Objectives: The authors identify demographic and clinical characteristics associated with noncompliance in patients beginning medical therapy for the treatment of glaucoma in a managed care setting.
Methods: The authors describe a retrospective cohort study in a group-model health maintenance organization in Massachusetts. Patients were members of the health maintenance organization who were newly initiated on topical drug therapy to treat open-angle glaucoma during the period January 1, 1987 through December 31, 1990, who met eligibility requirements, and who had evidence of health services utilization for a 12-month follow-up period. For all study subjects, we determined the number of days without available therapy for glaucoma during the 12-month period. Study subjects who did not fill prescriptions adequate to provide medication to cover at least 80% of days during the study period were considered noncompliant. Logistic regression analysis was used to assess demographic and clinical factors independently associated with noncompliance among patients initiated on medical therapy for the treatment of glaucoma.
Results: Of 616 subjects who met inclusion criteria, 152 (24.7%; 95% confidence interval, 21.3%-28.1%) met the study definition for noncompliance. These patients had an average number of days without therapy during the 12-month study period of 103.9 +/- 70.0 days compared with 6.8 +/- 19.5 days for those categorized as compliant. Of a variety of selected demographic and clinical characteristics, having fewer visits with an ophthalmologist during the study period (< 2) was most strongly related to noncompliance (odds ratio 2.99; 95% confidence interval 2.03, 4.40). There were no differences in average intraocular pressure between the compliant and noncompliant groups during the study period.
Conclusions: Noncompliance with prescribed medical therapy for glaucoma was found to be common in a managed care setting characterized by essentially unrestricted access to health care and medications. It remains difficult to identify noncompliant patients based on demographic and clinical characteristics. The use of automated prescription data to identify noncompliant patients is feasible in large managed health care insurance programs where such data are collected routinely for administrative purposes.