Post-lumbar surgery prescription variation and opioid-related outcomes in a large US healthcare system: an observational study

Abstract

Background context: Spinal decompression and fusion procedures are one of the most common procedures performed in the United States (US) and remain associated with high postsurgical opioid burden. Despite guidelines emphasizing nonopioid pharmacotherapy strategies for postsurgical pain management, prescribing practices are likely variable and guideline-incongruent.

Purpose: The purpose of this study was to characterize patient-, care-, and system-level factors associated with opioid, nonopioid pain medication, and benzodiazepine prescribing variation in the US Military Health System (MHS).

Study design/setting: Retrospective study analyzing medical records from the US MHS Data Repository.

Patient sample: Adult patients (N=6,625) undergoing lumbar decompression and spinal fusion procedures from 2016 to 2021 in the MHS enrolled in TRICARE at least a year prior to their procedure and had at least one encounter beyond the 90-day postprocedure period, without recent trauma, malignancy, cauda equina syndrome, and co-occurring procedures.

Outcome measures: Patient-, care-, and system-level factors influencing outcomes of discharge morphine equivalent dose (MED), 30-day opioid refill, and persistent opioid use (POU). POU was defined as dispensing of opioid prescriptions monthly for the first 3 months after surgery and then at least once between 90 and 180 days after surgery.

Methods: (Generalized) linear mixed models evaluated multilevel factors associated with discharge MED, opioid refill, and POU.

Results: The median discharge MED was 375 mg (IQR 225, 580) and days' supply was 7 days (IQR 4, 10); 36% received an opioid refill and 5%, overall, met criteria for POU. Discharge MED was associated with fusion procedures (+151-198 mg), multilevel procedures (+26 mg), policy release (-184 mg), opioid naïvty (-31 mg), race (Black -21 mg, another race and ethnicity -47 mg), benzodiazepine receipt (+100 mg), opioid-only medications (+86 mg), gabapentinoid receipt (-20 mg), and nonopioid pain medications receipt (-60 mg). Longer symptom duration, fusion procedures, beneficiary category, mental healthcare, nicotine dependence, benzodiazepine receipt, and opioid naivety were associated with both opioid refill and POU. Multilevel procedures, elevated comorbidity score, policy period, antidepressant receipt, and gabapentinoid receipt, and presurgical physical therapy were also associated with opioid refill. POU increased with increasing discharge MED.

Conclusions: Significant variation in discharge prescribing practices require systems-level, evidence-based intervention.

Keywords: Benzodiazepines; Health services research; Opioids; Pain medication; Prescribing practices; Spinal surgery.

Figure 1.

Variation in median discharge morphine equivalent dose (MED) by provider. Each point represents a provider with at least 10 lumbar surgeries performed at any of the 18 treatment facilities included in the study. Treatment facilities are ordered from highest (top) to lowest (bottom) patient volume. The dashed line indicates the median MED for all opioid prescriptions (375 mg) dispensed at time of discharge.

Figure 2.

Variation in median discharge morphine equivalent dose (MED) by lumbar surgery and surgeon type. Dashed lines indicate the median MED for lumbar decompression-only surgeries (315 mg) and posterior- (450 mg) and anterior-only-fusions (450 mg) for opioid prescriptions dispensed at discharge. Treatment facilities are ordered from highest (top) to lowest (bottom) patient volume. Note – certain facilities only have one type of surgeon.

Figure 3.

Variation between treatment facilities in discharge morphine equivalent dose following lumbar surgery. Facilities are ordered by patient volume from highest (top left) to lowest (bottom right).

Figure 4.

Variation between treatment facilities in discharge opioid and non-opioid prescriptions following lumbar surgery. Facilities are ordered by patient volume from highest (top left) to lowest (bottom right).