Abstract
Background
Opioid exposure during cancer therapy may increase long‐term unsafe opioid prescribing. This study sought to determine the rates of coprescription of benzodiazepine and opioid medications and new persistent opioid use after surgical treatment of early‐stage cancer.
Methods
A retrospective cohort study was conducted among a US veteran population via the Veterans Affairs Corporate Data Warehouse database. Participants were opioid‐naive persons aged ≥21 years with a new diagnosis of stage 0–III cancer between January 1, 2015, and December 31, 2016. Outcomes were days of coprescription of benzodiazepines and opioids in the 13 months posttreatment and new persistent opioid use. The exposure was total morphine milligram equivalents (MMEs) attributed to treatment and prescribed from 30 days before through 14 days after the index surgical procedure.
Results
Among 9213 veterans, coprescription of benzodiazepines and opioids occurred in 366 patients (4.0%) and new persistent opioid use in 981 patients (10.6%). In a linear model adjusting for patient, clinical, and geographic factors, persons in the highest quartile compared to no opioid exposure had increased days with coprescription of benzodiazepines and opioids (mean difference, 1.0; 95% CI, 0.3–1.7). In a discrete time survival analysis, persons in the highest quartile of MME exposure compared to none had a greater risk of new persistent opioid use (hazard ratio, 1.6; 95% CI, 1.3–1.9).
Conclusions
More than one of 10 opioid‐naive veterans undergoing curative‐intent surgical treatment for cancer developed new persistent opioid use. Optimizing cancer treatment pain management strategies to mitigate long‐term opioid‐related health risks is crucial.
Keywords: cancer survivorship, curative‐intent surgical cancer treatment, early‐stage cancer, new persistent opioid use
Short abstract
Among opioid‐naive veterans with early‐stage cancer who underwent curative‐intent surgical therapy, increased exposure to opioids in the treatment period was associated with coprescription of opioids and benzodiazepines and incidence of new persistent opioid use in the first year postsurgery.
INTRODUCTION
In response to the growing opioid epidemic and recognition that pain may be undertreated in certain populations, guidelines for the use of opioid medications in pain management have evolved over recent years. 1 , 2 Although effective pain management during the cancer diagnostic and treatment period is essential, prescription opioid practices associated with cancer treatment may lead to unsafe and long‐term opioid treatment (LTOT) and the related adverse outcomes of opioid use disorder, opioid overdose, and nonoverdose adverse events, including suicide.
New persistent opioid use (NewPersOU) has occurred after surgical procedures for both noncancer conditions and early‐stage cancer. 3 , 4 NewPersOU has been reported at 10.4% among patients with a cancer diagnosis undergoing curative‐intent surgery. 5 Cancer survivors are at risk for LTOT. This risk is further increased among those who have low incomes, experience unemployment at diagnosis, live in urban areas or areas with lower educational attainment, and have higher comorbidity or are tobacco users. 6 , 7 Polypharmacy is also a concern among cancer survivors, 8 with one study estimating that 10% of cancer survivors were coprescribed opioid and benzodiazepine medications 1 year after cancer therapy. 9
Veterans may face an even greater risk of adverse outcomes associated with opioid use attributable to cancer treatment. The incidence of cancer among persons who receive health care in US Veterans Affairs medical centers (VAMCs) approximates that of the general US male population. 10 However, veterans may face a particularly high risk of adverse outcomes associated with opioid use in the treatment period because of the higher baseline incidence of mental health, tobacco use, and substance use disorders among this population. 11 , 12 A study of 106,732 US veterans diagnosed with early‐stage cancer between 2000 and 2015 reported persistent opioid use in 8.3% and 3.5% of the total cohort and opioid‐naive subset, respectively. 13 However, this study did not limit the sample to those receiving a curative‐intent surgical procedure, and did not evaluate the outcome of coprescription of benzodiazepines and opioids (CoBenzOp), a measure of unsafe opioid prescribing practices.
This study seeks to evaluate the impact of opioid prescribing practices attributable to the index surgical treatment across common cancer types in the US veteran population. We elucidate the incidence of CoBenzOp and NewPersOU in the year after definitive surgical treatment for stage 0–III cancers among an opioid‐naive cohort of US veterans. In addition, we seek to identify patient, clinical, and treatment factors associated with the outcomes of CoBenzOp and NewPersOU.
MATERIALS AND METHODS
Study design
We conducted a national retrospective cohort study of a US veteran population receiving care at VAMCs. Inclusion criteria were age 21 years or older, a diagnosis of early‐stage (stage 0–III) cancer between January 1, 2015, and December 31, 2016, treatment including a definitive surgical procedure, diagnosis and first course of treatment received at a VAMC, and at least one visit in primary care at a VAMC in the 12 months before the cancer diagnosis. 11 , 14 Eleven cancer types were selected on the basis of their prevalence and having surgical treatments associated with postoperative pain: prostate, bladder, lung, oropharynx, colon, rectum, esophagus, stomach, liver, pancreas, and soft tissue sarcoma. Breast cancer was not included because the VA population is primarily male, and the majority of breast cancers diagnosed among veterans are treated outside of VAMCs. 15 The cohort was restricted to veterans who were opioid naive, defined as having no opioid prescriptions 12 months to 30 days before the index surgical procedure. This restriction included veterans prescribed buprenorphine formulations indicated for pain management in this time frame. Outpatient prescriptions for methadone were most often prescribed at non–hospital‐based clinics, and were not available in the VA Corporate Data Warehouse (CDW) database. Therefore, methadone prescriptions in this time period were not an exclusion criterion. Additional exclusion criteria were (1) a prior cancer diagnosis, with the exception of a diagnosis of nonmelanoma skin cancer (a common condition with a favorable prognosis), 16 and (2) a hospice referral in the year before diagnosis, during the treatment period, or during the follow‐up period because goals of care may differ.
Databases
Databases included the CDW Cancer Cube and raw oncology file, 17 inpatient and outpatient pharmacy data files, and inpatient and outpatient encounter files. In the CDW files, race and ethnicity data were captured after being self‐identified by the veteran, and were reported as one response for race and a second response for ethnicity. Additional databases included the Master Veteran Index table to assess mortality, and the linked VA and Centers for Medicare & Medicaid Services (CMS) data set to identify veterans with access to benefits from Medicare Part D. Cancer diagnosis, surgery, and prescription data were obtained from the CDW database. VA databases are designed for accurate linkages and connections across multiple data tables, and up‐to‐date patient identifiers were used to ensure >95% accuracy in these linkages as determined by VA data quality analytic teams.
Opioid use in the baseline treatment period
We defined the baseline treatment period as 30 days before through 14 days after the date of the index surgical procedure. We considered prescriptions in the month before surgery to be associated with the index procedure because prescriptions for postoperative pain may be issued in advance of the surgical date. 3 , 5 We identified all inpatient and outpatient prescriptions of opioid drug formulations used to treat pain and issued to veterans in this baseline treatment period. The following drugs were included in our definition of opioid medications: buprenorphine formulations used for pain, codeine, fentanyl, hydrocodone, hydromorphone, levorphanol, meperidine, morphine, oxycodone, oxymorphone, tapentadol, and tramadol. We excluded codeine formulations used primarily for cough and buprenorphine formulations used for opioid use disorder. Given that outpatient prescriptions for methadone issued by the VA are primarily managed in a separate clinic system and not captured in the CDW database, we did not include methadone prescriptions in the analysis. We determined the total morphine milligram equivalent (MME) of opioids prescribed via Centers for Disease Control and Prevention (CDC) opioid oral MME conversion factors. 18 We also identified veterans who received an opioid medication within 24 h of discharge, including oral, intravenous (iv), and cartridge administrations, for use in a sensitivity analysis of the impact of opioid exposure within 1 day of hospital discharge on the outcomes of interest.
Covariates
We characterized sociodemographic factors, the region in which surgical treatment was received, clinical factors, baseline medical conditions, and use of nonopioid pain medications (Table 1). Lower socioeconomic status (SES) was defined as having a VA priority group 5 level, which indicates VA pension benefits, eligibility for Medicaid, or an annual income below the adjusted income limits based on zip code of residence. Clinical factors included tumor type and stage, and receipt of adjuvant radiation, immunotherapy, or chemotherapy. Baseline medical conditions including mental health disorders, chronic pain, opioid use disorder, substance use disorder, a history of suicide ideation, self‐inflicted harm, and opioid overdose were identified if coded in the CDW database in the year before diagnosis (Tables S1 and S2). Comorbidity was measured with the Elixhauser Comorbidity Index. Nonopioid pain medication was categorized as the following: nonopioid analgesics, nonsteroidal anti‐inflammatory analgesics, topical analgesics, local anesthetics, urinary analgesics, skeletal muscle relaxants, and anticonvulsants. Enrollment in Medicare Part D as identified in the VA/CMS database was used as a covariate in our analyses to account for potential dual‐system receipt of prescriptions from the VA and/or Medicare. 19 , 20
TABLE 1.
Description of study population‐stratified opioid prescription in the treatment period.
| Covariate | Total | Baseline treatment period | p | |
|---|---|---|---|---|
|
N = 9213, No. (%) |
Any opioids, n = 6970 (76%), No. (%) |
No opioids, n = 2243 (24%), No. (%) |
||
| Age at diagnosis, years | ||||
| <45 | 107 (1.2) | 79 (1.1) | 28 (1.3) | .002 |
| 45–64 | 3004 (32.6) | 2340 (33.6) | 664 (29.6) | |
| >65 | 6102 (66.2) | 4551 (65.2) | 1551 (69.2) | |
| Sex: male | 9036 (98.1) | 6842 (98.2) | 2194 (97.8) | .296 |
| Race | ||||
| Black | 1687 (19.1) | 1299 (19.5) | 388 (18.1) | .065 |
| White | 6965 (79.0) | 5237 (78.5) | 1728 (80.5) | |
| Other a | 162 (1.8) | 132 (1.9) | 30 (1.4) | |
| Hispanic ethnicity | 456 (5.1) | 324 (4.8) | 132 (6.0) | .019 |
| Elixhauser comorbidity score | ||||
| <0 | 2184 (23.7) | 1652 (23.7) | 532 (23.7) | .149 |
| 0 | 3671 (39.9) | 2785 (39.9) | 886 (39.5) | |
| 1–3 | 1151 (12.5) | 895 (12.8) | 256 (11.4) | |
| >3 | 2207 (23.9) | 1638 (23.5) | 569 (25.3) | |
| Cancer type | ||||
| Bladder | 2302 (24.9) | 1680 (24.1) | 622 (27.7) | <.0001 |
| Colorectal | 2393 (25.9) | 1620 (23.2) | 773 (34.5) | |
| Lung | 1252 (13.6) | 1138 (16.3) | 114 (5.1) | |
| Prostate | 2594 (28.2) | 2029 (29.1) | 565 (25.2) | |
| Other | 672 (7.3) | 503 (7.2) | 169 (7.5) | |
| Cancer stage | ||||
| 0 | 1758 (19.1) | 1110 (15.9) | 648 (28.9) | <.0001 |
| I | 2790 (30.3) | 2135 (30.6) | 655 (29.2) | |
| II | 3021 (32.8) | 2399 (34.4) | 522 (27.7) | |
| III | 1644 (17.8) | 1326 (19.0) | 318 (14.2) | |
| Adjuvant treatment | ||||
| Surgery only | 6495 (70.5) | 1667 (74.3) | 4828 (69.3) | <.0001 |
| Surgery + radiation | 253 (2.8) | 52 (2.3) | 201 (2.9) | |
| Surgery + immunotherapy | 452 (4.9) | 108 (4.8) | 344 (4.9) | |
| Surgery ± radiation ± chemotherapy ± immunotherapy | 2013 (21.9) | 416 (18.6) | 1597 (22.9) | |
| OUD in year before diagnosis: yes | 76 (0.8) | 60 (0.86) | 16 (0.71) | .501 |
| Baseline pain medication use: yes | 4340 (47.1) | 3381 (48.5) | 959 (42.8) | <.0001 |
| Alcohol use disorder: yes | 787 (8.5) | 627 (9.0) | 160 (7.1) | .006 |
| Tobacco use: yes | 1964 (21.3) | 1609 (23.1) | 355 (15.8) | <.0001 |
| Anxiety/depression/PTSD: yes | 2050 (22.3) | 1595 (22.9) | 455 (20.3) | .011 |
| Bipolar/schizophrenia: yes | 404 (4.7) | 308 (4.4) | 99 (4.4) | .992 |
| Dual enrollment with Medicare Part D: yes | 1932 (20.9) | 1398 (20.1) | 534 (23.8) | .0001 |
| Service connection: yes b | 4359 (47.4) | 3283 (47.2) | 1076 (48.2) | .402 |
| Lower SES: yes | 2904 (31.9) | 2273 (32.9) | 631 (28.4) | <.0001 |
| Region | ||||
| Continental | 1522 (16.5) | 1152 (16.5) | 370 (16.5) | <.0001 |
| Midwest | 2153 (23.4) | 1639 (23.5) | 514 (22.9) | |
| North Atlantic | 2029 (22.0) | 1455 (20.9) | 574 (25.6) | |
| Pacific | 1460 (15.9) | 1172 (16.8) | 288 (12.8) | |
| Southeast | 2049 (22.2) | 1552 (22.3) | 497 (22.2) | |
Note: Missing data: race (n = 399), ethnicity (n = 186), service connectedness (n = 24), and SES (n = 94). Lower SES is defined as being in VA priority group 5, which indicates either having a non–service‐connected disability or an annual income below the VA’s geographically adjusted income limit.
Abbreviations: OUD, opioid use disorder; PTSD, posttraumatic stress disorder; SES, socioeconomic status; VA, Veterans Affairs.
Other race categories included American Indian or Alaskan Native (0.67%), Asian (0.46%), and Native Hawaiian or Other Pacific Islander (0.63%).
Service connection indicates the veteran is receiving medical coverage for a disability that was caused or aggravated by their military service.
Outcome measures
We evaluated opioid use outcomes in the 13 months after the treatment period following a framework used in previous studies. 3 , 5 , 21 The outcome of CoBenzOp for a cumulative 30 days in the general adult population is identified by the Pharmacy Quality Alliance as a quality measure for medication safety. 22 NewPersOU was defined as receiving at least one opioid prescription in the 90–180 days after the treatment period (Figure 1). 5 We also determined the occurrence of opioid‐related adverse effects, including opioid use disorder, opioid overdose, and nonoverdose adverse events, via validated algorithms with International Classification of Diseases, Ninth and Tenth Revision codes (Table S2).
FIGURE 1.
Timeline for measurement of exposures and outcomes.
Analysis plan
We defined 13 monthly (30‐day) follow‐up intervals (days 1–390) starting the day after the treatment period (Figure 1). Veterans’ opioid prescription data from the CDW administrative database were captured in each 30‐day time interval until the veteran had a cancer recurrence or died, at which time they were censored from the analysis. To calculate CoBenzOp, we summed the days of coverage with concurrent prescriptions for opioid and benzodiazepine medications over the follow‐up period. To determine the incidence of NewPersOU, we identified veterans who had 1 or more days of a prescribed opioid medication in each 30‐day interval. Criteria for NewPersOU were met when the veteran was identified in interval 4, 5, or 6, which corresponded to 90–180 days in the posttreatment period (Figure 1).
We conducted χ 2 analysis and analysis of variance to analyze differences in CoBenzOp and NewPersOU outcomes in the cohort across quartiles of baseline opioid MMEs prescribed. We used a linear model to estimate the association between baseline opioid exposure and days of CoBenzOp in the follow‐up period. We then used a discrete time survival analysis to estimate the association between baseline opioid exposure and NewPersOU. Patients with missing data for any covariate were excluded from the analytic models.
To control for factors that could be associated with prescriptions for pain medication, we included the following in the analytic models: patient‐level sociodemographic factors, cancer type, stage, and treatment, baseline comorbidity and use of nonopioid pain medication, and region in which treatment was received. In the discrete time survival analysis, cases were censored because of recurrence or death in each of the 13 monthly intervals after the treatment period. In the analysis of NewPersOU, cases censored before 90 days were excluded from the analysis. For the linear model, we limited inclusion to those not censored in the 13‐month follow‐up period in the analysis. To characterize the dose and chronicity of those with NewPersOU, we examined the trajectory over time and the distribution of MME/day for each month in which a veteran had 1 or more days of prescribed opioids.
We conducted a prespecified sensitivity analysis defining NewPersOU as receipt of any opioid prescription in days 60–180 after the treatment period. We conducted a second prespecified sensitivity analysis defining opioid exposure in the treatment period as any opioid prescribed (oral, iv, or cartridge formulation) within 24 h of discharge.
RESULTS
Study cohort
The cohort included 9213 veterans. Of these, 9036 (98%) were male, 1687 (19%) were Black or African American, 6965 (79%) were White, and 62 (1.8%) were other races. Among the total cohort, 456 (5.1%) were of Hispanic ethnicity. The most common cancer types were prostate (n = 2594; 28%), colorectal (n = 2393; 26%), bladder (n = 2302; 25%), and lung (n = 1252; 4%). The remaining tumor types included other gastrointestinal, soft tissue, and head and neck cancers, with a total of 672 (7%). The distribution of cancer stages was as follows: stage 0, 1758 (19%); stage I, 2790 (30%); stage II, 3021 (33%); and stage III, 1644 (18%). Surgical therapy alone was received in 6495 veterans (71%), with the remaining having a surgical procedure in addition to adjuvant chemotherapy, immunotherapy, and/or radiation therapy (Table 1).
A majority of the cohort, 6364 (69%), were in the hospital on the day of their index surgical procedure. Among the total cohort, 6970 (76%) were prescribed opioids at some time in the baseline treatment period. Those prescribed opioids at some time in the treatment period differed from those who were not prescribed opioids with respect to cancer type, cancer stage at diagnosis, receipt of adjuvant treatment, baseline comorbidity, nonopioid pain medication use, SES, and region in which they received cancer treatment (Table 1). Among veterans prescribed any opioids in the treatment period, the median MME prescribed was 172.5 (interquartile range [IQR], 100–285).
Opioid use outcomes in the follow‐up period
Among the study cohort, CoBenzOp occurred in 366 persons (4.0%). The number of days of concurrent prescriptions of benzodiazepines and opioids in the follow‐up period among the total cohort ranged from 0 to 338, with a mean (SD) of 1.06 (9.2). The range of possible days was 0–390. The days of CoBenzOp (mean [SD]) in the follow‐up period increased with increased MMEs of opioids in the treatment period as follows: none, 0.48 (6.7); quartile 1 (Q1) (MME, >0 and ≤100), 1.15 (9.5); Q2 (MME, >100 and ≤172.5), 0.87 (6.4); Q3 (MME, >172.5 and ≤285), 0.84 (10.0); and Q4 (MME, >285), 2.1 (12.6) (p < .0001).
The incidence of NewPersOU among the total cohort was 981 (10.6%). Higher quartiles of MME opioid exposure in the baseline treatment period were associated with increased rates of NewPersOU as follows (number in the quartile [% of NewPersOU among the quartile]): no opioids prescribed, 174 (8%); Q1, 235 (13.5%); Q2, 160 (9.6%); Q3, 192 (11.4%); and Q4, 227 (13.2%) (p < .0001). The rates of opioid nonoverdose adverse events in the 13‐month follow‐up period were as follows: opioid use disorder, 72 (0.78%); nonoverdose adverse events, three (0.03%); and opioid overdose, 0.
Linear model of coprescription of benzodiazepine and opioid medications
In the linear model, veterans in the highest quartile of MME exposure in the treatment period versus no opioid exposure had more days of CoBenzOp in the 13‐month follow‐up period (mean, 0.99; 95% CI, 0.32–1.66). Lower SES, baseline nonopioid pain medication use, and a baseline history of anxiety, depression, or posttraumatic stress disorder (PTSD) were also associated with days of CoBenzOp (Table 2).
TABLE 2.
Linear model of unsafe coprescribing of benzodiazepine and opioid medications.
| Variable | Mean difference in number of days | 95% CI | p | |
|---|---|---|---|---|
| Lower | Upper | |||
| MME (reference, 0 [n = 2243]) | ||||
| >0 and ≤100 (n = 1831) | 0.45 | −0.20 | 1.11 | .18 |
| >100 and ≤172.5 (n = 1687) | 0.16 | −0.50 | 0.82 | .63 |
| >172.5 and ≤285 (n = 1701) | 0.18 | −0.48 | 0.84 | .59 |
| >285 (n = 1751) | 0.99 | 0.32 | 1.66 | <.01 |
| Age at diagnosis, years (reference, >65) | ||||
| <45 | −0.85 | −2.79 | 1.10 | .39 |
| 45–64 | 0.32 | −0.17 | 0.81 | .20 |
| Year of diagnosis (reference, 2016) | ||||
| 2014 | 0.44 | −3.30 | 4.15 | .82 |
| 2015 | 0.38 | −0.046 | 0.80 | .08 |
| Elixhauser comorbidity score (reference, <0) | ||||
| 0 | 0.38 | −0.20 | 0.96 | .20 |
| ≥1 and ≤3 | 0.16 | −0.61 | 0.93 | .69 |
| ≥4 | 0.16 | −0.49 | 0.81 | .62 |
| Race (reference, other) | ||||
| White | 0.02 | −1.57 | 1.60 | .98 |
| Black | −0.95 | −2.60 | 0.70 | .26 |
| Hispanic ethnicity (reference, non‐Hispanic) | 0.13 | −0.89 | 1.14 | .81 |
| Male sex (reference, female) | −1.19 | −2.83 | 0.46 | .16 |
| Lower SES (VA priority group 5) a | 0.71 | 0.14 | 1.28 | .02 |
| Adjuvant treatment (reference, surgery only) | ||||
| Surgery + radiation | 0.59 | −0.72 | 1.90 | .38 |
| Surgery + immunotherapy | −0.13 | −1.25 | 0.99 | .82 |
| Surgery + chemotherapy (±radiation and ±immunotherapy) | 0.66 | 0.01 | 1.30 | .05 |
| Service connection: yes b | −0.06 | −0.61 | 0.48 | .82 |
| Cancer type (reference, prostate) | ||||
| Bladder | −0.13 | −1.10 | 0.83 | .79 |
| Colorectal | 0.07 | −0.62 | 0.76 | .84 |
| Lung | 0.40 | −0.48 | 1.29 | .37 |
| Other | 0.36 | −0.65 | 1.36 | .49 |
| Cancer stage (reference, 0) | ||||
| I | 0.28 | −0.46 | 1.01 | .46 |
| II | 0.39 | −0.45 | 1.23 | .37 |
| III | 0.23 | −0.73 | 1.19 | .63 |
| Dual enrollment with Medicare Part D | 0.11 | −0.43 | 0.64 | .70 |
| Alcohol use disorder | −0.19 | −0.98 | 0.60 | .64 |
| Tobacco use | 0.01 | −0.54 | 0.57 | .96 |
| Chronic pain in year before diagnosis | 0.07 | −0.38 | 0.53 | .75 |
| Baseline nonopioid pain medication use | 1.26 | 0.80 | 1.72 | <.001 |
| Anxiety/depression/PTSD | 1.62 | 1.03 | 2.21 | <.001 |
| Bipolar/schizophrenia | 0.23 | −0.82 | 1.27 | .67 |
| Region (reference, Southeast) | ||||
| Continental | 0.04 | −0.65 | 0.72 | .92 |
| Midwest | −0.43 | −1.06 | 0.21 | .19 |
| North Atlantic | −0.11 | −0.74 | 0.53 | .75 |
| Pacific | −0.60 | −1.32 | 0.11 | .10 |
Note: MME reflects the total MME of opioids prescribed in the baseline treatment period.
Abbreviations: MME, morphine milligram equivalent; PTSD, posttraumatic stress disorder; SES, socioeconomic status; VA, Veterans Affairs.
Lower SES was indicated by having a VA priority group 5 level, which indicates VA pension benefits, Medicaid eligibility, or an annual income below the adjusted income limits based on zip code of residence.
Service connection indicates the veteran is receiving medical coverage for a disability that was caused or aggravated by their military service.
Discrete time survival analysis of NewPersOU
In the discrete time survival analysis, veterans in the highest quartile of MME exposure in the treatment period versus no opioid exposure had a higher risk of NewPersOU (hazard ratio [HR], 1.6; 95% CI, 1.2–1.9; p < .001). Treatment with adjuvant chemotherapy, with or without radiation, or immunotherapy increased the risk of NewPersOU (HR, 1.5; 95% CI, 1.2–1.8; p < .001). Additional factors associated with NewPersOU were a diagnosis of bladder, colorectal, lung, or other cancer type versus prostate cancer, stage I, II, or III compared to stage 0, age 45–64 years versus older, lower SES, baseline nonopioid pain medication use, and a baseline history of anxiety, depression, or PTSD (Table 3).
TABLE 3.
Discrete time survival analysis of new persistent opioid use.
| Variable | Hazard ratio | 95% CI | p | |
|---|---|---|---|---|
| Lower | Upper | |||
| MME (reference, 0) | ||||
| >0 and ≤100 | 1.33 | 1.08 | 1.65 | <.01 |
| >100 and ≤172.5 | 1.09 | 0.87 | 1.37 | .45 |
| >172.5 and ≤285 | 1.37 | 1.10 | 1.71 | <.01 |
| >285 | 1.55 | 1.25 | 1.93 | <.001 |
| Age at diagnosis, years (reference, >65) | ||||
| <45 | 0.91 | 0.47 | 1.75 | .77 |
| 45–64 | 1.18 | 1.02 | 1.38 | .03 |
| Year of diagnosis (reference, 2016) | ||||
| 2014 | 0.95 | 0.29 | 3.08 | .93 |
| 2015 | 1.12 | 0.98 | 1.28 | .11 |
| Elixhauser comorbidity score (reference, <0) | ||||
| 0 | 0.95 | 0.78 | 1.14 | .56 |
| ≥1 and ≤3 | 1.22 | 0.97 | 1.54 | .09 |
| ≥4 | 1.20 | 0.99 | 1.46 | .07 |
| Race (reference, other) | ||||
| White | 0.98 | 0.59 | 1.64 | .94 |
| Black | 0.90 | 0.53 | 1.55 | .71 |
| Hispanic ethnicity (reference, non‐Hispanic) | 0.80 | 0.56 | 1.15 | .23 |
| Male sex (reference, female) | 0.87 | 0.55 | 1.38 | .55 |
| Lower SES (VA priority group 5) a | 1.64 | 1.34 | 1.98 | <.001 |
| Adjuvant treatment (reference, surgery only) | ||||
| Surgery + radiation | 1.44 | 0.96 | 2.18 | .08 |
| Surgery + immunotherapy | 1.39 | 1.03 | 1.87 | .03 |
| Surgery + chemotherapy (±radiation and ±immunotherapy) | 1.49 | 1.25 | 1.77 | <.001 |
| Service connection: yes b | 1.32 | 1.10 | 1.59 | .004 |
| Cancer type (reference, prostate) | ||||
| Bladder | 3.57 | 2.66 | 4.80 | <.001 |
| Colorectal | 1.95 | 1.53 | 2.49 | <.001 |
| Lung | 1.93 | 1.45 | 2.58 | <.001 |
| Other | 2.57 | 1.90 | 3.47 | <.001 |
| Cancer stage (reference, 0) | ||||
| I | 1.48 | 1.18 | 1.85 | <.001 |
| II | 2.10 | 1.64 | 2.69 | <.001 |
| III | 2.03 | 1.52 | 2.72 | <.001 |
| Dual enrollment with Medicare Part D: yes | 0.87 | 0.73 | 1.04 | .12 |
| Alcohol use disorder: yes | 1.18 | 0.94 | 1.48 | .14 |
| Tobacco use: yes | 1.06 | 0.90 | 1.25 | .50 |
| Chronic pain in year before diagnosis: yes | 1.26 | 1.09 | 1.47 | .002 |
| Baseline nonopioid pain medication use: yes | 1.30 | 1.13 | 1.50 | <.001 |
| Anxiety/depression/PTSD: yes | 1.12 | 0.94 | 1.33 | .21 |
| Bipolar/schizophrenia: yes | 0.65 | 0.45 | 0.93 | .02 |
| Region (reference, Southeast) | ||||
| Continental | 1.10 | 0.90 | 1.36 | .36 |
| Midwest | 0.89 | 0.73 | 1.08 | .24 |
| North Atlantic | 0.87 | 0.71 | 1.06 | .17 |
| Pacific | 0.88 | 0.71 | 1.11 | .28 |
Abbreviations: MME, morphine milligram equivalent; PTSD, posttraumatic stress disorder; SES, socioeconomic status; VA, Veterans Affairs.
Lower SES was indicated by having a VA priority group 5 level, which indicates VA pension benefits, Medicaid eligibility, or an annual income below the adjusted income limits based on zip code of residence.
Service connection indicates the veteran is receiving medical coverage for a disability that was caused or aggravated by their military service.
Sensitivity analyses
Expanding the definition of NewPersOU to receiving at least one opioid prescription in days 60–180 after the treatment period increased the incidence of NewPersOU to 1261 (13.7%), with findings from the multivariable analysis similar to the primary analysis. Defining opioid exposure as the receipt of an opioid (oral, iv, or cartridge formulation) within 24 h of discharge versus not, in contrast to quartiles of MMEs in the treatment period, led to similar findings in our multivariable analyses of CoBenzOp and NewPersOU as in the primary analysis (Tables S3 and S4).
Description of dose in the survivorship period
The percentage of the cohort receiving prescription opioids declined over time. However, the MME/day of opioids prescribed among those with NewPersOU remained stable from month 1 to 12 (median, 20; IQR, 12–30; to median, 30; IQR, 15–40) (Figure 2).
FIGURE 2.
(A) The percentage of those with NewPersOU (at least one opioid prescription in the 90–180 days of the follow‐up period) who also received at least one opioid prescription in each of the 30‐day follow‐up time periods. Although the % receipt of opioid prescriptions declines, Panel B indicates that among those who continue to be prescribed opioids, the MME/day remains similar across this period of time. (B) Values indicate the MME of opioids prescribed per 30‐day period for all of those who were defined as having NewPersOU (at least one opioid prescription in a given 30‐day period). The values in the boxes indicate the median MME/day of opioids prescribed in each 30‐day period for this cohort of persons with NewPersOU. As seen above, this median ranges from 20 to 30 MME/day. Number of participants in each 30‐day interval: interval 1, n = 428; interval 2, n = 292; interval 3, n = 349; interval 4, n = 565; interval 5, n = 530; interval 6, n = 502; interval 7, n = 302; interval 8, n = 198; interval 9, n = 182; interval 10, n = 148; interval 11, n = 149; interval 12, n = 138; and interval 13, n = 117. MME indicates morphine milligram equivalent; NewPersOU, new persistent opioid use.
DISCUSSION
In this retrospective cohort study of an opioid‐naive cohort of US veterans with a new cancer diagnosis who underwent a surgical procedure with curative intent, 366 (4.0%) experienced CoBenzOp and 981 (10.6%) developed NewPersOU in the 13‐month follow‐up period. Consistent with our hypothesis, the intensity of exposure to opioid prescriptions attributed to treatment was positively associated with these outcomes. Persons undergoing treatment for cancer have specific needs for pain control, given multimodality treatments often including surgery, radiation, chemotherapy, and immunotherapy. 23 , 24 For those undergoing therapy with curative intent, there is the potential for a substantial impact of opioid‐related outcomes on quality of life. Minimizing opioid exposure associated with treatment while providing adequate pain control may decrease long‐term risks, including unsafe opioid coprescribing with benzodiazepines and LTOT.
The incidence of CoBenzOp in our cohort was 3.6% overall, and higher among those with lower SES, baseline use of nonopioid pain medications, and comorbidities of anxiety or depression. This incidence of CoBenzOp is lower than reported in two previous studies among cancer survivors that did not limit the cohorts to opioid‐naive persons. One study reported rates of CoBenzOp of 6.1% at 1 year. 9 A second study of a cohort treated for early‐stage breast cancer reported rates of CoBenzOp of 16% at 90 days after chemotherapy. 25 CoBenzOp is identified as a safety concern by the Pharmacy Quality Alliance, and CDC 2022 guidelines recommend caution when coprescribing these medications. However, both this statement and guideline excluded patients with cancer. 1 , 26 Although we report a relatively low risk of CoBenzOp in this opioid‐naive cancer cohort, the risk was higher among more vulnerable populations.
We report the incidence of NewPersOU in our cohort to be 10.6%. This finding is similar to the 10.4% incidence of NewPersOU reported by Lee et al. in a cancer cohort in the general US population. 5 Our finding indicates a higher incidence of NewPersOU among patients with cancer than the 5.9%–6.5% reported in a study of noncancer surgical populations undergoing both minor and major surgical procedures. 3 As hypothesized, a higher intensity of exposure to opioid medications in the treatment period was associated with a higher risk of NewPersOU in a multivariable analysis that controlled for patient, clinical, treatment, and geographic factors. This association is consistent with previous studies in both non‐VA and VA cancer populations. 13 , 21 Together, these studies emphasize the importance of limiting opioid exposure associated with treatment to the degree possible while providing adequate pain control. 25
Additional factors associated with NewPersOU in our multivariable models included cancer type and treatment, most notably among patients with bladder cancer and those who received adjuvant chemotherapy. These findings are considered exploratory but suggest that future studies may identify specific cancer populations at risk for these outcomes.
We described the trajectory of continued opioid use in our cohort over the 13‐month period after treatment. Although the proportion of survivors continuing on opioids after the 90‐ to 180‐day period declined, the dose of opioids prescribed remained stable, similar to other studies of cancer survivors who were maintained on LTOT. 5 , 27
Notably, 24% of patients in our cohort did not receive any opioids in the treatment period, which raises important questions regarding hospital‐level practice variation (e.g., in the use of opioid‐sparing pathways) and differences in the attributes of patients treated with and without opioids. Our findings show that patients who received opioids were more likely to be younger, be diagnosed with lung cancer, have a more advanced stage of cancer, have baseline pain medication use, have lower SES, and to differ geographically compared with patients not treated with opioids. In the absence of patient‐level data on pain control, it is difficult to ascertain the quality of pain management among patients with and without opioids. Nonetheless, this variation indicates an opportunity for the identification and development of interventions to optimize pain management associated with treatment and in the first year of survivorship via system‐, provider‐, and patient‐level interventions.
American Society of Clinical Oncology guidelines emphasize that opioids be offered to patients with moderate to severe pain related to cancer or active cancer treatment unless contraindicated, initiated at the lowest dose providing acceptable treatment, and with early assessment and frequent titration. 8 However, guidelines regarding safe opioid prescribing practices, including the 2022 CDC clinical practice guidelines for prescribing opioids for pain 1 and the Department of Veterans Affairs, 2 exclude patients with cancer. Although patients with cancer have specific needs for pain control, there may be benefits to extending general guidelines regarding prescription opioids to those facing cancer treatment in consideration of health and well‐being during the years of survivorship. Postoperative opioid prescribing guidelines have been effective in reducing opioid MMEs, days covered, and quantity of pills prescribed for noncancer procedures. 28 Education and opioid prescribing guidelines for patients undergoing breast cancer and melanoma procedures have been effective in decreasing the quantity of opioids prescribed at discharge without an increase in the need for refills. 29 In addition, intraoperative modifications for pain management may improve pain outcomes, and could be better explored in cancer surgery. 30 , 31 , 32 , 33 Finally, psychosocial and behavioral interventions for pain management can be effective as part of pain management strategies for patients with cancer. 34 , 35 , 36 , 37 , 38 , 39 Our study and others can alert providers to risk factors for NewPersOU in this population that can inform decision‐making for pain management.
Our study has limitations. Our analysis was conducted in a VA population, and did not analyze opioid prescriptions obtained outside of the VA system. Our findings may therefore underestimate the opioid prescriptions received in the pretreatment, baseline, or follow‐up periods if paid for by Medicare Part D, commercial insurance, or Medicaid. However, dual enrollment in the VA and Medicare Part D was not associated with opioid use outcomes in our analysis. In addition, our analysis was based on filled opioid prescriptions, and actual patient consumption of opioids was unknown. The regional variation noted in the prescription of any opioids in our cohort suggests areas for future study on variations in practice regarding pain management in this population, including the use of opioid‐sparing pathways. 40 Because the majority of outpatient prescriptions for methadone are not captured within the VA CDW database, we could not assess whether persons were prescribed methadone as an outpatient in the year before the index surgery or in the follow‐up period. Evaluation of the providers who were prescribing the opioids was outside the scope of our study but offers an important area of future study as we design provider‐facing interventions to optimize pain management in this group of veterans. Our study was also limited in not having details of the treatment received, such as the total dose of radiation provided. Further work may be warranted to evaluate the association of the intensity of cancer therapy with these outcomes. Our data reflect outpatient prescriptions and pharmacy refills and does not include patient self‐reported use of as needed medications, which may underestimate the outcomes of interest. Finally, our study did not include patients with breast cancer, which limits the generalizability of our findings. However, our estimates were similar to prior studies among breast cancer cohorts. 21
In conclusion, we found that approximately one of 10 veterans who were opioid naive before the diagnosis of early‐stage cancer and underwent a definitive surgical procedure developed NewPersOU, and one of 25 veterans received a CoBenzOp in the first year of cancer survivorship. Those with a prior history of chronic pain, greater comorbidities, and lower SES and who received adjuvant chemotherapy were at an even higher risk. Although a cancer diagnosis, treatment, and associated pain syndromes will require specific pain management strategies, efforts should be taken to mitigate long‐term opioid use and its potential adverse effects in this population. This is especially true because cancer is increasingly recognized as a chronic condition, and survivors of cancer are living longer. Both system‐level changes that involve preoperative evaluation planning as well as increased knowledge, awareness, and education among providers and patients about the risk of long‐term opioid use can guide strategies for effective and safe pain management.
AUTHOR CONTRIBUTIONS
Marilyn M. Schapira: Conceptualization; investigation; funding acquisition; writing—original draft; methodology; writing—review and editing; formal analysis; supervision; resources; and project administration. Sumedha Chhatre: Formal analysis; writing—review and editing; and investigation. Patience M. Dow: Conceptualization; formal analysis. Charles E. Leonard: Conceptualization; investigation; methodology; and writing—review and editing. Peter Groeneveld: Conceptualization and writing—review and editing. Jason M. Prigge: Writing—review and editing and project administration. Christopher Roberts: Conceptualization; visualization; formal analysis; and data curation. Zachary F. Meisel: Conceptualization and writing—review and editing. Ravi B. Parikh: Conceptualization and writing—review and editing. Ravishankar Jayadevappa: Conceptualization. Emily C. Paulson: Conceptualization and writing—review and editing. Robert S. Krouse: Conceptualization and writing—review and editing. Katie J. Suda: Conceptualization and writing—review and editing. Pallavi Kumar: Conceptualization and writing—review and editing. Visala Muluk: Conceptualization and writing—review and editing. Rebecca A. Hubbard: Conceptualization; formal analysis; writing—review and editing; and supervision.
CONFLICT OF INTEREST STATEMENT
Charles E. Leonard reports consulting for Moderna and TriNetX. Ravi B. Parikh reports consulting for Genetic Chemistry, Thyme Care, Biofourmis, Onc.AI, Credit Suisse, Main Street Health, ConcertAI, Medscape, and G1 Therapeutics. The other authors declare no conflicts of interest.
Supporting information
Supplementary Material
ACKNOWLEDGMENTS
This study was funded by Department of Veterans Affairs Health Services Research & Development Grant 1121HX003273‐0.
Schapira MM, Chhatre S, Dow PM, et al. The impact of opioid use associated with curative‐intent cancer surgery on safe opioid prescribing practice among veterans: an observational study. Cancer. 2025;e70009. doi: 10.1002/cncr.70009
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available because of privacy or ethical restrictions.
REFERENCES
- 1. Dowell D, Ragan KR, Jones CM, Baldwin GT, Chou R. CDC clinical practice guideline for prescribing opioids for pain—United States, 2022. MMWR Recomm Rep. 2022;71(3):1‐95. doi: 10.15585/mmwr.rr7103a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Sandbrink F, Murphy JL, Johansson M, et al. The use of opioids in the management of chronic pain: synopsis of the 2022 updated U.S. Department of Veterans Affairs and U.S. Department of Defense clinical practice guideline. Ann Intern Med. 2023;176(3):388‐397. doi: 10.7326/M22-2917 [DOI] [PubMed] [Google Scholar]
- 3. Brummett CM, Waljee JF, Goesling J, et al. New persistent opioid use after minor and major surgical procedures in US adults. JAMA Surg. 2017;152(6):e170504. doi: 10.1001/jamasurg.2017.0504 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Santosa KB, Hu HM, Brummett CM, et al. New persistent opioid use among older patients following surgery: a Medicare claims analysis. Surgery. 2020;167(4):732‐742. doi: 10.1016/j.surg.2019.04.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Lee JS, Hu HM, Edelman AL, et al. New persistent opioid use among patients with cancer after curative‐intent surgery. J Clin Oncol. 2017;35(36):4042‐4049. doi: 10.1200/JCO.2017.74.1363 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Jones KF, Fu MR, Merlin JS, et al. Exploring factors associated with long‐term opioid therapy in cancer survivors: an integrative review. J Pain Symptom Manage. 2021;61(2):395‐415. doi: 10.1016/j.jpainsymman.2020.08.015 [DOI] [PubMed] [Google Scholar]
- 7. Check DK, Avecilla RAV, Mills C, et al. Opioid prescribing and use among cancer survivors: a mapping review of observational and intervention studies. J Pain Symptom Manage. 2022;63(4):e397‐e417. doi: 10.1016/j.jpainsymman.2021.10.015 [DOI] [PubMed] [Google Scholar]
- 8. Paice JA, Bohlke K, Barton D, et al. Use of opioids for adults with pain from cancer or cancer treatment: ASCO guideline. J Clin Oncol. 2023;41(4):914‐930. doi: 10.1200/JCO.22.02198 [DOI] [PubMed] [Google Scholar]
- 9. Murphy CC, Fullington HM, Alvarez CA, et al. Polypharmacy and patterns of prescription medication use among cancer survivors. Cancer. 2018;124(13):2850‐2857. doi: 10.1002/cncr.31389 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Zullig LL, Jackson GL, Dorn RA, et al. Cancer incidence among patients of the U.S. Veterans Affairs health care system. Mil Med. 2012;177(6):693‐701. doi: 10.7205/milmed-d-11-00434 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Hoggatt KJ, Harris AHS, Washington DL, Williams EC. Prevalence of substance use and substance‐related disorders among US Veterans Health Administration patients. Drug Alcohol Depend. 2021;225:108791. doi: 10.1016/j.drugalcdep.2021.108791 [DOI] [PubMed] [Google Scholar]
- 12. Agha Z, Lofgren RP, VanRuiswyk JV, Layde PM. Are patients at Veterans Affairs medical centers sicker? A comparative analysis of health status and medical resource use. Arch Intern Med. 2000;160(21):3252‐3257. doi: 10.1001/archinte.160.21.3252 [DOI] [PubMed] [Google Scholar]
- 13. Vitzthum LK, Riviere P, Sheridan P, et al. Predicting persistent opioid use, abuse, and toxicity among cancer survivors. J Natl Cancer Inst. 2020;112(7):720‐727. doi: 10.1093/jnci/djz200 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Arnold J, Zhao X, Cashy JP, et al. An interrupted time‐series evaluation of the association between state laws mandating prescriber use of prescription drug monitoring programs and discontinuation of chronic opioid therapy in US veterans. Med Care. 2021;59(12):1042‐1050. doi: 10.1097/MLR.0000000000001643 [DOI] [PubMed] [Google Scholar]
- 15. Cordasco KM, Moreau JL, Chawla N. Coordinating care for reproductive health malignancies in the Veterans' Health Administration: promising practices, ongoing challenges, and future research. Semin Reprod Med. 2019;37(1):32‐42. doi: 10.1055/s-0039-1692398 [DOI] [PubMed] [Google Scholar]
- 16. Muzic JG, Schmitt AR, Wright AC, et al. Incidence and trends of basal cell carcinoma and cutaneous squamous cell carcinoma: a population‐based study in Olmsted County, Minnesota, 2000 to 2010. Mayo Clin Proc. 2017;92(6):890‐898. doi: 10.1016/j.mayocp.2017.02.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Zullig LL, Jazowski SA, Chawla N, et al. Summary of Veterans Health Administration cancer data sources. J Registry Manag. 2019;46(3):76‐83. [PubMed] [Google Scholar]
- 18. Opioid National Drug Code and Oral MME Conversion File Update. Centers for Disease Control and Prevention. Accessed July 23, 2025. https://archive.cdc.gov/#/results?q=https://www.cdc.gov/opioids/data‐resources/index.html&start=0&rows=10 [Google Scholar]
- 19. Thorpe JM, Thorpe CT, Schleiden L, et al. Association between dual use of Department of Veterans Affairs and Medicare Part D drug benefits and potentially unsafe prescribing. JAMA Intern Med. 2019;179(11):1584‐1586. doi: 10.1001/jamainternmed.2019.2788 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Moyo P, Zhao X, Thorpe CT, et al. Dual receipt of prescription opioids from the Department of Veterans Affairs and Medicare Part D and prescription opioid overdose death among veterans: a nested case‐control study. Ann Intern Med. 2019;170(7):433‐442. doi: 10.7326/M18-2574 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Roberts AW, Fergestrom N, Neuner JM, Winn AN. New‐onset persistent opioid use following breast cancer treatment in older adult women. Cancer. 2020;126(4):814‐822. doi: 10.1002/cncr.32593 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Pharmacy Quality Alliance: Advancing the Quality of Medication Use. Pharmacy Quality Alliance. Accessed July 23, 2025. https://www.pqaalliance.org/assets/Measures/PQA_ [Google Scholar]
- 23. Swarm RA, Paice JA, Anghelescu DL, et al. NCCN clinical practice guidelines in oncology adult cancer pain, version 3.2019. J Natl Compr Canc Netw. 2019;17(8). [DOI] [PubMed] [Google Scholar]
- 24. Shah A, Hoffman EM, Mauermann ML, et al. Incidence and disease burden of chemotherapy‐induced peripheral neuropathy in a population‐based cohort. J Neurol Neurosurg Psychiatry. 2018;89(6):636‐641. doi: 10.1136/jnnp-2017-317215 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Yau WH, Roeland EJ, Revta C, Ale‐Ali A, Ma JD. Opioid and benzodiazepine relative use in breast cancer patients before, during, and after curative chemotherapy. J Pain Palliat Care Pharmacother. 2018;32(1):27‐29. doi: 10.1080/15360288.2018.1479328 [DOI] [PubMed] [Google Scholar]
- 26. PQA Opioid Measure Set. Pharmacy Quality Alliance. Accessed July 23, 2025. https://www.pqaalliance.org/assets/Measures/PQA_Opioid_Core_Measure_Set_Description.pdf [Google Scholar]
- 27. Chen L, Chubak J, Yu O, et al. Changes in use of opioid therapy after colon cancer diagnosis: a population‐based study. Cancer Causes Control. 2019;30(12):1341‐1350. doi: 10.1007/s10552-019-01236-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Hlavin C, Muluk S, Muluk V, Ryan J, Wagner J, Dhupar R. A program to reduce post‐operative opioid prescribing at a Veteran's Affairs hospital. J Clin Med. 2022;11(18):5453. doi: 10.3390/jcm11185453 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Lee JS, Howard RA, Klueh MP, et al. The impact of education and prescribing guidelines on opioid prescribing for breast and melanoma procedures. Ann Surg Oncol. 2019;26(1):17‐24. doi: 10.1245/s10434-018-6772-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Doan LV, Li A, Brake L, et al. Single‐dose of postoperative ketamine for postoperative pain after mastectomy: a pilot randomized controlled trial. J Pain Res. 2023;16:881‐892. doi: 10.2147/JPR.S389564 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Othman AH, El‐Rahman AMA, El Sherif F. Efficacy and safety of ketamine added to local anesthetic in modified pectoral block for management of postoperative pain in patients undergoing modified radical mastectomy. Pain Physician. 2016;19(7):485‐494. [PubMed] [Google Scholar]
- 32. Machado FC, Vieira JE, de Orange FA, Ashmawi HA. Intraoperative methadone reduces pain and opioid consumption in acute postoperative pain: a systematic review and meta‐analysis. Anesth Analg. 2019;129(6):1723‐1732. doi: 10.1213/ANE.0000000000004404 [DOI] [PubMed] [Google Scholar]
- 33. Kumar K, Kirksey MA, Duong S, Wu CL. A review of opioid‐sparing modalities in perioperative pain management: methods to decrease opioid use postoperatively. Anesth Analg. 2017;125(5):1749‐1760. doi: 10.1213/ANE.0000000000002497 [DOI] [PubMed] [Google Scholar]
- 34. Sheinfeld Gorin S, Krebs P, Badr H, et al. Meta‐analysis of psychosocial interventions to reduce pain in patients with cancer. J Clin Oncol. 2012;30(5):539‐547. doi: 10.1200/JCO.2011.37.0437 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Somers TJ, Winger JG, Fisher HM, et al. Behavioral cancer pain intervention dosing: results of a sequential multiple assignment randomized trial. Pain. 2023;164(9):1935‐1941. doi: 10.1097/j.pain.0000000000002915 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Kelleher SA, Winger JG, Fisher HM, et al. Behavioral cancer pain intervention using videoconferencing and a mobile application for medically underserved patients: rationale, design, and methods of a prospective multisite randomized controlled trial. Contemp Clin Trials. 2021;102:106287. doi: 10.1016/j.cct.2021.106287 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Darnall BD, Ziadni MS, Krishnamurthy P, et al. “My surgical success”: effect of a digital behavioral pain medicine intervention on time to opioid cessation after breast cancer surgery—a pilot randomized controlled clinical trial. Pain Med. 2019;20(11):2228‐2237. doi: 10.1093/pm/pnz094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Azizoddin DR, Knoerl R, Adam R, et al. Cancer pain self‐management in the context of a national opioid epidemic: experiences of patients with advanced cancer using opioids. Cancer. 2021;127(17):3239‐3245. doi: 10.1002/cncr.33532 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Azizoddin DR, Wilson JM, Flowers KM, et al. Daily pain and opioid administration in hospitalized patients with cancer: the importance of psychological factors, recent surgery, and current opioid use. Pain. 2023;164(8):1820‐1827. doi: 10.1097/j.pain.0000000000002880 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Lund BC, Ohl ME, Hadlandsmyth K, Mosher HJ. Regional and rural‐urban variation in opioid prescribing in the Veterans Health Administration. Mil Med. 2019;184(11‐12):894‐900. doi: 10.1093/milmed/usz104 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available because of privacy or ethical restrictions.