To examine the effect of initial family interviews by health care providers on patients' use of health services, 177 patients and their families were randomly assigned to an interviewed (I) or a non-interviewed (NI) subgroup. Initial interviews would encourage families to bring their children in for health supervision visits, but discourage families from making after-hour telephone calls, using the emergency room and bringing the children to the clinic frequently for problem visits. The I families had an initial interview, attended by all family members. Both a physician and a nurse elicited patient histories and explained use of the emergency room, when to make after-hour calls, how to schedule appointments and other information about the clinic. If I families failed to have an initial interview, they were deleted from the study. In the NI subgroup, patient histories were elicited during a routine health supervision visit without the entire family in attendance, and information about emergency room visits, after-hour calls and appointment scheduling was provided during the same visit. After one year (1987) into the study and two years (1988) into the study, all patient charts were examined. Data analysis was performed using analysis of variance for repeated measures (ANOVA) and step-wise multiple regression of Statistical Analysis System. For 1987, the interview intervention explained a significant (p = 0.01) amount of variance in the number of problem visits (less in I) after controlling for months in the study and age of the child.(ABSTRACT TRUNCATED AT 250 WORDS)