A study of respiratory diseases was conducted with 10,938 persons, 10,628 males, employed as active union carpenters for at least 3 months between 1988 and 1989. The mean age at entry into the cohort was 36 years. The occurrence of lung diseases, as defined by the International Classification of Diseases, 9th revision (ICD-9) code, in the cohort was determined by linking the subjects to union administration records and private health insurance and workers' compensation claims. The subjects completed a questionnaire on respiratory health history, smoking, and potential workplace and home exposures. The incidence rates for chronic bronchitis and asthma were compared with data obtained in the occupational health supplement of the 1988 National Health Interview Survey (NIHS). The incidence of lung cancer was compared with data obtained in the Surveillance, Epidemiology, and End Results (SEER) program. A nested case control study of the identified asthma cases was also conducted using the ICD-9 code and the Burney case definitions of asthma. A total of 931 cases of lung disease among 812 subjects occurred during the study period. Acute bronchitis, asthma, chronic obstructive pulmonary disease, and chronic bronchitis were the most frequently occurring diseases, 511, 213, 90, and 51 cases being detected, respectively. The crude incidence rates were 1.81, 0.76, 0.32, and 0.18%. The incidence rates increased with increasing age. The incidence rates of chronic bronchitis and asthma were generally similar to those reported by the NIHS. Thirteen cases of lung cancer occurred versus 10.33 expected from the SEER database, representing a standardized cancer incidence ratio of 1.3. The case control analysis of the asthma cases defined by the ICD-9 code indicated significantly increased risks associated with exposure to hay, epoxy paints, animal fur or wastes, enzymes, and molds, odds ratios (ORs) after adjusting for age, sex, smoking history, and length of union membership varying from 1.5 to 2.2. When defined according to Burney criteria, exposure to a much larger number of agents such as most dusts, metal dusts, glues, acids, alkalis, and concrete form oils was associated with a significantly increased asthma risk, adjusted ORs varying from 1.6 to 5.1. The authors conclude that these analyses illustrate the way combined data sources (administrative, health insurance, and workers' compensation records) can be used for epidemiologic research.