New quality indicators introduced for hospitals

New quality indicators for hospitals and the software "Qlize!" were developed from several research projects at the University of Lucerne funded by the Swiss Agency for Innovation Promotion (Innosuisse). The development was led by Dr. Dr. Michael Havranek, Research Director of the Competence Center for Health Data Science, in cooperation with industry partner INMED and various hospital partners. The indicators have now become part of the mandatory measurement plan of the National Association for Quality Development in Hospitals and Clinics (ANQ). ANQ licenses and finances the software system and makes it available to all Swiss hospitals and clinics as well as to the cantonal health departments. The final part of the program's "rollout" took place on February 17.

Initial focus on unplanned re-entries

In an initial phase, thirteen quality indicators on unplanned readmissions after hospitalization will be used to assess the quality of treatment provided by hospitals. However, the software includes more than 30 additional quality indicators on complications and deaths during hospitalization, which can also be used in the future. All indicators were developed using internationally established principles, which were further developed based on the conditions of the Swiss healthcare system and tested together with seven hospital partners.

The special feature of the software developed is that it enables hospitals to analyze their own treatment quality down to the smallest detail. On the one hand, the hospitals can check their own quality results in a statistical comparison with other hospitals. On the other hand, they can also narrow down their results across different patient groups down to the individual case and relate them to expected rates from prediction models. The prediction models used for this purpose were calculated on the basis of all Swiss hospitalizations and use methods from the field of artificial intelligence.

Data protection compliant approach

In order to be able to offer such detailed evaluation options at all, despite the strict data protection requirements, it was necessary to develop a two-stage procedure. First, the data supplied by the Federal Statistical Office (FSO) for the whole of Switzerland is evaluated annually in order to provide the hospitals with their official quality results in statistical comparison with the other hospitals. Subsequently, the calculated prediction models are applied to the data provided by the hospitals themselves to enable them to make case-based comparisons with the expected rates.

In this way, it is possible for hospitals to analyze targeted groups of patients or even individual cases. For example, they can identify cases in which a low probability of a quality-relevant event (such as an unplanned readmission) was predicted, but such an event nevertheless occurred (e.g., due to a surgical complication). In order to communicate the diverse evaluation options in this regard, Michael Havranek had held two nationwide training events with simultaneous translation in French and Italian at the beginning of February on behalf of the ANQ, which were attended by around 150 hospitals.

Source: University of Lucerne