BACKGROUND: Medication adherence is a serious problem and an important health care challenge. Despite evidence indicating benefit, many patients do not take prescribed medications. To increase understanding within a large, integrated health system, we studied patient characteristics associated with medication adherence. OBJECTIVE: To use automated pharmacy records to assess patient characteristics associated with adherence across 8 diseases. METHODS: We identified all members older than 18 years with at least 2 prescription fills for greater than a 28 days supply from January 2007 through March 2009. The integrated health system covers approximately 750,000 members. Medications (n = 128) to treat the following diseases were identified: depression, hypertension, hyperlipidemia, diabetes, asthma/COPD, multiple sclerosis, cancer, or osteoporosis. Diagnoses were identified by ICD-9-CM codes and merged with pharmacy claims. We examined differences by age, sex, race, comorbidity, and geo-code. Patients with 1 condition and 1 prescribed medication (n = 15,334) were analyzed. Analyses were repeated on patients with any number of conditions and medications among those specified (n = 31,636 patients). Adherence was calculated by the medication possession ratio at a threshold of 80%. RESULTS: Adherence differed significantly by patient characteristics. Males were more adherent than females. Males, Caucasians, older patients, and patients living in areas having a higher proportion of residents with a high school education, lower poverty, and higher family income were associated with increased adherence. Among patients with any number of conditions, adherence increased with lower Charlson score and fewer conditions and medications. Substantial variation in adherence was found by condition. Specifically, multiple sclerosis, hypertension, hyperlipidemia, osteoporosis, and cancer had a high percentage of adherent patients (75% +). Adherence rates for depression, diabetes, and asthma/COPD were 62%, 51%, and 33%, respectively. CONCLUSIONS: We expected adherence to be less significant within an integrated health system because pharmacy is a covered benefit, and members have enhanced medication access. However, significant differences were found by patient characteristics. Great room for improvement remains, specifically for diabetes and asthma. Patient-specific, targeted efforts might be considered to improve adherence.