Description

Systematic provider comparisons are part of the national health care quality strategy in many countries. Providers may be compared amongst each other or to pre-specified benchmarks. Results may be used for public reporting, quality improvement or financial or regulatory incentives or sanctions. As the focus on quality measurement is growing, the need to perform health care provider comparisons in a reliable, adequate, and robust manner is increasing. The topic of provider comparison features a number of characteristic statistical problems. First, caseloads may vary widely between providers, and thus statistical uncertainty of estimated effects also varies widely. Both Bayesian and frequentist approaches play an important role to regularize results. However, shrinkage leads to the risk of underdiagnosing poor performers, especially among smaller providers. Second, medical treatments are processes in time. Patients have more than one treatment, and competing outcomes and censoring must be taken into account. Finally, measuring treatment outcomes may be challenging and costly. The wish for rich datasets for optimal risk adjustment must be balanced against the documentation burden. Some outcomes may best be reported by patients (patient reported outcome measures, PROMs), and so the number of initiatives that gather PROMs is growing. To compare providers based on PROMs, answers from questionnaires must be suitably aggregated and normalized. Limitations of the data such as unknown hierarchical structures or insufficient information on relevant patient characteristics are sources for bias that need to be understood and addressed.

The workshop extends the contributed session TC5: “Statistical issues in health care provider comparisons” in order to stimulate the exchange between statisticians working in this area. We will give an overview over different initiatives, discuss current open topics and plan further activities to advance the field and strengthen the community.

For further information, please contact: providercomparisons2026@iqtig.org