Performance Indicators in Health Care

Programme

___________________________________________________________________________________ 

Abstracts

Jan Mainz: Introduction to conference themes

Andrew Hale:  Key performance indicators in safety critical industries

Key performance indicators need to be linked carefully to the safety management system at points where the behaviour of those judged on those indicators can have a significant influence on the safety and health outcomes. This presentation will discuss in general the nature of good indicators and present a case study and evaluation of an organisation in the steel industry which introduced a system of KPIs and achieved a significant improvement in safety performance.

Mary Dixon-Woods:  How hard is it to know whether care is safe?

This talk will explore the challenges in identifying whether organisations are providing safe, high quality care. The fugitive nature of much intelligence about safety, harm and risk will be explored, and the role of performance indicators in both revealing and suppressing intelligence will be characterised. The intended and unintended consequences of use of performance indicators to secure improvements in quality and safety of care will be explored.  It will be shown that measurement is a social as well as a technical practice, but ‘gaming’ does not provide a full account of why performance measures between different organisations often lack comparability. The talk will conclude by offering suggestions on how to optimise use of performance indicators in generating intelligence about whether care is safe.

Richard Lilford:  Evaluation of quality improvement: design and measurement issues

I shall discuss basic study designs for service delivery improvement, with an emphasis on the advantages of step-wedge cluster studies. Then I shall discuss measurement issues, with special emphasis on the distinction between outcomes, clinical processes, intervening variables, and system level processes. I shall draw a distinction between targeted management interventions and generic management interventions. I shall bring it all together under a Bayesian epistemological framework that I consider an advantage over the ‘realist’ idea.

Ulrik Gerdes:  HSMR is a poor instrument for measuring healthcare quality

HSMR figures for all Danish hospitals including aggregated results for each of the five regions and the whole country have been published quarterly since 2009. Danish Regions and the Danish Safer Hospital Programme use HSMR to set targets for quality improvement (a 10% and a 15% reduction, respectively) and for monitoring trends.  There are two basic problems with HSMR which combine to make the parameter a poor instrument to measure quality: The proportion of deaths in hospitals attributable to variation in quality of care is small and the validity of the data behind the HSMRs is questionable.

Mette Nørgaard: HSMR - strengths and weaknesses

Avoidance of unnecessary deaths matters to policy makers, hospital managers, clinicians, and patients. Use of summary hospital mortality indicators, such as HSMR, reflects this interest. HSMR has, however, several limitations. It cannot, as an example, distinguish between inevitable and preventable deaths and the method in itself makes comparison between hospitals inappropriate. Still, The HSMR may contribute as an important tool for surveillance purposes when combined with other types of information. Caution must, however, be used, to avoid drawing unsupported conclusions.

PhD project session

Christian Sørup; Evaluation of emergency department performance

The overall research questions of the project is: (i)  How can a holistic framework be developed, which, using selected indicators, can measure if a patient receives better quality and efficiency in care and treatment by implementing new initiatives, such as triage? The derived sub-questions are: (i) Which indicators can be used with benefit and what are their relations? (ii) How can a framework be developed, measuring quality and efficiency on the basis of knowing about a series of indicators and their respective relations? (iii) Can the developed framework be generalised?

Caroline Raben: Indicators for improving patient safety  

Based on a literature survey of proactive use of indicators in healthcare and other safety critical industries, the project will (i) develop a method for mapping specialized processes related to patient safety, and (ii) analyse these processes in order to identify candidate indicators.

Jan Kenneth Poulsen: Health economics models of cost - quality relationships

Paul Bartels: Perspectives and conclusion