OBJECTIVES:
* Determine whether rates of surveillance colonoscopy can be substantially increased by linking computerized data on endoscopic screening procedures to computerized pathology data to identify patients with prior adenomatous polyps who are due for repeat colonoscopy.
* Evaluate whether the impact of the intervention varies by patient characteristics, including age, sex, or site of primary care.
OUTLINE: This is a randomized, controlled, crossover study. Patients are stratified according to age (≤ 65 years of age vs \> 65 years of age), gender, time elapsed since last colonoscopy, and primary care physician's (PCP) membership in Harvard Vanguard Medical Associates (yes vs no). Patients are randomized to 1 of 2 arms.
* Arm I: PCPs are mailed response forms (regarding the disposition of their eligible patients \[e.g. patient letter mailed, patient called, scheduled for colonoscopy, colonoscopy completed elsewhere, patient refused procedure, clinical contraindication, patient deceased\]) and letters reminding them that their patients are due for a follow-up colonoscopy. PCPs who do not return response forms within 4 weeks are sent a second set of reminders and patient letters.
* Arm II: Six months after the beginning of the study, PCPs are mailed response forms and patient letters once. A second set of reminders and patient letters are not sent.
PROJECTED ACCRUAL: A total of 800 patients will be accrued for this study.
