience data

Abstract

Objective: To determine the factors contributing most to variability in patient experience in order to present approaches for fairer benchmarking of hospitals and for quality improvement.

Design: Secondary analysis of data from a widely used survey on patient experience.

Setting: Inpatients from all 24 acute hospitals in the Canton of Bern in Switzerland. Data collection followed the standardized and validated Picker Institute methodology for a period of 13 weeks in the fall/winter of 2005.

Participants: Inpatients age 18 years and older (n = 14 089), discharged within the sampling period.

Main outcome measures: 'Patient experience', measured by the total Picker Problem Score (PPS) and by six domain scores (care, communication, respect, cooperation, organization, discharge management).

Results: In regression analysis, the patient factors self-reported health, age and education explained the highest proportion of variability in the PPS (4.8, 2.2 and 0.7%, respectively). Multiple linear adjustment for factors associated with patients removed between 29 and 33% of variability between hospital categories. The domain score means varied from under 5% for 'respect towards the patient' to 34% for 'discharge management'. Ranking of hospitals by domain scores differed from the ranking based on the total PPS.

Conclusions: Statistical adjustment for patient mix and additional stratification for some hospital factors make benchmarking using patient survey data fairer and more transparent. Use of our approach for presenting quality data may make interpretation easier for the different target groups and may enhance the relevance of such information for decision-making.