Value of the unaided clinical diagnosis in dyspeptic patients in primary care.

OBJECTIVES: Attempts to establish a clinical diagnosis in dyspeptic patients have generally been unrewarding. However, studies in unselected dyspeptic patients are lacking. The aim of this study was to determine the value of the unaided clinical diagnosis by general practitioners (GP) and by experienced gastroenterologists (GA) in unselected dyspeptic patients in primary care.
METHODS: Three hundred forty-seven patients with epigastric pain/discomfort for more than 2 wk who were consulting general practitioners (n = 73), but without alarm symptoms. GPs and GAs gave a provisional diagnosis based on an unstructured interview. All patients underwent endoscopy within 5 days of referral. Validity of the provisional diagnoses was measured using the endoscopic diagnoses as the gold standards.
RESULTS: For GPs, the sensitivity of a provisional diagnosis of peptic ulcer was 61% [95% confidence intervals (CI): 46-74%]; for specificity 73%, the 95% CI was 68-78%; and for positive predictive values, it was 28%, the 95% CI was 20-37%. GAs were more reluctant to predict ulcer, leading to a higher specificity: 84% (95% CI: 79-88%), but a similar sensitivity: 55% (95% CI: 40-69%). The GPs were unable to distinguish between functional and organic dyspepsia (chance-corrected overall validity: 9%; 95% CI: 0-18%). GPs and GAs agreed in their provisional diagnosis in only 45% of the patients, in whom the diagnosis was confirmed by endoscopy in 2/3.
CONCLUSION: The unaided clinical diagnosis given by the GP and by the GA in dyspeptic patients in primary care is unreliable. Nearly half of patients with ulcer or esophagitis were misclassified, despite a high susceptibility to organic disease. Different patients were problematic for GPs and GAs, which may indicate that most dyspeptic patients do not present with symptoms characteristic of a specific disease.