OBJECTIVE: To test the effectiveness of customized, family-oriented reminder letters in activating patients to seek appropriate preventive services. DESIGN: Randomized clinical trial. One group received computer-generated, customized letters explaining recommended preventive procedures for each family member. A second group received a form letter listing recommendations for all preventive procedures for all age and sex groups. A third group (control group) received no letters. SETTING: A private medical centre, without university affiliation, in rural Quebec. PARTICIPANTS: From 8770 patients who met study criteria, 719 families were randomly selected. Data were available for 1971 of 1998 patients in these families. MAIN OUTCOME MEASURES: The Family Received Index is the proportion of all procedures for which a family was overdue that they received. The Family End-of-study Up-to-date Index is the proportion of procedures for which the family was eligible and for which they were up-to-date at the end of the study. RESULTS: The Family Received Index for families mailed customized letters was more than double the index for patients not mailed letters (Kruskal-Wallis P = .0139). Comparison of the Family End-of-study Up-to-date indices also demonstrated that families of patients sent customized letters were more likely to be up-to-date than families not sent letters (Kruskal-Wallis P = .0054). No statistically significant difference appeared between the number of preventive measures received by the control group and the form-letter group. CONCLUSIONS: This study demonstrates a clinically small but statistically significant value to customizing reminder letters.