Increased Nonopioid Chronic Pain Treatment in the Veterans Health Administration, 2010–2016

Expert guidelines recommend prioritizing nonopioid treatments and avoiding initiation of long-term opioid therapy for the management of chronic pain. The Veterans Health Administration (VHA) has implemented multiple system-level programs to support guideline-concordant chronic pain care. In this Commentary, we describe trends in pain-related treatment utilization from 2010 to 2016 among veterans with incident chronic pain in the VHA. From 2010 to 2016, utilization of seven of nine nonopioid treatment modalities increased. The proportion of veterans utilizing multiple modalities also increased during the study period. Future work should support patient-centered integration of these modalities and assess the impact of multimodal care on patient outcomes. Chronic pain is common and disproportionately impacts veterans. An estimated 66% of veterans report pain, and 9% report severe pain [1]. Expert guidelines recommend integrated, multimodal, multidisciplinary chronic pain care, individualized to the specific needs of the patient [2–4]. Treatment modalities include medications and nonpharmacologic therapies such as psychological/behavioral, movement, and complementary and integrative approaches. In the VHA, the delivery of multimodal chronic pain care is guided by the stepped care model for pain management. The stepped care model calls for population-based screening, assessment, and management of chronic pain using the full range of low-intensity interventions delivered in primary care settings with timely access to pain specialty teams providing more intensive treatments targeted to individuals with more complex chronic pain, in collaboration with the primary care team [5]. Since 2003, guidelines from the Departments of Veterans Affairs and Defense have emphasized nonopioid treatments as firstline for chronic pain [6]. In recent years, there has been increasing scrutiny of opioid medication prescribing, given inadequate evidence of long-term benefit and growing evidence of harms [7]. In 2013, the VHA began implementing its Opioid Safety Initiative, a system-wide program to improve the safety of opioid prescribing, which has reduced high-risk medication prescribing in the VHA [8,9]. Meanwhile, chronic pain remains an important challenge for the VHA, which cares for an aging population and a growing number of veterans with pain related to recent military service. Little is known about recent changes in utilization of nonopioid chronic pain treatment modalities. To identify veterans receiving primary care in the VHA with incident chronic pain between January 1, 2010, and December 31, 2015, we adapted a published algorithm developed to identify patients with chronic pain using electronic health record data (Supplementary Data) [10]. We defined chronic pain as at least one of three criteria: 1) persistent moderate to severe pain, 2) diagnostic codes “likely to represent chronic pain,” or 3) receipt of long-term opioid therapy. We defined persistent moderate to severe pain as ≥4 using a 0 (no pain) to 10 (worst pain imaginable) numeric rating scale at outpatient encounters within a single year with ≥30 days between ratings. Second, we identified encounters using International Classification of Diseases, 9th Revision (ICD-9), diagnosis codes “likely to represent chronic pain.” Individuals with two or more outpatient encounters within a two-year period or a single inpatient encounter with a chronic pain diagnosis met this criterion. We mapped ICD-9 codes to ICD-10 codes using the General Equivalence Mappings tool created by the Centers for Medicare and Medicaid Services [11]. Third, we defined long-term opioid therapy as receipt of ≥90 days of an opioid medication within a 365-day period. The chronic pain index date was defined as the earliest date on which any of these criteria was met. To identify incident chronic pain, we excluded individuals who met chronic pain criteria during a two-year washout period from January 1, 2008, to December 31, 2009 (i.e., “prevalent” chronic pain) (Figure 1). We also excluded individuals with no outpatient primary care visits or any palliative care visits in the year before their index date, individuals whose date of death was within one day of their index date, and individuals with missing demographic data. Finally, this study was undertaken in conjunction with a study of facility-level care patterns; we therefore excluded individuals if their facility was linked to fewer than 50 veterans in the study cohort. We used clinical and administrative data from the VHA Central Data Warehouse (CDW) from January 1, 2008, to December 31, 2016, encompassing two years of baseline data and one year of follow-up data for all cohort members. Cohort flow diagram. We adapted prior VHA operational research on adherence to clinical practice guidelines for opioid therapy to define 10 potentially pain-related treatment modalities: psychosocial treatment, physical therapy/occupational therapy (PT/OT), specialty pain clinic, complementary and integrative health (CIH), anticonvulsants, antidepressants, skeletal muscle relaxants, nonsteroidal anti-inflammatory drugs (NSAIDs), topical analgesics, and opioid medications (Table A1) [12]. Gender, date of birth, and race/ethnicity were obtained from administrative files. We assessed rurality of individuals’ residence using the VHA’s urban/rural/highly rural classification scheme [13]. We assigned every cohort member to the VHA facility where they were observed to receive most of their primary care during the year before and after the chronic pain index date. We categorized facilities as medical centers, urban clinics, or rural clinics according to the VHA’s site classification scheme [14]. We identified pain-relevant diagnoses (back pain, neck and other joint pain, neuropathy, migraine, fibromyalgia), mental health diagnoses (depression, anxiety, post-traumatic stress disorder, bipolar disorder), and substance use disorder diagnoses (opioid use disorder, alcohol use disorder, other substance use disorders), categorized consistent with prior work (Table A1) [15,16]. We recorded diagnoses and implemented the Quan-Charlson Comorbidity Index in the two years before or six months after veterans’ chronic pain index date [17,18]. Characteristics of veterans with incident chronic pain, Veterans Health Administration, 2010–2016 (N = 2,095,938) SUD = substance use disorder. Race categories not mutually exclusive. Characteristics of veterans with incident chronic pain, Veterans Health Administration, 2010–2016 (N = 2,095,938) SUD = substance use disorder. Race categories not mutually exclusive. We identified utilization of each treatment modality in the year following each veteran’s chronic pain index date and calculated the total nonopioid treatment modalities utilized during this period. To estimate adjusted annual utilization for each modality, we fit quasi-Poisson general linear models for each modality, with treatment modality utilization as the outcome. Models were adjusted for calendar year of index date, age at index date, gender, race/ethnicity, rurality of patient residence, pain diagnoses, mental health diagnoses, substance use disorder diagnoses, and Quan-Charlson Comorbidity Index. Individuals were censored during follow-up for mortality or if they utilized palliative care. We assigned a full weight of 365 days to those who were not censored and a smaller weight corresponding to follow-up time in days for those who were censored. The activities undertaken in this project supported VHA operational programs and did not constitute research; institutional review board approval was therefore not required. Of 2,095,938 veterans with incident chronic pain, 91% were male; 47% were aged 61 years or older; 78% were white; 27% had a diagnosis of back pain, 34% neck and other joint pain, 5% migraine, and 3% neuropathic pain; and 19% had diagnoses of depression and 14% post-traumatic stress disorder (Table 1). Among this cohort, 85% met inclusion criteria for chronic pain diagnoses, 51% met criteria for persistently elevated pain intensity ratings, and 25% met criteria for long-term opioid therapy (Figure A1). From 2010 to 2016, unadjusted utilization of seven of nine nonopioid treatment modalities increased (Figure 2). The largest absolute increase was seen in the utilization of PT/OT, from 30.9% in 2010–2011 to 36.7% in 2015–2016, an adjusted relative annual increase of 2.9% (95% confidence interval [CI] = 2.8%–2.9%) (Table 2). The largest relative increase was seen in the utilization of topical medications, an adjusted relative annual increase of 9.3% (95% CI = 9.1%–9.4%). Utilization of skeletal muscle relaxants and NSAIDs decreased during the period studied. Similarly, utilization of any opioid medication and of long-term opioid therapy decreased. Initiation of long-term opioid therapy declined from 23.7% in 2010–2011 to 9.9% in 2015–2016, an adjusted relative annual decrease of 13.9%. The proportion of veterans with incident chronic pain who utilized three or more treatment modalities increased from 35% in 2010–2011 to 41% in 2015–2016. Adjusted trends in utilization of pain-related treatment among veterans with incident chronic pain,* Veterans Health Administration, 2010–2016† CI = confidence interval; CIH = complementary/integrative health; NSAIDs = nonsteroidal anti-inflammatory drugs; PT/OT = physical therapy/occupational therapy. Incident chronic pain defined as at least one of three criteria: 1) persistent moderate–severe pain defined as at least three pain numeric rating scores ≥4 within a single year with ≥30 days between scores, 2) diagnosis codes “highly likely” to be related to chronic pain, and 3) receipt of long-term opioid therapy defined as ≥90 days during a 365-day period. Sample size across all years: N = 2,095,938. Adjusted estimates generated using a quasi-Poisson generalized linear model, adjusting for age at baseline (<30, 31–60, >60 years), gender, race/ethnicity (white, black, other race/ethnicity, missing), baseline calendar year, home facility type (medical center, urban clinic, rural clinic), pain diagnoses (back pain, neck/other joint pain, neuropathy, migraine, fibromyalgia), mental health diagnoses (depression, anxiety, post-traumatic stress disorder, bipolar disorder), substance use disorder diagnoses (alcohol use disorder, opioid use disorder, other substance use disorder), and Quan-Charlson Comorbidity Index. Adjusted trends in utilization of pain-related treatment among veterans with incident chronic pain,* Veterans Health Administration, 2010–2016† CI = confidence interval; CIH = complementary/integrative health; NSAIDs = nonsteroidal anti-inflammatory drugs; PT/OT = physical therapy/occupational therapy. Incident chronic pain defined as at least one of three criteria: 1) persistent moderate–severe pain defined as at least three pain numeric rating scores ≥4 within a single year with ≥30 days between scores, 2) diagnosis codes “highly likely” to be related to chronic pain, and 3) receipt of long-term opioid therapy defined as ≥90 days during a 365-day period. Sample size across all years: N = 2,095,938. Adjusted estimates generated using a quasi-Poisson generalized linear model, adjusting for age at baseline (<30, 31–60, >60 years), gender, race/ethnicity (white, black, other race/ethnicity, missing), baseline calendar year, home facility type (medical center, urban clinic, rural clinic), pain diagnoses (back pain, neck/other joint pain, neuropathy, migraine, fibromyalgia), mental health diagnoses (depression, anxiety, post-traumatic stress disorder, bipolar disorder), substance use disorder diagnoses (alcohol use disorder, opioid use disorder, other substance use disorder), and Quan-Charlson Comorbidity Index. Utilization of individual pain-related treatment modalities among veterans with incident chronic pain, Veterans Health Administration, 2010–2016 (sample size across all years: N = 2,095,938). CIH = complementary and integrative health; NSAIDs = nonsteroidal anti-inflammatory drugs; PT/OT = physical therapy/occupational therapy. The modalities examined in this study align with prior studies and can be interpreted in the context of broader trends in the use of nonopioid medications and nonpharmacologic therapies. In a nationally representative sample of outpatient visits for back pain from 1999 to 2010, prescribing of over-the-counter analgesics (NSAIDs or acetaminophen) decreased by 34% while prescribing of neuropathic pain agents (e.g., gabapentin) increased by 132%, a trend that appears to have accelerated in recent years [19,20]; referrals to physical therapy did not change, whereas referrals to other providers doubled. In the current study, utilization of PT/OT and specialty pain care services increased. In a national survey, use of several complementary and integrative health modalities (e.g., yoga, tai chi) increased from 2002 to 2012 [21]. Among veterans with incident chronic pain receiving care in the VHA, we found that this trend has continued since 2010. Potential explanations of greater nonopioid treatment utilization include changes in patient preferences, provider behaviors, and the VHA’s capacity to support multimodal chronic pain care. Veterans and VHA providers have identified interest in nonpharmacologic therapies but have also identified barriers to uptake [22–24]. During the study period, providers’ decision-making in chronic pain management was likely influenced by evolving guidelines and evidence. In 2010, the Departments of Veterans Affairs and Defense (VA/DOD) updated their guidelines on the management of opioid therapy, which reinforced an emphasis nonopioid treatments as firstline for chronic pain [25]. Since 2010, emerging evidence on the benefits and risks of specific medication classes such as NSAIDs and skeletal muscle relaxants highlighted the limitations of these medications for chronic pain [4]. Of note, the utilization trends seen in this national cohort of veterans precede the 2016 Centers for Disease Control and Prevention guideline and the 2017 VA/DOD guideline on opioid prescribing for chronic pain [2,3]. Future work should continue to assess changes in chronic pain management as patients and providers respond to more recent expert guidelines and more current evidence. Additionally, future work must examine the effects of these changes on important patient outcomes including pain-related function and quality of life, as well as broader behavioral health outcomes. During the study period, the VHA implemented several system-wide programs that were designed based on the principles of academic detailing and tailored to the needs of primary care providers [26]. These programs sought to improve the effectiveness and safety of chronic pain care through pain management education, clinical support and oversight, and expanded nonpharmacologic resources. The Pain ECHO program offered participants longitudinal video-conference-based consultation and education [15,27]. The Pain Mini-Residency programs provided pain management education tailored to the needs of primary care providers [26]. The VHA’s Opioid Safety Initiative, supported by the implementation of an academic detailing program, enhanced oversight of opioid prescribing practices for providers, facilities, and system-wide [8]. Real-time data are made available to providers and facility leaders using complementary dashboards such as the patient-level Opioid Therapy Risk Report and Stratification Tool for Opioid Risk Mitigation. Monitoring of high-risk opioid prescribing practices (e.g., high-dose opioid prescribing or concurrent benzodiazepine prescribing) and of utilization of risk mitigation strategies (e.g., urine drug testing) guides the implementation of efforts to improve the safety of pain management for veterans prescribed opioid medications. The VHA has also taken steps to expand access to nonopioid treatment options. For example, beginning in 2014, the VHA and the Military Health System partnered to develop and implement a system-wide curriculum to train front-line providers in acupuncture [28]. Nonopioid pain care options will likely further increase in the VHA with implementation of the 2016 Comprehensive Addiction and Recovery Act [29]. Among other provisions, this law mandated that each VHA medical facility designate a pain management team of health care professionals to coordinate pain management therapy. Several sections authorized expanded research, education, and delivery of complementary and integrative health. More recently, a 2017 VHA directive established a policy that VA practitioners “proactively offer and include, as appropriate,” any of several evidence-based CIH approaches [30]. However, measurement of utilization of nonpharmacologic modalities may become more challenging with ongoing implementation of the Veterans Choice Program, which will facilitate greater utilization of community-based resources outside VHA settings [31]. The trends in this study can provide lessons and benchmarks for non-VHA health care systems, where initiatives have sought to improve the safety of opioid prescribing [32–34]. First, the VHA has implemented initiatives across multiple fronts and sustained many of them over multiple years. As the VHA and its priorities have evolved during this period, it has been important that these initiatives be adequately adaptable to the unique needs of individual facilities. Second, it is critical that the VHA and non-VHA health care systems collaborate to align approaches to evaluation and share best practices. A recently funded interagency partnership between the US Departments of Health and Human Services, Defense, and Veterans Affairs is an example of a promising initiative to support this goal [35]. This study has several limitations. First, we cannot identify treatment that was received in other health care systems or in community settings, over-the-counter pain medications, or pain self-management strategies [36]. Therefore, pain treatment utilization is likely underestimated. Second, this large national cohort of veterans with incident chronic pain may not be representative of nonveterans, veterans not enrolled in VHA health care, or patients with longstanding chronic pain. Third, several treatment modalities have primary indications other than chronic pain (e.g., psychosocial treatment, anticonvulsants, antidepressants), and we are not able to identify the indication for each treatment modality for an individual veteran. These modalities may have been delivered with a primary indication of depression or insomnia, which commonly co-occur in veterans with chronic pain. In a biopsychosocial model of chronic pain, treatment of co-occurring conditions can be considered a component of multimodal pain care. Finally, the National Pain Strategy calls for pain care that is integrated, multimodal, and multidisciplinary [37]. We have measured utilization of multiple treatment modalities but cannot characterize whether treatments were integrated from the patient perspective. In summary, among more than 2 million veterans with incident chronic pain from 2010 to 2016, utilization of multiple nonopioid treatment modalities increased, as did the proportion of veterans utilizing multiple modalities. Future work should implement and evaluate models to support patient-centered integration of these modalities. Additionally, the effects of integrated, multimodal pain care on important outcomes such as function and quality of life require further study. Funding sources: This work was supported by the Department of Veterans Affairs Quality Enhancement Research Initiative (QUE 15-468). Dr. Frank was supported by a VA Health Services Research & Development Career Development Award (HSR&D CDA 15-059). Dr. Kerns was supported in part by the VA Health Services Research & Development Center of Innovation (CIN 13-047). The funding organizations had no role in the design or conduct of the study; in the collection, analysis, or interpretation of the data; in the writing of the report; or in the decision to submit for publication. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the US Government. Conflicts of interest: No potential conflicts exist. Supplementary data are available at Pain Medicine online. Cohort inclusion by chronic pain criteria. LTOT = long-term opioid therapy.