Background: Poor patient attendance to scheduled flexible sigmoidoscopy or colonoscopy may contribute to deficient colorectal cancer screening.

Objective: To examine the association of physician appointment-keeping behavior with attendance to scheduled endoscopic studies of the colon.

Design: Retrospective cohort.

Setting: 23 sites performing endoscopic procedures in a health care system.

Patients: 11 803 patients scheduled for a first colon study with 3 or more scheduled physician visits from June 1999 through November 2001.

Measurement: 2 outcomes from health system computerized records: 1) attendance at the first scheduled colon study and 2) among nonattendees, attendance at the study rescheduled within 6 months. Physician visit adherence was defined as the proportion of physician visits kept, grouped by quartile. Adjusted associations were examined in conditional logistic regression.

Results: Of 11 803 patients, 62% attended the first colon study. Of the 4496 nonattendees, 2739 (61%) rescheduled and, of these, 64% kept that appointment. Compared with the highest quartile of physician visit adherence (>85%), the adjusted odds ratio of attending the first colon study decreased as physician visit adherence decreased: Adjusted odds ratios were 0.94 (95% CI, 0.89 to 1.00) for 76% to 85% adherence, 0.87 (CI, 0.81 to 0.92) for 66% to 75% adherence, and 0.79 (CI, 0.73 to 0.85) for adherence of 65% or less. Among nonattendees who rescheduled, the lowest quartile of physician visit adherence (< or =65%) was the only statistically significant predictor of attending the rescheduled study (adjusted odds ratio, 0.87 [CI, 0.78 to 0.98]).

Limitations: The adherence measure applies only to patients with at least 3 scheduled visits. Persons having a colon study outside of the system could have been misclassified.

Conclusion: Physician appointment-keeping behavior predicted attendance to colorectal endoscopic studies in this cohort and may help identify persons who need interventions to promote adherence.