Pain is highly prevalent in Veterans and one of the most common reasons they seek medical care. Increasingly, schedule II and III opioids are a main treatment modality for chronic pain with duration of therapy lasting potentially for years. Despite increasing rates of opioid prescribing, evidence supporting its use for chronic pain is modest and serious safety and addiction issues appear to be increasing. Regarding unsafe medication combinations, benzodiazepines potentiate the sedative properties of opioids and are listed as co-ingestions in a large proportion of opioid overdose deaths. Given these potential harms, experts have called for restraint in opioid use, targeting deimplementation of high-dose therapy, avoidance of opioid-benzodiazepine co-prescribing and promotion of non-pharmacological treatment, to which VHA has responded with a proactive, multi-pronged approach.
VHA has implemented a multi-pronged approach to proactively address this burgeoning public health crisis, including publication of safety-oriented prescriber guidelines. Most recently, VHA has implemented a comprehensive Opioid Safety Initiative, a multifaceted program that, among other components, requires facilities to contribute controlled substance prescribing data to state prescription monitoring programs (PMPs). PMPs are provider-searchable databases that contain prescription data, listed by patient, of all controlled substance prescriptions filled in a given state.
Outside-of-VHA access to opioid therapy for Veterans may expand as the VHA expands coverage of care in the community through programs such as the Veterans Access, Choice and Accountability Act of 2014 (VACAA), which covers payment for qualifying Veterans' visits to authorized private providers and the accompanying prescriptions written. The purpose of this study is to examine whether multiple sources of prescription payment - as a surrogate marker for non-VHA sources of care - are associated with receipt of risky opioid therapy.
We performed a cross-sectional study of Veterans with VHA-paid opioid prescriptions from Oct. 1, 2014 through Sep. 30, 2015 to examine whether having additional sources of payment was associated with risky opioid therapy. This study was part of a quality improvement project and was determined not to be human subjects research by the VA Connecticut Healthcare System Investigational Review Board (IRB) and the Kentucky Cabinet for Health and Family Services' IRB.
We examined prescription data from the Kentucky All Schedule Prescription Electronic Reporting (KASPER) system, a PMP that maintains a data repository containing details of all controlled substance prescriptions dispensed in Kentucky's retail and VHA pharmacies. Kentucky is the 26th most populated state, contains a sizeable rural population, a higher than average proportion of Veteran residents, and has higher than average non-VHA opioid prescribing rates.
Cohort inclusion criteria:
We included all Veterans with at least one VHA-paid opioid prescription during the specified time period. Because individuals may have entered the VHA system during the study period, and thus had a period of non-VHA prescriptions followed by a period of VHA-prescriptions, we started individuals' observation periods on the date of their first VHA-paid opioid prescription. "VHA-paid" was determined as follows: KASPER contains a "source of payment" field, in which "military/VHA" is a category. To ascertain whether the prescription was VHA-paid rather than military, we examined the number of military/VHA prescriptions before and after the date when VHA pharmacies began contributing data to KASPER. Since there was a marked spike in military/VHA source prescriptions (~5000%), and negligible numbers of military/VHA prescriptions before that date, we assumed all military/VHA source prescriptions after that date to be VHA.
Independent variable of interest: source of payment
We divided the study sample into three categories: 1) those for whom the only source of payment was VHA ("sole source"); 2) those for whom sources of payment were VHA and at least one cash payment ("VHA + cash payment(s)") whether or not there was a third source of payment; and 3) those who had at least one other non-cash source: Medicare, Medicaid, or private insurance ("VHA + non-cash source(s)").
Outcomes of interest: Two types of risky opioid therapy
We examined two types of risky opioid therapy at the patient level: combination opioid/benzodiazepine therapy and high-dose opioid therapy, defined as morphine equivalent daily dose (MEDD) 90 mg, consistent with current guidelines. For assessing combination therapy, each prescription's dates were calculated from the start date plus number of days supplied. Any Veteran with overlapping dates of an opioid with a benzodiazepine prescription was considered to have combination therapy. Regarding opioid dose, we converted all oral or transdermal opioid prescriptions to mgs MEDD by multiplying number of dosing units by dose prescribed by CDC 2015 conversion factors and dividing by days' supplied; for individuals who had > 1 prescription on any day, mgs MEDD of each prescription were added together. As our population of interest was Veterans receiving opioid pain treatment, we excluded dispensing of liquid methadone and buprenorphine products, both treatments for opioid use disorder.
We examined sex and age at the time of an individual's index VHA-paid opioid prescription, according to KASPER records.
We performed bivariate analyses, using chi-square tests to compare the difference in covariates between payment source categories. We then examined the bivariate associations between each variable of interest and the two risky opioid outcomes. Finally, we performed two logistic regression models, controlling for sex and age, to examine these outcomes (combination therapy, high-dose opioids).
Of the 14,795 individuals in the analytic sample, there were 81.9% in the sole source category, 6.6% in the VHA + cash payment(s) category, and 11.5% in the VHA + non-cash source(s) category. The sole source category had a higher proportion of men as well as a higher proportion of older individuals than did the other two payment groups (both p < .0001). Women had higher prevalence of combination opioid/benzodiazepine therapy (chi-square statistic = 50.67, df = 1, p < 0.0001), while those aged 80+ years had lower prevalence of high-dose therapy (chi-square statistic = 7.08, df =1, p < 0.0001). Sole source patients' prevalence of combination opioid/benzodiazepine therapy was 13% compared to 42% prevalence among patients with VHA + cash payment and 33% for patients with VHA + non-cash source(s) (overall chi-square statistic = 850.27, df = 2, p < 0.0001). In terms of high-dose therapy, the prevalence among sole source participants was 6.1% compared to 12.5% among patients with VHA + cash payment(s) and 8.7% for patients with VHA+ non-cash source(s) (overall chi-square statistic = 67.71, df = 2, p < 0.0001).
In logistic regression, controlling for age and gender, persons with multiple payment sources had significantly higher odds of combination opioid/benzodiazepine therapy than those with only VHA payments (VHA + cash compared to sole source: OR=4.75; 95% CI: 4.14-5.46; VHA + non-cash payment compare to sole source: OR=3.25; 95% CI: 2.89-3.64). Comparing the two multiple payment groupings, VHA + cash had significantly higher odds of having combination therapy than the VHA + non-cash group (OR=1.46; 95% CI: 1.25-1.72). Female gender and age group 55-64 were associated with significantly greater odds of being prescribed combination therapy. With respect to high-dose opioid therapy, both multiple payment source groups were significantly more likely to be prescribed high-dose opioids than the single source group (OR=2.25 for VHA + cash and OR=1.51 for VA + non-cash respectively). With respect to the two types of multiple payment(s) groups, the VHA + cash payment group had significantly higher odds of high dose opioids than the VHA + non-cash source(s) (OR: 1.49; 95% CI: 1.15-1.92). Female Veterans were less likely to receive high-dose opioids. All younger persons were significantly more likely to receive high-dose opioids than those aged 80+ (range of ORs = 2.05-3.62, all statistically significant).
We described the results of an evaluation where we merged patient-level state PDMP data with VHA EHR data to demonstrate that there is significant dual use of VHA and non-VHA sources of controlled substance prescribing. This has important implications for ongoing quality improvement efforts, including mandated checking of the state PDMP and outreach efforts to non-VHA prescribers to improve their appreciation of this critical quality of care issue.
- Gaither JR, Goulet JL, Becker WC, Crystal S, Edelman EJ, Gordon K, Kerns RD, Rimland D, Skanderson M, Justice AC, Fiellin DA. The Association Between Receipt of Guideline-Concordant Long-Term Opioid Therapy and All-Cause Mortality. Journal of general internal medicine. 2016 May 1; 31(5):492-501.