We present a review of specific health status measures, including symptoms, physical examination, and laboratory tests (exclusive of lung function tests), in terms of their suitability for assessing the presence and severity of asthma in epidemiologic and clinical research. We focus on the validity, reliability, and responsiveness to clinical intervention of these measures. Several adult questionnaires designed for epidemiologic research include questions on asthma and wheezing that have demonstrated repeatability and validity against concurrent measurements of nonspecific airway responsiveness. The International Union Against Tuberculosis Bronchial Symptoms Questionnaire was designed specifically to detect asthma and airway hyperresponsiveness in adult populations, and its reliability and validity have been well documented. A childhood questionnaire developed by Australian investigators has been demonstrated to provide information on asthma and wheezing that is reliable and valid against the criterion of concurrently measured nonspecific airway responsiveness. Although suitable for epidemiologic research, these questionnaires do not provide sufficient data on the severity of current asthma symptoms (aspects of which include intensity, duration, and frequency of symptoms) to be useful for clinical research involving subjects with established asthma. Many different methods of obtaining and analyzing symptom data have been used in clinical trials, but these have not received the methodologic scrutiny that allow the recommendation of a "best" approach for evaluating symptoms in clinical trials of interventions for asthma. The use of daily symptom diaries in short-term drug trials is common, but the optimal symptom-reporting interval for such studies has not been established. Similarly, a particular approach to integrating different symptoms (wheeze, dyspnea, cough, sputum) and the different aspects of these symptoms (intensity, duration, frequency) cannot be recommended on the basis of available data. Physical examination findings have little utility as asthma outcome measures because they may be normal between symptom episodes, they have relatively poor interobserver reliability, and they are relatively poor predictors of the outcome of emergency room visits for asthma. The finding of an elevated arterial PCO2 has utility as an indicator of a severe asthma attack, but arterial blood gas measurements have little other utility as asthma outcome measures. The chest radiograph is generally normal in patients with asthma and therefore not useful as an asthma outcome measure.(ABSTRACT TRUNCATED AT 400 WORDS)