Figure 1. Cohort identification and data collection.
Patients in the current study were adult patients who had ≥1 outpatient opioid prescription and ≥1 ICD-9 code indicating musculoskeletal pain between July 1, 2011 and June 30, 2012. We excluded patients with any outpatient opioid prescriptions in the 1 year before their first or incident prescription in the cohort window. Data were collected for 1 year after the incident prescription. Our final cohort comprised patients with 1 opioid prescription in each quarter after their incident prescription. Dates were converted to days since each patient’s incident prescription.