This study examined the direct association between medical costs and cigarette smoking and modification of interest in visiting medical facilities. The study sample consisted of 966 males 30 years of age or older in a rural town in Kyoto Prefecture, who were members of the National Health Insurance. The subjects were respondents to a previous questionnaire about smoking in February, 1989. There were 544 current smokers, 226 ex-smokers and 141 nonsmokers. Fifty-five subjects did not answer the question about smoking habits. Outpatient medical costs paid by National Health Insurance for the fiscal year 1989 were analyzed. Average medical costs of ex-smokers were the highest of all groups-current, ex-, and non-smokers (30-49, 50-69, and 70+ years of age). Those of current smokers and of the combined current or ex-smokers group were higher than those of non-smokers (30-49 and 50-69 years of age). None of these differences were significant by nonparametric test. The effect of smoking on medical costs was evaluated by odds ratio of the largest class of costs (50,000+ yen for 30-49 years of age, 200,000+ yen for 50-69 years of age, and 500,000+ yen for 70+ years of age) for various types of smokers compared with non-smokers by logistic regression. The odds ratios of current, ex-, and current or ex-smokers 30-49 and 50-69 years of age were between 2 to 3, but not significant. The odds ratio of subjects 70 years of age or older was smaller than unity and not significant.(ABSTRACT TRUNCATED AT 250 WORDS)