Evaluating the Effects of Opioid Prescribing Policies on Patient Outcomes in a Safety-net Primary Care Clinic

Background After decades of liberal opioid prescribing, multiple efforts have been made to reduce reliance upon opioids in clinical care. Little is known about the effects of opioid prescribing policies on outcomes beyond opioid prescribing. Objective To evaluate the combined effects of multiple opioid prescribing policies implemented in a safety-net primary care clinic in San Francisco, CA, in 2013–2014. Design Retrospective cohort study and conditional difference-in-differences analysis of nonrandomized clinic-level policies. Patients 273 patients prescribed opioids for chronic non-cancer pain in 2013 at either the treated (n=151) or control clinic (n=122) recruited and interviewed in 2017–2018. Interventions Policies establishing standard protocols for dispensing opioid refills and conducting urine toxicology testing, and a new committee facilitating opioid treatment decisions for complex patient cases. Main Measures Opioid prescription (active prescription, mean dose in morphine milligram equivalents [MME]) from electronic medical charts, and heroin and opioid analgesics not prescribed to the patient (any use, use frequency) from a retrospective interview. Key Results The interventions were associated with a reduction in mean prescribed opioid dose in the first three post-policy years (year 1 conditional difference-in-differences estimate: −52.0 MME [95% confidence interval: −109.9, −10.6]; year 2: −106.2 MME [−195.0, −34.6]; year 3: −98.6 MME [−198.7, −23.9]; year 4: −72.6 MME [−160.4, 3.6]). Estimates suggest a possible positive association between the interventions and non-prescribed opioid analgesic use (year 3: 5.2 absolute percentage points [−0.1, 11.2]) and use frequency (year 3: 0.21 ordinal frequency scale points [0.00, 0.47]) in the third post-policy year. Conclusions Clinic-level opioid prescribing policies were associated with reduced dose, although the control clinic achieved similar reductions by the fourth post-policy year, and the policies may have been associated with increased non-prescribed opioid analgesic use. Clinicians should balance the urgency to reduce opioid prescribing with potential harms from rapid change.