Necessity of Clinical Information in Surgical Pathology: A College of American Pathologists Q-Probes Study of 771 475 Surgical Pathology Cases From 341 Institutions

To our knowledge, there have been no studies examining the nature and extent of diagnostic and reporting problems occurring secondary to lack of information in surgical pathology. As Juan Rosai, MD, eloquently stated:

These principles are understood in surgical pathology and are satisfied by the majority of clinical physicians. They are codified in the laboratory accreditation standards outlined by the Laboratory Accreditation Program of the College of American Pathologists (question 08.1105 of the 1998 Laboratory Checklist) and the Joint Commission on Accreditation of Healthcare Organizations 1998 Standard QC.2.1.1.2,3 Both standards require that each surgically removed specimen is accompanied by pertinent clinical information and, to the degree known, by the preoperative and postoperative diagnosis.

In a previous Q-Probes study concerning surgical pathology specimen identification and accessioning,4 we noted that no clinical history or clinical diagnosis was present on requisition slips in 2.4% of cases. In the current study, we examine the effects of inadequate clinical information. Our focus, however, is on cases in which the pathologist felt that more clinical information was necessary to render an appropriate diagnosis, regardless of the information that was present on the requisition slip. Therefore, the deficiency rate differs from the previous Q-Probes study.4 Our aims are to (1) define the magnitude of the problem, (2) determine problematic organ systems or types of specimens where clinical information is essential for accurate diagnosis, (3) assess delay in diagnosis attributed to this lack of communication, (4) evaluate the methods of obtaining additional information, and (5) correlate these data with practice parameters for peer group comparison and benchmarking.

In 1996, 341 participants in the College of American Pathologists voluntary subscription quality improvement program, Q-Probes, prospectively documented all surgical pathology cases in which the requisition slip failed to provide adequate clinical data. The Q-Probes format for data collection and handling has been described in detail previously. Cases were considered to have inadequate clinical data on the requisition slip if the pathologist required additional clinical information before a diagnosis could be made, regardless of the amount of information already present. Cases that had no clinical information on a requisition slip were not counted if the lack of history did not prevent the rendering of a diagnosis and the sign-out of the case. The data collection period lasted for 3 months or until data were collected on 40 surgical pathology cases that required additional clinical information, whichever occurred first. The following data were recorded for each case: (1) anatomic site of specimen, (2) type of procedure used to obtain the specimen, (3) the nature of the disease, (4) method of obtaining additional information, (5) the importance of the clinical information obtained, and (6) the length of delay in the final diagnosis due to the need for additional clinical information. Participants selected answers to these various parameters from a predetermined menu of items. Participants also completed a questionnaire regarding their surgical pathology practices, including the surgical pathology volume accessioned by the laboratory in 1995, the number of cases accessioned during the study period, the number of full-time equivalent surgical pathologists, whether space was available on the requisition slip for clinical history, whether special requisitions were used for specific organs, whether specific policies existed requiring that clinical information be included on requisition forms, and how this policy was enforced.

Univariate nonparametric Wilcoxon and Kruskal-Wallis analyses were used to test for a difference in the distribution of the rate of inadequate clinical information with respect to multiple variables. A semiparametric multiple regression analysis was performed to determine which combinations of variables were significantly associated with the ranked values of the rate of inadequate clinical information. Fisher exact tests were performed to test whether proportions were significantly different for 2 groups, namely, the overall number of surgical pathology cases and the cases in which additional information led to substantial changes in diagnosis or a revised report. Statistical significance was assumed for P < .05 for all of these analyses.

Of the 341 institutions that submitted data for this study, 322 were from the United States, 7 from Canada, 5 from the United Kingdom, 2 from Australia, and 1 each from Hong Kong, Belgium, and New Zealand. Eighty-four percent reported that their laboratories were accredited by the College of American Pathologists, and 69% were accredited by the Joint Commission on Accreditation of Healthcare Organizations. The average occupied bedsizes of participating institutions were categorized as follows: 1–150 beds (34.8%), 151–300 beds (39.3%), 301–450 beds (14.8%), 451–600 beds (5.6%), and more than 600 beds (5.6%). Private nonprofit organizations accounted for 65.8% of the institutions; 7.5% were private, for profit; 4.0% were independent laboratories; 3.1% were university hospitals; 18.9% were governmental (county, 5.6%; city, 3.4%; state, 1.9%; veterans, 4.0%; armed forces, 1.2%); and 1.9% were from other nongovernmental organizations. City hospitals accounted for 61.5%, 21.7% were suburban, 15.2% rural, 1.2% federal, and 0.3% were other. A total of 34.1% considered themselves to be teaching hospitals, 65.9% did not, and 16.8% had a pathology residency program.

During the study period, participants accessioned and signed out a total of 771 475 surgical pathology cases. Of these, 5594 (0.73%) required additional clinical information before the case could be completed. The median institutional rate of cases with inadequate clinical information for diagnosis was 0.62%, the rate at the 10th percentile was 3.01%, and at the 90th percentile, 0.08%. The highest rate of cases with inadequate clinical information was 20%.

Three factors showed a significant inverse relationship with the rate of requisitions lacking adequate clinical information. Institutions with greater average occupied bedsize, a greater number of cases accessioned per year, and a greater number of pathologists in the practice all had lower percentages of cases lacking adequate clinical data (Table 1). Multiple practice characteristics were not associated with the rate of surgical pathology requisitions lacking adequate clinical information, including using special requisition slips for certain anatomic sites or organs, allowing sufficient space on requisition slips for clinical history, and having a policy requiring clinical information before sign-out.

A written departmental policy requiring documentation of clinical history on the requisition slip before sign-out of the case was in place in 60.7% of the institutions. Enforcement of this policy was accomplished by various mechanisms, including telephoning the appropriate nurse or physician for the missing information (16 institutions), specimens rejected and returned to the submitting physician for completion (23 institutions), documentation or report to a quality assurance committee or surgical case review committee (26 institutions), and other multiple responses in 42 institutions. The remaining 100 institutions did not respond to this question.

In the assessment of reporting delay associated with obtaining additional clinical information required for diagnosis, 32% of cases were judged to have been delayed; 15.1% of cases experienced a delay of more than 1 day. The vast majority (68%) had no delay, and 16.2% experienced a delay that was 1 day or less.

The most common method of obtaining additional clinical information was through direct communication with the physician (49.6%), followed by obtaining information through a computerized medical information system (11.9%), communication with nursing (10.5%), previous surgical pathology reports (8.0%), communication with other health care personnel (5.8%), chart review (5.3%), and other (8.9%).

In 59.4% of cases the additional clinical information confirmed the initial diagnostic impression, and in 25.1% of the cases it was not relevant to the pathologic diagnosis. In 4.2% of the cases the diagnosis was substantially changed because of the additional clinical information, and a revised report was issued in 1.9% of cases. Thus, 6.1% of cases that required additional clinical information for diagnosis were substantially changed or a revised report was issued once the adequate information was obtained. In 2.2% of cases no additional information could be obtained. In 7.2% of cases the response was “other”.

Two practice characteristics were found to be statistically associated with the impact of the additional clinical information on the diagnosis. Institutions with a greater number of surgical pathology cases accessioned in 1995 had a significant association with an increased percentage of cases in which the additional clinical information helped to confirm the initial diagnosis (P = .0104). A greater number of surgical pathology cases accessioned in 1995 (P = .0015) and a greater number of full-time equivalent pathologists (P = .0031) were significantly associated with a greater percentage of cases in which the obtained information led to a substantial change in diagnosis.

Anatomic sites significantly associated with a changed diagnosis or a revised report included the small bowel, the bronchus/lung, and the ovary (Table 2). Resections and segmental resections were also significantly associated with a higher rate of changed diagnoses or revised reports (Table 3), as were 2 disease conditions, malignant neoplasms and therapy-induced changes (Table 4). For cases in which the specimen was obtained through biopsies and curettings, significantly more cases with inadequate clinical information were found to result in a changed diagnosis or a revised report if a malignant neoplasm or therapy-induced change was present (Table 5). Cases in which the specimen was obtained through resection procedures were also significantly more likely to result in a changed diagnosis or revised report if a malignant neoplasm was involved (Table 6).

In his article “Limitations of Histologic Diagnosis,” Rambo6 states:

Published in 1962, these words faithfully echo the essence of this study.6 

In this Q-Probes study we examined the extent and severity of problems arising from inadequate clinical information, and to our knowledge, this is the first multi-institutional study of its kind. Using this multi-institutional design, we have established the magnitude of the problem (0.73% of cases), which may serve as a benchmark applicable for future comparisons in quality improvement programs. We have also demonstrated that smaller hospitals and smaller laboratories tend to have higher rates of cases with inadequate clinical information than larger hospitals and laboratories. This is important for peer group comparisons. The difference may be due to the more general nature of practice in a smaller hospital and to a lack of specific clinical specialists and pathologist “experts” in various organ systems who form close working relationships with small groups of clinicians. It may also be postulated that additional clinical information may be less of an issue in larger hospitals and in multipathologist groups staffed by subspecialized pathologists, who have enhanced diagnostic abilities and a more in-depth understanding of disease processes and clinical situations in specific clinical areas.7 This is best exemplified in settings in which the pathologist is an integral part of a clinical treatment team, where knowledge is shared on a continual basis. A multipathologist group is also more likely to have routine intradepartmental consultation between members.8 

Larger laboratories were also associated with a higher percentage of cases in which obtained clinical information helped confirm the initial diagnosis and led to a changed diagnosis. In other words, the additional information was more often useful. While the reasons for these findings are not clear, some of the characteristics of larger groups may also contribute to the pathologist's ability to find useful clinical information more often.

Another aim of the study was to identify anatomic sites for which additional clinical information was more likely to lead to a change in diagnosis. We identified 3 organ systems (small bowel, lung, and ovary) that were more often associated with a change in diagnosis or a revised report when additional clinical information was obtained. In addition, resection specimens were more likely to lead to a changed diagnosis or a revised report if additional information was obtained. Furthermore, cases involving malignant neoplasms and therapy-induced changes had significantly higher rates of changed diagnoses or revised reports when additional information was obtained. The reasons that these characteristics were associated with a revised report or changed diagnosis were not specifically evaluated in this study and are likely to be multifactorial. From our own experience, therapy-induced changes in biopsy tissue best exemplifies the pathologist's need for clinical information. In numerous conditions (eg, inflammatory bowel disease, prostate cancer, and allograft rejection), therapy has been shown to change the histologic appearance of the lesion and may lead to an improper diagnosis.9–11 

Recently, several articles have highlighted the role of inadequate clinical information contributing to diagnostic errors and amended reports in surgical pathology. In a College of American Pathologists Q-Probes study addressing specimen identification and accessioning, 2.4% of cases submitted to surgical pathology had no clinical history provided on the requisition slip at all.4 In a subsequent Q-Probes study, which quantitated amended surgical pathology reports, it was noted that 10% of the amended reports resulted directly from additional clinical information unknown to the pathologist at the time of original sign-out. An additional 20% of cases came to the pathologist's attention because of a clinicopathologic discrepancy recognized by the clinician.12 In our practice, when clinicians ask for a review of a case, they usually bring additional clinical information to the pathologist.

Inadequate clinical information has also been the underlying cause of malpractice claims brought against pathologists. In a review of 53 pathology malpractice claims, Troxel and Sabella13 documented that failure to obtain all relevant information (which often is not provided on the pathology request slip) contributed to one fifth of the diagnostic errors (11 cases). Similarly, McBroom and Ramsay14 noted that clinical information affected 7.5% of surgical pathology reports that were amended during review of cases for clinicopathologic conferences. They also pointed out that a significant proportion of cases that were amended were the result of the special expertise of the pathologists reviewing these cases. This expertise very likely reflects the expert's enhanced visual diagnostic ability and also reflects his or her knowledge of the clinical disease process occurring in the particular area of interest.7,8 In other words, a full understanding of the clinical picture optimizes the pathologist's ability to see his or her way to the most accurate diagnosis. As Goethe said, “you only see what you know.”