Utility of Clinical Examination in the Diagnosis of Emergency Department Patients Admitted to the Department of Medicine of an Academic Hospital

The claim that high-quality history and physical examination are diagnostic for most patients is based on old studies and ambulatory patients.^1,2 We examined in a prospective study the utility of basic clinical information available on admission for the diagnosis of adult patients admitted to the department of medicine.

All consecutive patients newly admitted from the emergency department (ED) to 1 academic department of medicine over 53 days were prospectively included in the study. Planned admissions or readmissions were excluded. A senior resident (SR) with 4 years' training (L.P.) examined all patients within 24 hours of admission, including a full history taking, physical examination, and review of ancillary test findings done at the ED (basic hematology and chemistry tests, urinalysis, electrocardiography [ECG], and chest radiography [CXR]). Additional tests (troponin, C-reactive protein, and international normalized ratio) and computed tomography or ultrasonography (when preformed), medical charts from previous admissions, and all medications and vital signs were also reviewed. The SR then determined her main diagnosis at the highest degree of resolution possible (eg, syncope due to orthostatic hypotension) and indentified the modalities that were most helpful in making the diagnosis (eg, history + ECG). Once determined, the diagnosis was sealed and unknown to others. A hospital physician (HP) then repeated the same procedure, with no other data, and did not communicate the results. Participating HPs were active hospitalists and medical educators with 20 years' experience or more and were not involved in the care of the study patients. At least 1 month after discharge, the SR verified the patient's final diagnosis for the index admission by checking discharge summaries and records of any further hospital visits and called the patient's primary physician. These data were collected by the SR without being aware of her own or the HP diagnosis. The main diagnosis of the SR or HP compared with the final diagnosis constituted the primary outcome. Secondary outcomes included the value of different elements (history, physical examination, and basic tests) for the diagnosis and the prevalence and impact of imaging studies performed in the ED (other than CXR) on diagnosis. An experienced statistician analyzed the results.

Altogether, 442 eligible admitted patients (mean [SD] age, 66.9 [17.7] years; 51.4% male) were evaluated. Previous comorbidities were common (mean [SD] number, 1.8 [1.4]; 100 of 442 had none) including hypertension (57%), diabetes mellitus (34%), coronary disease (26%), stroke (16%), renal dysfunction (estimated glomerular filtration rate <60) (35%), and chronic lung disease (14%). Mean (SD) length of hospital stay was 4.5 (7.2) days. All patients had basic blood and urine tests performed in the ED, but only 15.5% had an ancillary test other than ECG and CXR (computed tomography, 11.8%; ultrasonography, 3.7%). The SR examined all patients within mean 14 hours of admission, spending approximately 40 minutes per patient (HP, ≤25 minutes). Follow-up and final diagnoses were obtained at a mean (SD) 2.0 (0.7) months after discharge and included a wide, diverse spectrum of illnesses typical of a department of general internal medicine. The SR was correct in 354 of 442 diagnoses (80.1%). The HPs made correct diagnoses in 373 patients (84.4%). They made identical correct diagnoses in 327 cases (73.9%); both were wrong in 42 patients (9.5%) (P = .04). The modalities considered to have been most useful in establishing the diagnosis were similar for both (Table). The patient's history emerged as the key element in formulating diagnosis either alone (approximately 20% of all diagnoses), in combination with the patient's examination (another 40%, approximately), or in addition to the basic tests with or without the physical examination (33%). The examination or basic tests alone were very seldom helpful. Used in conjunction, the physical examination doubled the diagnostic power of the history (19.5% to 39.0%; Table). The basic tests added a further 33%. Imaging was infrequently used in the ED (mainly head computed tomography) and had added little to determining diagnoses, being considered valuable in approximately 1 in 3 patients who had computed tomography performed.

Research continues to support the enduring value of the history and physical examination in diagnosis^3-5 and in deciphering problems with multiple diagnostic alternatives.^6-8 However, our study was the first to our knowledge to examine prospectively the value of the basic clinical methods for the diagnosis of the whole heterogeneous population of patients requiring an emergency admission to a general department of medicine.

We found that more than 80% of newly admitted internal medicine patients could be correctly diagnosed on admission and that basic clinical skills remain a powerful tool, sufficient for achieving an accurate diagnosis in most cases. Notwithstanding the great clinical diversity, 90% of all correct diagnoses were accomplished on presentation through a combination of the history, physical examination, and basic tests (excluding imaging studies) (Table). The use of this combination correctly diagnosed 3 of 4 admitted patients (329 [SR] and 347 [HP] of 442 consecutive admissions were properly diagnosed by the history, physical examination, and basic tests alone, or in combination [Table]). History was the most potent single tool identified. Its combination with physical examination of the patient alone was diagnostic in 60% of all admissions. Integrating the results of basic laboratory tests further increased the diagnostic yield (Table). The fact that a relatively small number of patients had ancillary investigations beyond ECG and CXR had no adverse effects on clinicians' performance. In contrast, the incorporation of the basic test results in the diagnostic considerations was found to be crucial because they were implicated in a third of all diagnoses. The SR performed nearly as well as the HPs, possibly owing to her 4 years' experience in the same department. One in 10 patients was misdiagnosed by both the SR and HPs, but diagnosis was achieved either during admission (for the majority of the patients) or after discharge and no patient harm occurred. Our results do not mean that sophisticated studies need not be used after admission, but they do suggest that their choice should be guided by the clinical data on presentation.^5,9 In conclusion, 4 of 5 of internal medicine inpatients could be accurately diagnosed close to their admission on the basis of little other than the traditional clinical information. Physicians may count more on their clinical faculties when making decisions about patients.

Correspondence: Dr Schattner, Department of Medicine A, Kaplan Medical Center, Rehovot and the Hebrew University Hadassah Medical School, POB 1, Rehovot 76100, Israel (amiMD@clalit.org.il).

Author Contributions:Study concept and design: Schattner. Acquisition of data: Paley, Zornitzki, Cohen, Friedman, Kozak, and Schattner. Analysis and interpretation of data: Paley, Zornitzki, and Schattner. Drafting of the manuscript: Paley and Schattner. Critical revision of the manuscript for important intellectual content: Zornitzki, Cohen, Friedman, Kozak, and Schattner. Statistical analysis: Schattner. Study supervision: Zornitzki, Cohen, Kozak, and Schattner.