Association Between Smoking Status and Opioid Dose in Prescriptions Written for Breast Cancer-related Pain.

BACKGROUND

Cancer-related pain has historically been undertreated. Prescription opioids have been shown to be an integral part of the treatment of cancer pain. Despite the significant amount of scientific evidence that smoking is associated with variation in pain expression and opioid misuse in both cancer and non-cancer populations, little is known about the association between smoking status and opioid utilization in cancer populations.

OBJECTIVES

To assess the association between smoking status and high-risk opioid-prescribing behaviors of oncologists prescribing opioids in the outpatient setting to patients with breast cancer-related pain.

STUDY DESIGN

A retrospective cross-sectional study of opioid prescriptions written by oncologists for breast cancer-related pain was conducted using the Patient Cohort Explorer (PCE) database at the University of Mississippi Medical Center (UMMC) from March 15, 2015 to March 15, 2017.

SETTING

Tertiary academic medical center.

METHODS

De-identified data from UMMC PCE were utilized for this study. Patient-level information, such as age, gender, race, insurance status, and smoking status, were also selected for each prescription. Prescription-level data, such as name of opioid, dose, frequency, route, and primary diagnosis, were also obtained. Prescriptions were included if they are written in the outpatient setting, for breast cancer-related pain, and for women 18 years or older. Prescriptions were excluded if they were written by a specialist other than a medical oncologist or if the information necessary to calculate morphine milligram equivalence (MME) was missing.

RESULTS

The sample consisted of 577 opioid prescriptions that were written in the outpatient setting to women ages 18 years and older for breast cancer-related pain. The majority of the sample were ages 46 to 64 years (60.5%), Nonwhite (75.2%), publicly insured (66.2%), and with nonmetastatic disease (86.1%). Almost one-fifth (19.6%) of the prescriptions were written to current smokers, 21.3% to former smokers, and 58.1% to nonsmokers. Nonsmoking status predicted an increased odds of receiving a prescription ≥ 50 MME (odds ratio [OR] = 1.98, 95% confidence interval [CI]: 1.08-3.60, P = 0.030) and ≥ 90 MME (OR = 6.29, 95% CI: 1.38-28.58, P = 0.017) compared to current smokers. Nonsmoking status also predicted an increased odds of receiving a prescription ≥ 90 MME (OR = 4.29, 95% CI: 1.43-12.92, P = 0.009) compared to former smokers.

LIMITATIONS

This cross-sectional sample was drawn from a single institution and only included the breast cancer population and may not be generalizable to other populations or institutions. Second, our sample was drawn from secondary data not collected for the purposes of our study. This limits the inclusion of other variables that may impact the opioid-prescribing behaviors of oncologists, potentially resulting in bias.

CONCLUSIONS

During a time of heightened awareness of opioid-related harm, as well as implementation of national opioid-prescribing guidelines, current smoking may potentially be impacting how oncologists evaluate the need for opioids to treat breast cancer-related pain. Further studies that examine the relationship between smoking status, perceived need for opioids, and evaluative need for opioids in cancer populations are warranted.

KEY WORDS

Cancer pain, opioids, smoking, breast cancer, opioid-prescribing guidelines, health policy, oncology, end of life.