Aspects of retrospective record review - A matter of patient safety

Sammanfattning: Background: Health Care is of great value but despite increased efforts to improve patient safety, many patients still suffer harm caused by healthcare, and even more patients have encountered incidents that could have caused harm. Adverse events can be detected by using retrospective record review. The Harvard Medical Practice Study and the Global Trigger Tool are such methods. Retrospective record review has shown better coverage than the commonly used clinical incident reporting system to identify patient safety information.Aims: The general aim of the thesis was to evaluate, compare and expand retrospective record review methods for clinical use in health care. The specific aims were;To evaluate the agreement in judgments of adverse events between well-trained Global Trigger Tool teams from different hospitals.To describe strengths and weaknesses, from team members perspectives of working with the Global Trigger Tool method of retrospective record review to identify adverse events causing patient harm.To evaluate the feasibility and capability of two common retrospective record review methods, the "Harvard Medical Practice Study" method and the "Global Trigger Tool" in detecting adverse events in adult orthopedic inpatients.To evaluate retrospective record review for the detection and characterization of no-harm incidents and compare findings with conventional incident reporting systems.Methods: A random sample of fifty patient records was reviewed by a team from each of five hospitals according to the Global Trigger Tool method (I). The teams were interviewed in focus groups concerning their experiences of the Global Trigger Tool method (II). In papers III and IV, a random sample of 350 orthopaedic admissions was examined. The outcomes from the Harvard Medical Practice Study and the Global Trigger Tool methods were compared (III). In paper IV the Harvard Medical Practice Study method was also used for identifying no harm incidents.Results: The number of identified adverse events differed between the teams, corresponding to a level of adverse events ranging from 27.2 to 99.7 per 1000 hospital days. Differences were also found in the assessment of level of harm and judgment of preventability. Four of the teams made similar assessments while the fifth identified three times as many adverse events compared to the other teams (I).Eight categories with their strengths and weaknesses emerged from the focus group interviews. The team members were constant in their generally positive experiences of the Global Trigger Tool method, even if the teams over time altered the application of the method (II).With the Harvard Medical Practice Study and the Global Trigger Tool methods combined, 160 adverse events were identified in 105 (30%) of the 350 records. The Harvard Medical Practice Study method identified 155 (97%) adverse events in 104 of 350 records compared with 137 (86%) adverse events in 98 records using the Global Trigger Tool method. The adverse events causing the greatest differences were the ones causing minimal or moderate impairment (III).In paper IV, results showed that 118 no harm incidents were detected in 91 (26%) of the 350 patient records. Ninety-four of the 118 (80%) no-harm incidents were classified as preventable. Sixteen no-harm incidents were identified by the five conventional incident reporting systems. Of these, ten no-harm incidents were also found by the Harvard Medical Practice Study method.Conclusions: Retrospective record review enabled detection of adverse events as well as identification of no harm incidents. It is considered a useful method. There were differences both in agreement between reviewer teams and between review methods. Joint preparations and discussions seemed to increase the level of agreement in judgment between reviewers. By adding retrospective record review for findings of no-harm incidents to conventional incident reporting, healthcare providers can gain new important information about commonly occurring, no-harm incidents in order to improve patient safety.