OBJECTIVE To examine differences between frequent and infrequent users in decisions to seek medical care based on symptom type and duration of experience of the symptom. METHODS Retrospective study. METHODS University family practice clinic. METHODS Fifty-eight of 130 patients returned the completed instruments, a response rate of 45%. METHODS The following two instruments were used: the Intentions to Seek Care questionnaire, which measures intention to see a physician, self-treat, or use medications for 11 different symptoms after three time periods--1 day, 3 days, and 1 week of symptom duration--and the SF-36 Health Survey measure. RESULTS Initial analyses suggested that patients perceive three clusters of symptoms, which we have labeled as serious, mild, and psychological. Scale scores for these groups of symptoms yielded significant reliability and were used in subsequent analyses. Multivariate analysis of variance comparing source of care, symptom duration, and frequent vs infrequent user groups indicated that both user groups are more likely to seek care from physicians for symptoms that are serious. Analysis of mild and psychological symptoms yielded significant three-way interactions, with frequent users more likely than infrequent users to seek care from physicians at a 1-day duration. Such differences disappear by 7 days. Significantly lower scores were obtained on five of the eight SF-36 Health Survey dimensions for the infrequent users. CONCLUSIONS Frequent users are more likely than infrequent users to report that they would seek care for minor symptoms. Infrequent users tend to self-treat for 3 to 7 days, after which time physician advice is sought. Efforts to teach patients optimal timing of use for different symptoms may be an effective intervention.