Statistische Methoden der Risikoadjustierung im Gesundheitswesen

Risikoadjustierung ist ein wichtiger Bestandteil bei der statistischen Analyse von medizinischen Behandlungsdaten. Das IQTIG veranstaltet zu diesem Thema am 18. März 2021 einen internationen Online-Workshop.

Insbesondere bei Bewertung und Vergleich von Leistungserbringern ist eine funktionierende Risikoadjustierung eine zentrale Grundvoraussetzung. Mit der internationalen Veranstaltung will das IQTIG renommierte Expertinnen und Experten in einem biostatistischen Workshop zur Risikoadjustierung zusammenbringen und aktuelle Entwicklungen diskutieren.

Statistical Methods for Risk Adjustment in Health Care

Effective risk adjustment is an essential prerequisite when evaluating and comparing health care providers. For this event, the IQTIG invites internationally renowned experts in statistical methods for Risk Adjustment to present and discuss recent developments in the field.

Sprecherinnen und Sprecher / Speakers

• Christian Günster (WIdO, AOK Research Institute, Germany)
  Titel / Title: Risk Adjustment Using Administrative Data: Experiences from 10 years of public reporting in Germany
  Zusammenfassung / Abstract: Risk adjustment of quality indicators using administrative data is challenging due to the inherent limitations of the data. On the other hand, administrative data provide access to information on preconditions and therapies prior to the treatment of interest. In recent decades, risk adjustment instruments developed for administrative data sets have increasingly been used. In the QSR (Qualitätssicherung mit Routinedaten) programme, further risk models have been developed by clinical experts for the purpose of public reporting for 10 surgical inpatient treatments in Germany.
• Prof. Els Goetghebeur (Ghent University, Belgium)
  Titel / Title: On the transportability of care center effects and the use of summary statistics for standardizing quality of care indicators
• Prof. Stijn Vansteelandt (Ghent University Belgium, and London School of Hygiene and Tropical Medicine, UK)
  Titel / Title: An overview of recent developments on risk adjustment from the causal inference literature
  Zusammenfassung / Abstract: Risk Adjustment forms a central theme within the causal inference literature, on which enormous progress has been made over the past decade. In this talk, I will give an overview of some of the recent developments that I believe to be most impactful, with particular relevance for assessing hospital performance. In particular, I will discuss problems of shrinkage bias that may result when attempting to stabilise estimates in response to low numbers of events in some hospitals, and how to temper this bias. I will moreover provide insight into the biased inferences that may arise as a result of model building, and how to accommodate this.
• Dr. Johannes Rauh (Institut für Qualitätssicherung und Transparenz im Gesundheitswesen, Germany)
  Titel / Title: Risikoadjustierung im IQTIG: Status quo und offene Fragen / Risk Adjustment at the IQTIG: Status quo and open problems
  Zusammenfassung / Abstract: The IQTIG manages an increasing number of currently around 60 risk adjustment models that are used in profiling of health care providers. Across the different medical areas, the complexity of the models (as expressed by the number of degrees of freedom) varies widely, as does the yearly caseload. Developing and updating these risk adjustment models and ensuring their statistical validity is a major task of the IQTIG. While most models are used to risk adjust quality indicators of SMR type, with rising numbers of quality indicators there is a growing need to construct valid risk adjustment models for aggregate measures (indices). This talk is about our methodological efforts to improve model validity and reliability for existing and new models and to increase transparency about our methods and our models.
  Prof. Arlene Ash (University of Massachusetts Medical School, USA)
• Titel / Title: The Push for Equity: Incorporating social risk factors into models to predict costs and other outcomes
  Zusammenfassung / Abstract: Topics to be covered: Ideas for identifying “social risk“ at the individual and ecological (neighborhood) levels. Issues to consider when developing a neighborhood-level measure of social risk (supervised vs. unsupervised learning). How can risk modelling be used to inform collaborators about the consequences (for vulnerable populations and those responsible for their health) of adjusting for social risk. The difficulty of comparing “inequity“ across populations served by different plans. The value of risk modelling (not just risk adjustment) for identifying and evaluating policies that can improve the health of vulnerable subgroups.

Für Rückfragen steht Ihnen Stefan Gehrig unter riskadjustment@iqtig.org gerne zur Verfügung.

For further questions, please contact Stefan Gehrig at riskadjustment@iqtig.org.