Prediction model shows potential for identifying cancer survivors at risk for opioid misuse
Click Here to Manage Email Alerts
A novel predictive model appeared to identify cancer survivors at increased risk for opioid misuse, according to study results published in Journal of the National Cancer Institute.
“ASCO recommends a risk-stratified approach to adherence monitoring and prevention [of opioid misuse], but there are limited data to estimate the risk for adverse opioid events in patients with cancer,” Lucas K. Vitzthum MD, physician in the department of radiation medicine and applied sciences at University of California, San Diego, told Healio. “This project was prompted by a desire to better understand the risk for persistent opioid use, abuse and toxicity among cancer survivors and to develop a predictive model to guide opioid management.”
For the study, Vitzthum and colleagues used the nationwide Veterans Affairs Informatics and Computing Infrastructure database to identify 106,732 cancer survivors diagnosed with bladder, breast, colon, esophageal, stomach, head and neck, kidney, liver, lung, pancreas or prostate cancer between 2000 and 2015. Patients had survived without recurrence 2 years after treatment initiation.
Researchers assessed rates of persistent posttreatment opioid use, diagnoses of opioid abuse or dependence and inpatient admissions for opioid toxicity. Persistent opioid use consisted of having filled at least 120 days’ supply or 10 or more opioid prescriptions from 1 to 2 years after initiation of curative treatment.
Investigators used multivariable logistic regression models to identify patient, cancer and treatment risk factors associated with adverse opioid-related outcomes. They also developed and validated predictive risk models using a least absolute shrinkage and selection operator (LASSO) regression technique.
Results showed an overall rate of persistent opioid use among cancer survivors of 8.3% (95% CI, 8.1-8.4), with rates ranging from 5.3% among prostate cancer survivors to 19.8% among liver cancer survivors.
Posttreatment diagnoses of opioid abuse or dependence occurred at a rate of 2.9% (95% CI, 2.8-3), with opioid-related admissions occurring at a rate of 2.1% (95% CI, 2-2.2).
Results of multivariable analysis showed several factors associated with increased odds of persistent opioid use, including younger age, white race, unemployment at the time of cancer diagnosis, lower median income, increased comorbidity, and current or prior tobacco use.
Those previously diagnosed with alcohol, nonopioid drug or opioid misuse, or depression demonstrated higher odds of persistent opioid use. Previous chronic and intermittent opioid use appeared associated with substantially higher odds of persistent use.
Among survivors who had not previously used opioids, those who had not received an opioid prescription during the diagnostic or treatment period had a lower risk for persistent opioid use than those who had a prescription. In addition, compared with prostate cancer survivors, those who had been diagnosed with bladder, breast, esophageal, stomach, head and neck, liver, lung and pancreas cancer had higher odds of persistent opioid use.
Researchers observed a high level of discrimination among predictive models when identifying individuals at risk for adverse opioid-associated outcomes, including persistent opioid use (area under the curve [AUC] = 0.85), future diagnoses of opioid abuse or dependence (AUC = 0.87), and inpatient admission for opioid abuse or toxicity (AUC = 0.78).
“We developed a predictive model, which can be found online at www.canceropioidrisk.org, that may help stratify patients by their risk for adverse opioid outcomes and guide adherence monitoring,” Vitzthum told Healio. “Most patients with cancer have a low-risk for adverse opioid events. Yet, there is a subset of high-risk patients that can benefit from more intensive risk-mitigation techniques.”
The predictive model developed in this study requires prospective validation in future research, according to Vitzthum.
“We are interested in studying whether behavioral questions used in previously developed screening tools would improve the accuracy of our risk model. It is also of significant interest how the prescribing patterns of individual providers are associated with patients’ risk for adverse opioid outcomes,” Vitzthum said. – by Jennifer Southall
For more information:
Lucas K. Vitzthum MD, can be reached at University of California, San Diego, Altman Clinical and Translational Research Institute Building, 9452 Medical Center Drive, La Jolla, CA 92037; email: lvitzthum@ucsd.edu.
Disclosures: The ASCO Conquer Cancer Foundation Young Investigator Award funded the study. Vitzthum reports no relevant financial disclosures. Please see the study for all other authors’ relevant financial disclosures.
Perspective
Back to Top
Jai N. Patel, PharmD, BCOP, CPP
Opioids are and will remain the mainstay treatment for cancer-related pain for years to come. With the current opioid epidemic, it is imperative that providers implement strategies to reduce the risk for opioid abuse and ensure appropriate access to opioids for those who need them.
Unfortunately, there are limited resources to screen for opioid abuse risk. Universal precautions such as drug screening, pain contracts and reviewing prescription drug monitoring databases are effective strategies to reduce opioid misuse. Nonetheless, objective screening tools validated in the cancer population are much needed, especially for those receiving curative treatment.
This large retrospective analysis of cancer survivors identified several clinical factors associated with persistent opioid use, many of which have been identified in prior studies, such as tobacco use, younger age, lower median income and history of opioid use. Risk scores using predictive modeling highly discriminated between low-, intermediate- and high-risk patients. Although less than 5% of patients were considered high risk, about three-quarters of these patients were persistent opioid users.
It is critical to preemptively identify these high-risk patients and implement risk-mitigation strategies early in the pain management course. Importantly, screening tools must strike the right balance between complexity, predictive value and usability in busy clinics. Clinical factors identified in this study and that have been previously reported should be used routinely as red flags to identify high-risk patients. These patients should not be denied opioids, but rather should be monitored closely using prescription drug-monitoring databases, receive focused education on risks, undergo random drug tests and receive other risk-reducing measures. A concerted effort should be made to externally validate the study’s risk score prospectively. Ideally, a simplified and automated version of the risk tool could be applied as decision support to assist with opioid prescribing and real-time monitoring in clinical practice.
Levine Cancer Institute at Atrium Health
Click Here to Manage Email Alerts