Objective: To determine and analyze the degree of agreement and disagreement in the diagnosis of bronchial asthma (BA) by respiratory disease specialists and generalists in regional hospital and primary care settings.
Material and methods: Ninety-six outpatients (16 to 70 years of age) were studied; all had been assigned a diagnosis of BA by the referring physician or by the respiratory disease specialist. We recorded 1) clinical symptoms, determining the initial probability of a diagnosis (IPD)of BA to be high, medium or low; 2) results of spirometry and bronchodilator testing (BDT), peak flow variability and methacholine challenge testing; 3) prick test results, eosinophil levels and total serum IgE levels. Three diagnoses were recorded: the initial diagnosis (ID) by the referring physician to whom follow-up data were unavailable; diagnosis by the respiratory disease specialist based only on clinical symptoms (RSS); and the final diagnosis(FD). To arrive at a FD of BA, it was necessary to have a high or medium IPD and a positive BDT. A Kappa test was used to analyze the degree of agreement among the three diagnoses. Group features associated with greater or lesser agreement were analyzed by chi-square tests and analysis of variance.
Results: Agreement was acceptable between RSS and FD (K = 0.63) but very low between ID and RSS and between ID and FD. In the latter two cases, agreement was greatest for patients diagnosed in hospital and for those with high IgE levels (p < 0.05), with high IPD, longer course of disease and a history of asthma (p < 0.01) (odds ratio =59.8). Diagnostic disagreement occurred mainly for patients for whom a BA diagnosis was gained later, the of under-diagnosis being 39%. The patients involved visited the physician only because they had observed an isolated symptom related to asthma (odds ratio = 119) and to arrive at a diagnosis bronchomotor tests other than BDT were required (p < 0.01).
Conclusions: a) The degree of agreement for a diagnosis of BA is low. b)The functional profile of patients for whom diagnostic agreement exists differs from that of patients for whom diagnosis is gained through testing. c) In the context of this study, a high rate of under-diagnosis is evident.