Malpractice claims: finding the silver lining.
This article should be read in context of the Institute of Medicine report, To Err is Human (National Academies Press), published in 2000. This report concluded that 44,000 to 98,000 patients die each year from medical errors in US hospitals and that we must focus on improving systems of healthcare delivery (in contrast to the malpractice system's goal of blaming individual providers for mistakes). The report led to increased research funding from the federal government, new regulations by accrediting organizations, a focus on quality by purchasers of healthcare, and innumerable safety programs in hospitals throughout the country. For example, many hospitals are encouraging providers to report errors and adverse events in order to learn more about the types of errors that occur and why they occur. Malpractice claims files can be considered a type of patient-initiated error reporting system, and one that offers valuable insights for how to improve patient safety.
Before considering the implications of claims file analyses in general, and the Holohan study in particular, readers should be reminded of the limitations of malpractice claims file analyses. First, the errors documented in claims files are probably not representative of all errors. Recall that a patient had to hire a lawyer and file a lawsuit to initiate the process that results in a paid claim. Thus, claims can be considered a type of patient initiated reporting system. Patients often sue because there has been a bad outcome or because of conflict with a physician, and not solely because of poor quality of care. So there are probably many types of errors in medicine that are not represented in claims files. Second, there is a time lag of several years between the occurrence of the error, the time the claim was filed, and the time it was paid (which was the final event that made claims eligible for review in this sample). So results of claims file analyses may have limited application to current medical care. Third, the data in the claim file may be incomplete and the facts may be disputed. These and other factors can lead to poor reliability when reviewers are asked to make judgments about negligence. Fourth, reviewers are more likely to find errors in the process of care if they know that there was a bad outcome, as is the case with malpractice claims. This is called hindsight bias and may mean that studies of claims files result in overcalling errors.
A final limitation is that claims file analyses do not allow us to calculate the incidence rate of events. We do not know the appropriate denominator to use to calculate a rate. It is possible that when compared to other quality problems, the rates of diagnostic errors are so low that we should invest resources elsewhere first. But whatever measure one chooses, the absolute number of diagnostic errors and deaths reported here is too many and deserves our attention.
A strength of claims files is that they provide a unique view of the quality of care, allowing one to efficiently identify a large number of errors. One would have to review tens of thousands of randomly selected medical records to find the number of diagnostic errors reported in this study. Claims file analysis is probably more efficient in identifying serious errors than other reporting systems.
Claims files should be seen as one of many sources of data about errors and adverse events. For example, from record review studies or direct observations, institutions may know that diagnostic errors are common in certain settings or for certain diagnoses. If so, then the data from the claims file analysis can provide an enriched sample of errors with more detail about the systems factors that lead to diagnostic errors. The medical records, depositions, and other data in claims files allow the identification of system issues that contribute to the occurrence of errors and adverse events (although these are not reported by Holohan and colleagues). With this information, interventions can be designed to rectify these system errors and improve patient safety.
Malpractice claims are justifiably viewed as undesirable events. But this study should remind researchers and quality improvement leaders of how they can be used to help quality improvement efforts. Claims file analysis is a useful adjunct to other data sources about errors and adverse events and the high prevalence of diagnostic errors in this dataset should add momentum to efforts to understand their frequency and causes, especially in the outpatient setting.
Eric J. Thomas, MD
From the University of Texas Medical School at Houston, Houston, TX
Reprint requests to Eric J. Thomas, MD, University of Texas Medical School at Houston, 6431 Fannin MSB 1.122, Houston, TX 77030. Email: email@example.com
Accepted August 25, 2005.