An Analysis of Morning Report: Implications for Internal Medicine Education

Methods

Morning Report and the Morning Report Database

Morning report at the University of California, Los Angeles Medical Center is a 60-minute conference held four times per week in which approximately 5% of admissions are selected for presentation. All second- and third-year residents working on ward rotations are required to attend; residents on intensive care unit rotations, consult services, and outpatient clinic rotations routinely are present. The chief residents lead the report and several faculty members, including the chief or associate chief of service, are present. Residents present cases of their choosing that were admitted to their inpatient service. Diagnostic dilemmas, difficulties in management, rare illnesses, and unusual presentations of common diseases are preferred cases, and presentation at admission or early in the hospital course is the rule. Case presentations include a complete history and physical and laboratory findings. For diagnostic cases, questions are asked of the presenter followed by a problem-oriented discussion of the case. A differential diagnosis is constructed, and participants arrive at an initial morning report diagnosis or a prioritized list of initial morning report diagnoses.

A database was introduced to catalog morning report cases from July 1989 through June 1990. Essential elements of each case were recorded by a chief resident onto a data collection form during the conference, and the information was later keypunched into a database [Paradox 3.5, Borland International, Scotts Valley, California]. Data included 1) the patient's name and hospital identification number; 2) historical data including chief complaint and the history of present illness, physical findings, and laboratory results; and 3) the morning report differential diagnosis with the initial diagnosis or a codified list of initial diagnoses. This initial diagnosis represented the full consideration of the case by all morning report participants and their differential diagnosis on the day that the case was presented. Interesting, classic, or unusual chest roentgenograms, electrocardiograms, laboratory values, or microbiologic smears were also recorded on the report form.

Cases were rereviewed daily at each subsequent morning report with addition to the data collection form of new physical examination, historical, and laboratory data until, in the judgment of both chief residents based on the clinical evaluation and course, the diagnostic process was completed. This diagnosis was considered to be the final diagnosis and was recorded on the report form; the case was then retired from rereview. The chief residents then completed the data collection form by assigning organ system categories for the final diagnosis or diagnoses according to the general section heading under which the disease is classified in Harrison's Principles of Internal Medicine [6] and classifying the case as primarily diagnostic or management.

If, in the judgment of both chief residents, a final diagnosis was not reached by the time the patient was discharged from the Medicine service, the diagnosis based on the full morning report consideration of the case was recorded on the morning report data collection form as the best morning report diagnosis, and the patient was followed for further elucidation of the diagnosis. Cases for which a final diagnosis was not reached before the patient was transferred to another inpatient service (for example, Surgery or Obstetrics/Gynecology) were followed onto the new service and the chart reviewed every other day until a final diagnosis was reached. Cases that were discharged from the hospital or died before reaching a final diagnosis had their charts (including inpatient laboratory reports, outpatient visits, and autopsy reports, when applicable) reviewed after discharge. Charts were reviewed every 4 weeks for 3 months and then 6 months after discharge. If, during case review, both chief residents agreed that an accurate diagnosis was reached, this was entered into the database as the final diagnosis, and the information was presented to the residents at morning report. Cases for which a final diagnosis was not delineated 6 months after hospital discharge or death were classified as diagnosis uncertain.

Database Analysis

For the 1989-90 academic year, frequencies of case type, organ system category, and whether cases required post-morning-report chart review were tabulated. No inpatient or outpatient clinic charts were lost to follow-up.

To evaluate the diagnostic accuracy of the morning report process, we separately analyzed cases in which a final diagnosis was reached during morning report and those cases that required post-morning-report chart review. For cases in which a final diagnosis was reached during morning report, we compared initial and final diagnoses for each case and decided whether these differed. For cases that required chart review because there was no final diagnosis when the patient left the Medicine service, the best morning report diagnosis was compared with the final diagnosis after chart review. Diagnoses were considered to differ if the final diagnosis substantially changed the educational value or the teaching points of the case when compared with the initial morning report diagnosis for cases completed while the patient was on the Medicine service or when compared with the best morning report diagnosis for cases requiring follow-up after discharge. If the final diagnosis was listed as a possibility in the initial morning report differential diagnosis (or in the best morning report diagnosis), even if it was considered unlikely, the final diagnosis was not considered to be different.

Initial (or best) morning report diagnoses were compared with final diagnoses by an independent physician reviewer for a random sample of cases. There was good agreement concerning whether initial (or best) morning report and final diagnoses were the same or different (Kappa = 0.9).

The proportion of cases in which the initial and final diagnoses were different was calculated for those cases where the final diagnosis was reached during morning report case rereview. The proportion of cases in which the best morning report diagnosis and final diagnosis differed was calculated for those cases requiring post-morning-report chart review. Rates of different initial (or best) and final diagnosis were compared by organ system category using chi-square tests.

Results

Of the 6640 patients admitted to the Medicine service during the study period, 330 were presented as morning report cases and analyzed; 294 (89%) were diagnostic cases, and 36 (11%) were management cases (Figure 1). Organ system classification of the final diagnoses (a case may have more than one) listed in descending order of frequency are: infectious disease (22%), hematology/oncology (21%), gastroenterology (15%), pulmonary (11%), cardiology (7%), rheumatology (7%), neurology (5%), nephrology (5%), endocrinology (4%), and pharmacology, dermatology, psychiatry, and allergy (1% or fewer). Ten percent of the patients had diagnoses related to the human immunodeficiency virus.

View larger version:

• In this window
• In a new window

Figure 1. Schematic of morning report cases on the Medicine service between July 1989 and June 1990.

Among the 294 diagnostic cases, 76% reached a final diagnosis during the morning report case presentation or during case rereview while the patient remained on the Medicine service, whereas 24% of cases required chart review because a final diagnosis was not reached. Among the 222 cases for which the final diagnosis was arrived at during morning report, the initial diagnosis differed from the final diagnosis for 38 (17%) (Table 1). Examples include a case given an initial diagnosis of aortic dissection that was later changed to pulmonary embolus and another in which the initial diagnosis was salmonella infection but the actual diagnosis was trypanosomiasis.

Among the 72 diagnostic cases that required follow-up due to lack of a final diagnosis during morning report rereview, chart review by 6 months after morning report for 18 (25%) cases confirmed the best morning report diagnosis; for 26 (36%) revealed a final diagnosis that differed from the best morning report diagnosis (Table 2); and for 28 (39%) the final diagnosis remained uncertain. An example of a final diagnosis that differed from the best morning report diagnosis is a case in which a presumed parasitic infection of unclear cause was later recognized to be the eosinophilia-myalgia syndrome. In another case, a presumed Nocardia skin infection was found to be due to an atypical mycobacterium. Among the cases that required follow-up, 14 died during the 6-month study period, and 6 (43%) received autopsies. Three of these resulted in a final diagnosis that differed from the best morning report diagnosis. In one case, a presumed colorectal malignancy causing gastrointestinal bleeding was found to be lung carcinoma metastatic to the colon. In the second, a bowel obstruction was found to be due to scleroderma, and in the third a presumed viral encephalitis was found to be anoxic encephalopathy caused by an unrecognized myocardial infarction.

Organ system categories did not differ significantly in the percentage of cases that reached a final diagnosis or for which the final and initial (or best) diagnoses differed.

Discussion

Morning report is an important part of academic internal medicine training programs and reflects the emphasis on inpatient medical education. The data presented from one institution's morning report, a selective case-based conference, suggest that the brief nature of the inpatient stay chronicled by the conference may short-change the educational process. One quarter of the diagnostic cases presented at morning report were deemed by the chief residents not to have had their diagnosis clarified by the time the patient left the inpatient Medicine service. For this group of cases, the differential diagnosis could be constructed, but inadequate feedback in terms of diagnostic tests; clarification of signs, symptoms, or historical data; or development of the natural history of disease existed for a diagnosis to be established. In that clinical reasoning tends to follow the path of iterative hypothesis testing [7], this lack of closure seriously hampers the learning process for this substantial proportion of cases. Although it is not to be expected that all cases will spontaneously unravel to reveal their cause, 61% of these unfinished cases had a diagnosis established at later follow-up. Most importantly, for 26 cases36% of those followed-up and 9% of all diagnostic casesa different diagnosis was found at follow-up than was presumed when the patient left the Medicine service.

Diagnostic cases for which a diagnosis is not reached during the educational experience represent lost educational opportunities. Reasoning that created a differential diagnosis and prescriptions for further testing may not be critically evaluated if the test results, the next set of clues, and perhaps an ultimate diagnosis are lacking. In cases where the initial diagnosis would eventually have been validated, the lack of feedback hampers the educational experience and may adversely affect resident satisfaction. Even worse, in those cases where the last known diagnosis is wrong (such as when presumed abdominal malignancy turns out to be a pancreatic abscess), lessons of diagnostic reasoning may be mislearned and misinformation incorporated in one's memory of diagnosis and disease. By contrast, Wheeler [8] has pointed out that a recognized mistake is probably the best teacher. Unexpected diagnostic lessons provide an opportunity to re-evaluate the diagnostic process, to reconsider one's perception of disease prevalence, or to review the meaning of a sign, symptom, or diagnostic test result. Presentation of follow-up data allowed residents to update their memory of previous case presentations and to reconsider the diagnostic process. Because the follow-up was retrospective, the patients themselves did not necessarily benefit from the case closure, but future patients might be expected to benefit from the lessons learned.

Residents and medical students should not come to believe that all cases have tidy solutions. It is important to recognize and accept uncertainty in the diagnostic process [9]. Yet, although it is often not feasible to keep a patient in the hospital until a diagnosis is reached or even for residents to follow the patient into the outpatient setting in many training programs, follow-up of diagnosis and outcome for most cases is possible. Most unfinished cases can be completed with a simple data-base that tabulates cases and provides for easy retrieval of cases with a final diagnosis still pending.

An added advantage afforded by systematized follow-up of undiagnosed cases is that the autopsyan underused resourcethat might have an important effect on delivering quality medical care [10, 11] can play a larger role. Autopsies in this study helped to establish unexpected diagnoses for 21% of unresolved patients in morning report who died even though the overall Medicine service autopsy rate was only 15% for the study period, a rate similar to that reported by McPhee and colleagues [12] but lower than that reported by Goldman and colleagues [13] and by Dans [11].

Analysis of the morning report database also highlights some of the deficiencies in the teaching of medicine as a series of snapshots of inpatient medical patients. The natural history of disease, methodic diagnostic evaluations, and watchful waiting rarely are found in the inpatient setting today, where almost all medical training still occurs. In morning report, despite in-depth case analysis, fully 24% (64 of 266 cases that reached a final diagnosis) of cases had a different final diagnosis than that first proposed. Misconceptions of disease presentation and appropriate diagnostic evaluation are likely to occur if cases are never revisited.

Because morning report cases are inpatient by design and, at our institution, reflect what the residents find to be interesting and difficult rather than a sample of all service admissions, the selection is skewed. Roughly twice as many infectious disease and hematology/oncology cases as gastroenterology, pulmonary, cardiology, and rheumatology cases were considered. A core curriculum was not achieved by our conference [1].

Further, only a small proportion of the Medicine service cases are presented. The approximately two cases covered per day is well below the mean of 5.7 found by Parrino [1]. Our data may not be generalizable to other morning report formats. Perhaps a more significant limitation is that the chief residents subjectively evaluated whether a final diagnosis was reached; in some cases, the diagnostic process may still have been incomplete. However, if more cases required follow-up after discharge from the Medicine service, the finding of missed educational opportunities reported here would only be accentuated.

This analysis reveals that continuity is often lacking and that educationally valuable material is lost from the inpatient Medicine service. In this way, morning report may reflect the general shortcomings of the inpatient internal medicine educational process. Changes that restore continuity between the inpatient and outpatient settings may enhance medical education. Further evaluation of the effects of improved continuity on learning the process of diagnosis is needed.