Johan Elfström

Reporting of Sentinel Events in Swedish Hospitals: A Comparison of Severe Adverse Events Reported by Patients and Providers

BACKGROUND Mandatory and voluntary reporting of adverse events is common in health care organizations but a more accurate understanding of the extent of patient injury may be obtained if additional sources are used. Patients in Sweden may file a claim for economic compensation from the national insurance system if they believe they have sustained an injury. The extent and pattern of reporting of serious adverse events in a mandatory national reporting system was compared with the reporting of adverse events on the basis of patient claims. METHODS Regional sentinel event reports were compared with malpractice claims data between 1996 and 2003. A sample consisting of 113 patients with deaths or serious injuries was selected from the malpractice claims data source. The medical records of these patients were reviewed by three chief medical officers. RESULTS Of the deaths or injuries associated with the 113 patients-25 deaths, 37 with more than 30% disability, and 51 with 16/o-30% disability-23 (20%) had been reported by chief medical officers to the National Board of Health and Welfare as sentinel events. Most adverse events were found in orthopedic surgery, and orthopedic injuries had more serious consequences. None of the patient injuries caused by infections were reported as sentinel events. Individual errors were more frequent in cases reported as sentinel events. CONCLUSIONS Adverse events causing severe harm are underreported to a great extent in Sweden despite the existence of a mandatory reporting system; physicians often consider them to be complications. Health care organizations should consider using a portfolio of tools-including incident reporting, medical record review, and analysis of patient claims-to gain a comprehensive picture of adverse events.

High Rate of Implementation of Proposed Actions for Improvement With the Healthcare Failure Mode Effect Analysis Method: Evaluation of 117 Analyses

Objectives The aims of this study were to investigate what kind of impact the Healthcare Failure Mode Effect Analysis (HFMEA) had on the organization in 1 county council in Sweden and to evaluate the method of working for multidisciplinary teams performing HFMEA. Three main outcome measures were used: the quality of the documentation from the HFMEAs, fulfillment of the primary goal of the HFMEA, and, finally, whether proposed actions for improvement were implemented. Methods The study involved retrospective analysis of the documentation from 117 performed HFMEAs from 3 hospitals in the county council of Östergötland, Sweden, and interviews or questionnaires with team leaders and managers between 2006 and 2010. Results A proposed change in the organizational structure was the most common issue in the analyses. Eighty-nine percent of the written reports were of high quality. A median of 10 serious risks were detected, and 10 proposed actions (median) were made. In 78% of the HFMEAs, all or a large part of these had been implemented a few years afterward. We were unable to find factors that promoted the rate of implementation of proposed actions. Seventy-eight percent of the managers were completely satisfied with the results of the HFMEA. The mean cost per risk analysis was &OV0556;1909. Conclusions Most of the proposed actions were implemented. The use of HFMEA can be improved using fewer team leaders but with more experience. The work involved in writing a report can be reduced without loss of impact on the organization.

Learning From Incident Reporting? Analysis of Incidents Resulting in Patient Injuries in a Web-based System in Swedish Health Care

Objectives Incident reporting (IR) systems have the potential to improve patient safety if they enable learning from the reported risks and incidents. The aim of this study was to investigate incidents registered in an IR system in a Swedish county council. Methods The study was conducted in the County Council of Östergötland, Sweden. Data were retrieved from the IR system, which included 4755 incidents occurring in somatic care that resulted in patient injuries from 2004 to 2012. One hundred correctly classified patient injuries were randomly sampled from 3 injury severity levels: injuries leading to deaths, permanent harm, and temporary harm. Three aspects were analyzed: handling of the incident, causes of the incident, and actions taken to prevent its recurrence. Results Of the 300 injuries, 79% were handled in the departments where they occurred. The department head decided what actions should be taken to prevent recurrence in response to 95% of the injuries. A total of 448 causes were identified for the injuries; problems associated with procedures, routines, and guidelines were most common. Decisions taken for 80% of the injuries could be classified using the IR system documentation and root cause analysis. The most commonly pursued type of action was change of work routine or guideline. Conclusions The handling, causes, and actions taken to prevent recurrence were similar for injuries of different severity levels. Various forms of feedback (information, education, and dialogue) were an integral aspect of the IR system. However, this feedback was primarily intradepartmental and did not yield much organizational learning.