Abstract
Rationale: Increasing numbers of law suit cases against medical practitioners signal the whole Thai society to pay more attention to the problem. This study focuses on studying the nature of medical reviews as tools for monitoring and assessing the problems of adverse events (AEs) occurred in hospitals. Objectives: The study aims at 1. Evaluating inter-rater reliability of medical record reviews in detecting AEs. 2. Evaluating changes of inter-rater reliability of medical record reviews when more financial incentives are proposed to the reviewers. 3. Evaluating usefulness of medical record reviews in terms of types of quality issues identified and applicability to the decision making on patient compensation. Methods: the study is divided into 2 phases. Phase I : 279 medical records were independently reviewed by 23 auditors. Each record was examined by 3 auditors using a standardized review form. Agreements on the occurrence of AEs were assessed using Kappa statistic (κ). Phase II : Lessons learned from the first phase were used in modifying the review form and processes of medical reviews. 50 medical records were independently reviews by 13 auditors. Early in the process, 7 of the auditors were informed that they will be given 50 – 100 % extra fees if the agreement of assessing the AEs among them was high. The rest 6 auditors were informed the same message later after they had finished reviewing the first 25 records. Agreements on the occurrence of AEs were assessed using Kappa statistic (κ). Statistical differences between kappa were tested using Mann Whitney U Test. 95 % confidence interval of kappa was also calculated. Content analysis of opened questions was done to categorize the reasons for making AEs diagnosis given by the auditors. Modeling from the Norwegian system of patient compensation, the author had tested whether the results of the medical reviews can be used for the compensation decision. Results: Phase I: Agreement on determining AEs occurrence was fair (weighted κ = 0.34, 95%CI= 0.22-0.45). Agreement on determining AEs preventability was fair (weighted κ = 0.27, 95%CI= 0.16 to 0.39). Phase II: For Agreement on determining AEs occurrence was fair (κ = 0.27, 95% CI:0.24 ถึง 0.30). Agreement on determining AEs preventability was also fair (κ = 0.22, 95% CI:0.19 ถึง 0.25). Informing auditors that they would be given extra fees had no positive effect on the agreement on determining AEs occurrence and preventability (p-value > 0.05). Reasons for determining AEs occurrence and preventability have revealed a numbers of common quality related problems, for example, missed diagnosis, delayed intubation, inappropriate antibiotic, inadequate treatment, overloaded fluid, overdosed Insulin, Propanolol in the bronchospastic, inadequate electrolyte monitoring, one-lung intubation etc. 42 % of the patients were considered to have rights for compensations according to the criteria used by the Norwegian patient compensation system. Conclusions and Discussions: The reliability of medical record review to detect AEs seems to be influenced by the rate of the event occurrence as well as the variability among reviewers. Whereas, financial incentive has no effect. The use of multiple reviewers to detect AE will produce unstable rates of AE occurrence, but is a reasonable and beneficial tools to search for wide ranges of quality related problems. Reasons for AE occurrence and preventability given by reviewers can serve well as a basis for making decision on patient compensation using the Norwegian criteria. The tool developed by this study is recommended to be modified and used as a tool for voluntary AE reporting and quality improvement process.
