Learning from patient injury claims

University dissertation from Stockholm : Karolinska Institutet, Department of Learning, Informatics, Management and Ethics (Lime)

Abstract: Background: The Institute of Medicine report, To err is human, heightened attention to safety and quality performance in healthcare. This has led to demands on healthcare systems to collect data on safety and quality performance. Patient safety improvement requires learning at many levels in the system leading to changes in organizational structure and processes along many dimensions. Safety information systems support learning about the performance of a system by collecting, analyzing, and providing feedback of data. Other industries have come further than healthcare in measuring safety performance as well as in identifying industry specific knowledge about sources of vulnerabilities and hazards. In healthcare, evidence based measures are being developed such as incident reporting systems, medical chart reviews, patient safety indicators and malpractice claims data. The Swedish patient insurance claims database is a source of data on safety performance that has not yet been systematically studied. The aim of this thesis is to assess the potential contribution of patient injury claims have in supporting organizational learning in improving patient safety and to present a framework for the management of patient safety information in healthcare. Principal findings: Patient injury claims are, by themselves, not sufficient to serve as a sensor for vulnerabilities in healthcare. They do, however, provide a broad national source of patient generated information on negative outcomes of care which complements other healthcare generated reporting systems (Study I-II). Swedish healthcare leaders have a relatively high awareness of patient safety and give it high priority. However, few healthcare organizations actively involve patients in improving safety (Study III). Based on the assumption that analogies to known phenomena promote learning, the preservation of genomic integrity was presented as a model to describe different sources of variability, applicable also to patient safety (Study IV). Conclusions: Patient injury claims are less subject to bias than other sources of patient generated safety data (especially litigated malpractice claims), inexpensive, national, and allow for aggregation of data across many providers to identify rare complications. Analysis of the data can be done on both high level and granular levels of the system, which allows for organization specific feedback. From an organizational learning perspective, patient injury claims have both limitations and potential contributions. While there are limitations regarding timeliness, coverage and validity if they are used to provide an estimate of the rate of preventable adverse events, patient injury claims data contain useful information regarding adverse events and could act as a starting point for identifying areas in health care for further analysis in order to find vulnerabilities. Healthcare needs to develop comprehensive safety information systems that combine different sources of data to detect and learn from vulnerabilities.

  This dissertation MIGHT be available in PDF-format. Check this page to see if it is available for download.