Is clinical diagnosis of heart failure reliable?—Clinical judgement of cardiologists versus internists

Abstract

Background: Heart failure is a major cause of morbidity and mortality. Its diagnosis is mainly clinical. Most patients are seen by cardiologists and internists. However, it is not known whether clinical practices vary by specialty. We sought to evaluate the concordance in the clinical judgement of heart failure between cardiologists and internists. Methods: Clinical judgement was investigated in 17 cardiologists and 25 internists from four district hospitals using the probability assessment of heart failure diagnosis in 30 case histories based on real patients. Use of clinical information in the diagnostic assessment was defined by the regression coefficients of multiple regression analysis, in which the diagnostic probability was the dependent variable and the clinical criteria the independent variables. The importance attributed to clinical variables, as reported by doctors in a questionnaire, and that of clinical practice, as expressed by the magnitude of the regression coefficients, were compared. Results: We found no significant difference between cardiologists and internists. However, within each group of specialists, there was a wide inter-observer variation in the probability assessment of heart failure in the same case histories. The probability ranged from 25.6 to 83%. The relative importance of clinical variables actually used in diagnostic assessment was different from that reported by doctors. Conclusions: Cardiologists and internists do not differ in their clinical judgement of heart failure. However, within each group there can be wide discrepancies in the evaluation of the same case histories. This may be related to the different use of clinical information, as indicated by the wide confidence intervals of regression coefficients for clinical criteria. The way doctors use clinical information in practice differs from how they think they use it.