Author ORCID Identifier

https://orcid.org/0000-0003-4353-5235

Semester

Fall

Date of Graduation

2025

Document Type

Dissertation (Campus Access)

Degree Type

PhD

College

School of Pharmacy

Department

Pharmaceutical Systems and Policy

Committee Chair

Sabina Nduaguba

Committee Member

Khalid Kamal

Committee Member

Kimberly Kelly

Committee Member

Doug Thornton

Committee Member

Gerald Higa

Abstract

Potentially inappropriate prescription of opioids (PIP-OP) has been a major driver for the opioid epidemic and is associated with increased risk of opioid overdose and mortality. Given the lack of evidence of PIP-OP among patients with cancer, and within the sub-group (patients with cancer and chronic non-cancer pain (CNCP)), it is important to study the prevalence of PIP-OP and impact of PIP-OP among such populations.

This dissertation had four specific aims. The first aim was to conduct a systematic review with a meta-analysis of prevalence of PIP-OP and associated risk factors among patients with cancer. The study was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-analyses (PRISMA) guidelines through October 2024. We searched five electronic databases, namely PubMed, Scopus, Web of Science, CINAHL, and PsycINFO. A total of 11,248 articles were screened for title and abstract, and 509 articles were screened for full-text. Among them 35 studies were included. Within PIP-OP, fifteen articles were included for criteria I- high dose of opioid prescriptions (>=120MME per day), fifteen articles were included for criteria II- overlapping prescriptions of opioids and benzodiazepines only, and two articles for doctor shopping only. Further, two articles reported criteria I and criteria II together, and one article reported all three criteria together. Only six articles reported risk factors. Risk of bias was assessed using National Institute of Health quality assessment tool. Results of pooled estimate illustrate high prevalence of PIP-OP i.e., 13.21% (95%CI, 4.94%-33.49%). Results of sub-group analysis showed that prevalence of high dose of opioid prescriptions among patients with cancer was 13.21% (95%CI, 0.5%-80.40%), prevalence of overlapping opioid and benzodiazepine prescriptions was 14.39% (95%CI, 2.32%-54.36%), and prevalence of doctor shopping was 6.68%, (95%CI, 3.62%-65.66%). The results of our study showed that the prevalence of PIP-OP was high among patients with cancer, but evidence on risk factors was limited. Clinical strategies implemented to identify and prevent PIP-OP must be promoted and supported. Moreover, there was a major gap in prescribing practices and PIP-OP among patients with cancer and CNCP.

Therefore, our second aim was to determine the impact of CNCP on opioid prescribing patterns in patients with cancer. For this study, a retrospective cohort study was conducted utilizing Surveillance Epidemiology and End Results (SEER) – Medicare linked database from 2006 to 2019. Patients, at least 66 years old with histologically confirmed diagnoses of breast, lung, and colorectal cancer from 2007 to 2017, were included in the study. Patients with CNCP were identified in the baseline period using ICD9 and ICD10 codes. Patients without CNCP during the study period were included in the no-CNCP group. The follow-up period was one year after cancer diagnosis, and the outcomes were any opioid use, chronic opioid use, high-dose opioid use and total count of opioid prescriptions filled. Inverse probability weighting (IPTW) method was used to balance the covariates between the CNCP and non-CNCP group. Logistic and Poisson regressions were used to compare opioid use and opioid prescription patterns among the two groups. The results of the study showed that the prevalence of CNCP was higher in patients with lung cancer (41.21%), followed by breast cancer (40.10%), and colorectal cancer (36.42%, p< 0.01). Further, after weighting, the result of the logistic regression showed that the odds of chronic opioid use was high among patients with cancer and CNCP (breast cancer, (OR=2.77, 95%CI, (2.65-2.89)), lung cancer (OR=2.39, 95%CI, (2.27-2.51)), and colorectal cancer (OR=2.39, 95%CI, (2.25-2.53)), respectively when compared to patients without CNCP but their respective cancer. The study showed that CNCP adds burden in opioid prescription in active patients with cancer which necessitates a careful approach to opioid prescribing.

Next, our third aim was to determine the prevalence and risk factors of PIP-OP for patients with cancer and CNCP and investigate the association of PIP-OP with cardiovascular outcomes. This was a retrospective cohort study using SEER-Medicare database from 2006 to 2019. We included patients with histological confirmed and diagnosed primarily with lung, breast, and colorectal cancer from January 2007 to December 2017, who were at least 66 years of age. Further, patients diagnosed with CNCP, prescribed with opioids and enrolled continuously in Medicare part A, B and D during the study period were included. Patients enrolled in a health maintenance organization (HMO) and diagnosed with malignant pain or cancer related treatment pain during the study period were excluded. PIP-OP was defined based on Pharmacy Quality Alliance measures. Patients with PIP-OP were in exposure group, patients without PIP-OP with same inclusion and exclusion criteria were in no PIP-OP group. Additionally, a third control group was created consisting of patients with the same inclusion and exclusion criteria but with no opioid exposure during the study period. To calculate period prevalence and risk factors of PIP-OP, cohort 1 was created. For the second objective, association of PIP-OP with cardiovascular outcomes, cohort 2 was created with additional exclusion criteria. Patients were excluded if patients had cardiovascular events (i.e., myocardial infarction, coronary artery disease, and stroke) during the baseline period. Multivariable logistic regression was used to determine risk factors. We used 1:1 propensity score matching method was used in cohort 2 and Cox proportional hazard model was used to determine the association of PIP-OP with cardiovascular outcomes from the matched data. We observed the prevalence of PIP-OP was relatively low: 2.13% for high dose of opioid prescriptions, 2.60% for concurrent opioid and benzodiazepine prescriptions, and 0.30% for doctor shopping. Sociodemographic and clinical characteristics emerged as important risk factors of PIP-OP. Younger age, male sex, black race, and lower income were associated with higher risk of PIP-OP. Moreover, patients with lung cancer had higher risk of PIP-OP. Interestingly, PIP-OP is associated with increased risk of coronary artery disease when compared to patients with no opioid use. However, no significant difference was observed in terms of risk of overall cardiovascular events when compared to no PIP-OP group.

Lastly, our fourth aim was to estimate the economic burden associated with PIP-OP among patients with cancer and CNCP. We used the cohort 1 from aim 3 with same inclusion and exclusion criteria. Annual all-cause cost was calculated among patients with PIP-OP, no PIP-OP and no opioid use group. The secondary outcome was the annual all-cause health services utilization of PIP-OP and no PIP-OP group. Cost of PIP-OP was calculated from a payer’s perspective (Medicare) We used IPTW to balance the covariates between the CNCP and non-CNCP group. Using stabilized weighting, we used two-part models (logistic regression + generalized linear model) to estimate the annual cost. Additionally, we used multivariate logistic regression to determine the association of PIP-OP with HSU. Patients with PIP-OP had higher odds of emergency visit (OR = 1.29, 95%CI, 1.11-1.51) and hospitalization (OR = 1.19, 95%CI, (1.04-1.36) when compared to no PIP-OP group. Similarly, patients with PIP-OP group also had higher odds of emergency visit (OR = 1.56, 95%CI, 1.31-184), and hospitalization (OR = 1.53, 95%CI, (1.29-1.82)) when compared to patients with no opioid group. The total annual cost of PIP-OP group and no opioid group was $37,732 (95%CI, $29,298-$48,204) and $31,870(95% CI, $30,233 - $33,708) respectively per person. The annual incremental cost of PIP-OP when compared to no opioid group was $5,862 (95%CI, -$2,971–16,459). Similarly, the result of cost analysis of PIP-OP when compared to no PIP-OP group. The total annual cost of PIP-OP group and no PIP-OP group was $40,975 (95%CI, $36,007–$46,470) and $36,744(95%CI, $36,919–$37,674) respectively per person. The annual incremental cost associated with PIP-OP was $4,231 (95%CI, -$818– $9,701) respectively.

This dissertation provides a comprehensive evaluation of PIP-OP in patients with cancer, with a novel focus on those with CNCP. To summarize, the first aim provided insights into the prevalence and risk factors of PIP-OP among patients with cancer. The results showed that the prevalence of PIP-OP is higher among patients with cancer, but there was limited evidence on risk factors. To extend this work, the second aim focused on the subgroup of patients with cancer and CNCP. Findings demonstrated that CNCP substantially increases opioid utilization in patients with active cancer, highlighting the need for careful prescribing in this high-risk population. Building on the evidence from the first and second aim, the third and fourth aim investigated the prevalence, risk factors, cardiovascular outcomes, economic burden, and health-services utilization associated with PIP-OP. Although the overall prevalence of PIP-OP was relatively low, younger age, male sex, Black race, lower income, and lung cancer were risk factors of PIP-OP. Importantly, PIP-OP was associated with an elevated risk of coronary artery disease and increased health services utilization. From the results of the dissertation, we were able to bridge the critical gap in understanding PIP-OP among patients with cancer, prescribing practices for the treatment of CNCP among patients with cancer, provide the evidence of PIP-OP among patients with cancer and determine the impact of PIP-OP among patients with CNCP and cancer. These findings provide evidence to inform future research, policy development, and clinical interventions targeted for the optimization of pain management in cancer populations while reducing the risks of inappropriate opioid prescribing.

Download

Share

COinS