Content area
Full text
Medical records are the benchmark for assessing competence and determining what clinicians do in the course of patient visits. 1- 3 Despite their prominent place in quality measurement, chart abstraction is subject to important limitations, including the expense of abstraction and, for paper formats, illegibility and record unavailability. 4, 5 Perhaps the most important limitation of medical records as a measure of clinical performance is that physicians do not document everything they do. This recording bias contributes to a high false negative rate, meaning that chart abstraction may underestimate the actual quality of care. 6, 7
This observation of recording bias led us to ask if the medical record might also overestimate quality due to false positive reporting by clinicians. We hypothesised that the medical record not only lacks sensitivity (due to false negatives) but also specificity (due to false positives). If present, false positives would certainly raise additional questions about the reliability of the record as a quality measure and the integrity of physician documentation. False positives also raise substantive ethical questions, including the possibility of intentional deception and fraud. 8
Though a growing body of literature recognises the problem of recording bias and other causes of underreporting, few investigators have addressed the potential problem of erroneous inclusions in the medical record. This is primarily due to the methodological challenge inherent in measuring such errors. The use of the standardised patient (SP) encounter overcomes this obstacle, however, because SPs are a gold standard measure against which to measure not only false negatives but also false positives in the medical record. 9- 13 Thus, to determine if false positives exist in the medical record, we report on a study that compares the quality of care documented by physician subjects with the quality of care reported by actor patients (case-mix controlled). We then consider the ethical concerns that emerge from such an evaluation.
METHODS
Data was collected in the general medicine clinics of two VA medical centres between December 1996 and August 1997 using methods described elsewhere. 4 All second and third year residents and attending physicians in these clinics were eligible to participate; of these, 97% consented to participate. From consenting physicians, we randomly selected 20 participants, ten at each site.
Quality of care provided...