Background

Obesity, non-cancer pain and opioid use have been shown to be positively associated. However, little is known regarding the association between opioid use and opioid-related healthcare cost following bariatric surgery (BS). This study examined factors related to opioid use and opioid-related health care costs one-year post-BS.

Methods

This retrospective cohort study identified patients with BS, their opioid use and health care costs. Data was extracted using the Utah All Payer Claims Database (APCD), which includes data representing 80% of the Utah population and over 90% of the commercial health insurance market. The APCD avoids disconnect payment systems and promotes data capture of subjects visiting multiple clinics or hospitals across different delivery systems.

Results

Subjects (n=1482) had laparoscopic BS between 2013 and 2014 (gastric bypass, 569; banding, 190; sleeve, 445; and duodenal switch, 278). All patients had continuous insurance coverage and health care claims information one year before and one year after BS were analyzed. Opioid prescription one month after BS was excluded. Costs were calculated as the sum of reimbursement, co-insurance and out-of-pocket payments. All costs to reflect inflation were adjusted to 2015 dollars using the Personal Health Care Expenditure component of the National Health Expenditure Accounts. Mean age was 48 years and 78% were female. Overall pre-BS, mean (sd) number of opioid prescriptions was 0.61 orders (2.10) and 1.49 orders (4.09) one-year post-BS. Mean cost (sd) for opioid use one-year post-BS was $132.02 ($908.25). Subjects with sleeve had lower opioid use (1.03 orders) and health care cost ($47.59) post-BS as compared to subjects with gastric bypass (1.78; $160.95) and duodenal switch (1.47; $174.90). Using logistic regression, banding patients had lower odds for opioid prescription orders than gastric bypass patients (reference group) (OR, 0.90; p=0.04), while duodenal switch patients had higher opioid prescription orders (OR, 2.13; p-=0.05) compared to gastric bypass. Patients with opioid use history had higher odds of opioid prescription orders post-BS (OR, 1.5; p<0.001) compared to patients without previous history. Generalized linear regression with gamma distribution and log link function showed that age (coefficient= -0.02, p=0.02), opioid use cost pre-BS (coefficient=0.03, p<0.001) and duodenal switch (vs. gastric bypass) (coefficient=0.65, p=0.04) were statistically significant variables related to higher or lower opioid cost post-BS.

Conclusions

These data suggest examining why opioid use increases after surgery and whether or not this increase relates to increased risk behaviors (i.e. alcohol abuse and self-injury) post-BS.