TY - JOUR
AU - Lund, Brian C.
AB - Abstract Background. Understanding opioid prescribing trends requires differentiating clinically distinct short- and long-term receipt patterns. Objectives. Describe the one-year course of opioid receipt among new opioid recipients and determine the proportion with subsequent long-term opioid therapy. Discern variation in proportion with long-term therapy initiation by geographic region and across Veterans Health Administration (VHA) medical centers. Methods. Longitudinal course of opioid receipt was analyzed using a cabinet supply approach. Short-term receipt was defined as index treatment episode lasting no longer than 30 days; long-term therapy as treatment episode of >90 days that began within the first 30 days following opioid index date. Patients. All VHA pharmacy users in 2004 and to 2011 who received a new prescription for an opioid (incident opioid recipients) preceded by 365 days with no opioid prescribed. Results. The proportion of all incident recipients who met the definition for long-term therapy within the first year decreased from 20.4% (N = 76,280) in 2004 to 18.3% (N = 96,166) in 2011. The proportion of incident recipients with chronic pain was unchanged between 2004 and 2011. Hydrocodone and tramadol increased as a proportion of initial opioids prescribed. Median days initially supplied decreased from 30 to 20 days. A greater percentage of new opioid prescriptions were for 7 days or fewer (20.9% in 2004; 27.9% in 2011). The proportion of new recipients who initiated long-term opioid therapy varied widely by medical center. Medical centers with higher proportions of new long-term recipients in 2004 saw greater decreases in this metric by 2011. Conclusion. The proportion of new opioid recipients who initiated long-term opioid therapy declined between 2004 and 2011. Opioids, Chronic Non-Cancer Pain, Incidence, Prevalence, Veterans Background Opioid medications are well-accepted in short-term treatment of moderate to severe acute pain, and in long-term treatment for cancer-related pain or symptoms at end-of-life. While acceptance of opioids for chronic non-cancer pain (CNCP) gained momentum in the United States over prior decades [ 1–3 ], rapidly increasing opioid prescribing and parallel increases in adverse outcomes have brought long-term opioid therapy (LTOT) under closer scrutiny [ 4–9 ]. Recent guideline recommendations caution against opioids for CNCP, as evidence for benefit is insufficient weighed against concern for medical adverse effects and overdose risk, which increase with dose and duration [ 10 , 11 ]. Prior research in Veterans Health Administration (VHA) documented steady increases in the annual prevalence of opioid prescribing, from 18.9% in fiscal year (FY) 04 to 33.4% in FY 12 [ 4 ]. However, incidence rates remained relatively unchanged during this time frame, increasing from 8.8% to 9.8%, demonstrating that escalating opioid prescribing in VHA was not explained simply by an expanding base of new short-term recipients or increased frequency of short-term receipt. Instead, this pattern could be explained by increased rates of long-term therapy among veterans newly prescribed opioids, in addition to the simple accrual of patients on opioid therapy over a number of years. By directly measuring the duration of opioid therapy following incident prescription, we can distinguish if LTOT became more frequent between 2004 and 2011 among veterans newly receiving an opioid. Therefore, our primary objective is to describe the 1-year course of opioid receipt among new recipients in VHA, and specifically to determine the proportion of new recipients who began LTOT. Previous studies have suggested that long-term receipt patterns are established within the first year following initiation [ 12 ]. We hypothesized that the proportion of new opioid recipients prescribed long-term therapy would increase, reflecting a growing proclivity toward initiation of long-term opioid therapy. We similarly hypothesized that total calendar days of opioid receipt within 1 year following initiation would increase from 2004 to 2011. Secondary objectives were to contrast initial prescription characteristics over time (e.g., dose, days supplied), and to characterize variation in the proportion of LTOT by geographic region and across VHA medical centers. We used a novel cabinet supply approach to characterize prescription patterns following incident receipt. By describing the prescription patterns that culminate in long-term therapy, we inform policies and interventions aimed at encouraging safe and rational opioid prescribing; efforts to reduce rates of new long-term therapy initiation should be considered alongside those to taper dosing in patients on established LTOT. Understanding prescribing patterns following incident prescription that lead to persistent opioid receipt may aid clinicians in identifying and avoiding unintended long-term therapy before it becomes established. Methods Data Sources National administrative VHA data were obtained through the Austin Information Technology Center for FY 2003 through 2005 and 2010 through 2012 (Austin, TX). Decision Support System National Data Extracts were used to identify opioid medication dispensing events. Inpatient and outpatient Medical SAS datasets were used to identify diagnostic codes and health care services defined in the patient selection criteria. The study was approved by the University of Iowa Institutional Review Board and the Iowa City VA Health Care System Research and Development Committee. New Opioid Users New opioid recipients (i.e., incident recipients) were identified during 2 target years using identical methods, which produced two separate cohorts for comparison: a 2004 new recipient cohort and a 2011 new recipient cohort. The target years were the earliest and most recent years that comparable data were available, and preceded the widespread implementation of the Opioid Safety Initiative (OSI) aimed at reducing high-dose therapy in patients receiving opioids long term; we purposefully did not examine trends following the OSI as our primary aim was to better account for increased opioid prevalence concurrent with decreased incidence prior to 2011. Patients were required to have an opioid dispensed during the target year with no opioids dispensed during the prior 365 days, where the date of first receipt is referred to as the opioid index date. Opioids included non-injectable dosage forms of butorphanol, fentanyl, hydrocodone, hydromorphone, levorphanol, meperidine, methadone, morphine, oxycodone, oxymorphone, pentazocine, and tramadol. Only opioids prescribed in the outpatient setting were included. Additional characteristics were determined for the index opioid prescription, including the initial days’ supply dispensed, whether an extended-release dosage form was used, and the estimated daily dose in terms of morphine equivalents [ 13 ]. Patients were excluded if they had a diagnostic code for metastatic cancer (ICD-9 codes 196.x-199.x) or received hospice or palliative care services at any point in the year prior to or following the opioid index date. Hospice and palliative care services were identified in outpatient data using clinic stop codes 351 or 353 and in inpatient data using ICD-9 code V667 or bedsection code 96. Patients were also excluded if they had an outpatient visit to an opioid treatment program (stop code 523) at any point during the year prior to or within the first 30 days following the opioid index date. These criteria were applied to exclude opioid recipients with globally accepted indications for long-term opioid therapy. Course of Opioid Use The longitudinal course of opioid receipt following the index date was analyzed using a cabinet supply approach. On the opioid index date (Day 1) all patients were assigned a cabinet supply equal to the days’ supply of opioid dispensed in the index prescription. One day’s supply was subtracted on each sequential day until the cabinet supply reached zero, which was carried forward until another opioid prescription was dispensed. If an opioid prescription was dispensed prior to the cabinet supply reaching zero, the new supply was added to the existing supply. The cabinet supply approach thus created an array of 365 values corresponding to the estimated days’ supply of opioid the patient had on hand for each day over the course of the year. The total calendar days of receipt was determined by the number of non-zero cabinet days in the array. This array was further used to determine the duration of opioid treatment episodes. All patients began their first episode on the opioid index date (Day 1). This episode ended when, moving forward through the cabinet supply array, a period of sequential zero-supply days occurred that exceeded the days’ supply value of the most recent opioid prescription. An end date for the episode was assigned as the last observed non-zero cabinet supply date following this opioid prescription date. If another opioid prescription was observed after this point, it marked the onset of a new opioid treatment episode and the process for assigning episodes was continued forward through the 1-year follow-up period. Information regarding the timing and duration of opioid treatment episodes was used to define clinically meaningful treatment courses. The 1-year course was divided into two phases, the initial index phase and the subsequent follow-up phase. The index phase described the course of opioid treatment in the first 90 days following opioid initiation. Short-term therapy was defined as an index treatment episode lasting no longer than 30 days, and no opioids dispensed between Days 31 and 90. Long-term therapy was defined as a treatment episode of > 90 days that began within the first 30 days following the opioid index date. A minimum treatment episode of 91 days is consistent with existing literature as the conceptual definition of long-term or chronic opioid therapy [ 13–15 ]. The remaining initial opioid receipt patterns were categorized as intermediate. The follow-up phase then described the subsequent course of opioid receipt within the three index course groups. Initial short-term recipients were classified into the following three follow-up groups: (a) no further opioid receipt (no opioid prescriptions during Days 91–365); (b) new long-term recipient (having a subsequent treatment episode lasting >90 days); or (c) new intermediate recipient, which included all other opioid receipt patterns . Initial intermediate recipients were classified into the following three follow-up groups: (a) no further opioid receipt (no opioid prescriptions during Days 91–365); (b) new long-term recipient (having a subsequent treatment episode lasting >90 days); or (c) on-going intermediate recipient, which included all other opioid receipt patterns. Initial long-term recipients were classified into two follow-up groups: (a) on-going long-term receipt (≥200 days of non-zero cabinet supply during Days 91–365); (b) transition to intermediate receipt, which included all other opioid receipt patterns. These detailed course descriptions were also used to determine a single summary measure, the proportion of new opioid recipients who were classified as long-term recipients during the year following initiation, or LTOT proportion. Analysis Descriptive statistics were used to characterize patients and to describe the 1-year course of opioid receipt for both 2004 and 2011 cohorts. Inferential statistics were not used in cross-cohort comparisons because these cohorts were not samples, but rather the entire population of new opioid recipients in VHA. Two sets of sensitivity analyses were performed to assess the stability of the main findings concerning LTOT proportion estimates during 2004 and 2011. The first applied additional patient selection criteria to identify patients with a history of VHA service use in the year prior to opioid initiation. Such criteria are often applied when identifying new medication use to increase confidence that the medication of interest was not used prior to the observed start date. These criteria included the presence of at least one VHA primary care encounter in the year prior to opioid index and a history of at least 240 days of regular VHA medication use in the year prior to opioid index [ 4 ]. These additional criteria were applied as a sensitivity analysis, rather than in the primary selection criteria, because of concern that it may be overly restrictive, particularly for younger veterans with lower rates of chronic disease, and that including similar criteria had little impact on our findings in prior research [ 4 ]. The second set of sensitivity analyses examined alternative operational criteria to define long-term opioid use [ 13 , 15 , 16 ]. All analyses were conducted using SAS version 9.3 (SAS Institute, Inc., Cary, NC). Results The number of veterans receiving opioids in VHA nearly doubled from 684,425 in 2004 to 1,298,226 in 2011, which represented an increase in annual period prevalence from 16.3% to 27.4% ( Figure 1 ). New opioid recipients comprised 56.9% (389,409/684,425) of all opioid recipients in 2004, which fell to 42.1% (546,918/1,298,226) in 2011. After applying all selection criteria, our final study cohorts included 373,276 new opioid recipients in 2004 and 526,499 in 2011. Figure 1 Open in new tabDownload slide Cohort selection in 2004 and in 2011. Figure 1 Open in new tabDownload slide Cohort selection in 2004 and in 2011. Patient Characteristics The characteristics of new opioid recipients remained largely consistent during the observation period ( Table 1 ). In 2011, a greater proportion of new recipients were younger than 35 years, black, and from urban areas. The proportion of patients with mental health and substance abuse disorders, mental health visits, and with ≥ 4 comorbid conditions also increased. The proportion of new opioid recipients with chronic pain remained unchanged and most patients initiated opioids in the outpatient setting in both years. Table 1 Clinical characteristics of Veterans with new opioid use in FY 2004 and 2011 Clinical Characteristics . FY 2004 . FY 2011 . N = 373,276 . N = 526,499 . n (%) . n (%) . Demographics Age  18–34 18564 (5.0) 48803 (9.3)  35–49 65959 (17.7) 78650 (14.9)  50–64 149776 (40.1) 219431 (41.7)  65–79 95877 (25.7) 126891 (24.1)  ≥80 32617 (8.7) 46830 (8.9)  Unknown 10483 (2.8) 5894 (1.1) Sex  Male 337824 (90.5) 480273 (91.2)  Female 24979 (6.7) 44875 (8.5)  Unknown 10473 (2.8) 1351 (0.3) Race  White 246422 (66.0) 326373 (62.0)  Black 59611 (16.0) 99354 (18.9)  Other 26054 (7.0) 42134 (8.0)  Unknown 41189 (11.0) 58638 (11.1) Residence  Urban 259282 (69.5) 383740 (72.9)  Large rural 46611 (12.5) 64955 (12.3)  Small rural 27582 (7.4) 36159 (6.9)  Isolated 23694 (6.4) 30643 (5.8)  Unknown 16107 (4.3) 11002 (2.1) Medical conditions* Chronic pain 177125 (47.5) 250987 (47.7) Anxiety disorder 24835 (6.7) 49278 (9.4) Depressive disorder 35132 (9.4) 55575 (10.6) Substance abuse 37161 (10.0) 64932 (12.3) Count of comorbidities [ 27 ]  0–1 158128 (42.4) 207017 (39.3)  2–3 142503 (38.2) 201794 (38.3)  ≥4 72645 (19.5) 117688 (22.4) Health services utilization Health care setting of opioid initiation  Inpatient surgery 26685 (7.2) 26434 (5.0)  Other inpatient 13496 (3.6) 16545 (3.1)  Specialty pain clinic 2319 (0.6) 2100 (0.4)  Other outpatient 330776 (88.6) 481420 (91.4) Hospitalization count, year prior to opioid initiation  None 301595 (80.8) 449106 (85.3)  1 51829 (13.9) 59053 (11.2)  ≥2 19852 (5.3) 18340 (3.5) Mental health visit count, year prior to opioid initiation  None 282395 (75.7) 360199 (68.4)  1–2 32919 (8.8) 57964 (11.0)  ≥3 57962 (15.5) 108336 (20.6) Concurrent medications at opioid initiation [ 28 ]  0–2 104158 (27.9) 157913 (30.0)  3–4 82587 (22.1) 117025 (22.2)  5–6 68727 (18.4) 90773 (17.2)  ≥7 117804 (31.6) 160788 (30.5) Unique prescribers of concurrent medications  1 156355 (41.9) 227061 (43.1)  2 112352 (30.1) 153332 (29.1)  ≥3 104569 (28.0) 146106 (27.8) Clinical Characteristics . FY 2004 . FY 2011 . N = 373,276 . N = 526,499 . n (%) . n (%) . Demographics Age  18–34 18564 (5.0) 48803 (9.3)  35–49 65959 (17.7) 78650 (14.9)  50–64 149776 (40.1) 219431 (41.7)  65–79 95877 (25.7) 126891 (24.1)  ≥80 32617 (8.7) 46830 (8.9)  Unknown 10483 (2.8) 5894 (1.1) Sex  Male 337824 (90.5) 480273 (91.2)  Female 24979 (6.7) 44875 (8.5)  Unknown 10473 (2.8) 1351 (0.3) Race  White 246422 (66.0) 326373 (62.0)  Black 59611 (16.0) 99354 (18.9)  Other 26054 (7.0) 42134 (8.0)  Unknown 41189 (11.0) 58638 (11.1) Residence  Urban 259282 (69.5) 383740 (72.9)  Large rural 46611 (12.5) 64955 (12.3)  Small rural 27582 (7.4) 36159 (6.9)  Isolated 23694 (6.4) 30643 (5.8)  Unknown 16107 (4.3) 11002 (2.1) Medical conditions* Chronic pain 177125 (47.5) 250987 (47.7) Anxiety disorder 24835 (6.7) 49278 (9.4) Depressive disorder 35132 (9.4) 55575 (10.6) Substance abuse 37161 (10.0) 64932 (12.3) Count of comorbidities [ 27 ]  0–1 158128 (42.4) 207017 (39.3)  2–3 142503 (38.2) 201794 (38.3)  ≥4 72645 (19.5) 117688 (22.4) Health services utilization Health care setting of opioid initiation  Inpatient surgery 26685 (7.2) 26434 (5.0)  Other inpatient 13496 (3.6) 16545 (3.1)  Specialty pain clinic 2319 (0.6) 2100 (0.4)  Other outpatient 330776 (88.6) 481420 (91.4) Hospitalization count, year prior to opioid initiation  None 301595 (80.8) 449106 (85.3)  1 51829 (13.9) 59053 (11.2)  ≥2 19852 (5.3) 18340 (3.5) Mental health visit count, year prior to opioid initiation  None 282395 (75.7) 360199 (68.4)  1–2 32919 (8.8) 57964 (11.0)  ≥3 57962 (15.5) 108336 (20.6) Concurrent medications at opioid initiation [ 28 ]  0–2 104158 (27.9) 157913 (30.0)  3–4 82587 (22.1) 117025 (22.2)  5–6 68727 (18.4) 90773 (17.2)  ≥7 117804 (31.6) 160788 (30.5) Unique prescribers of concurrent medications  1 156355 (41.9) 227061 (43.1)  2 112352 (30.1) 153332 (29.1)  ≥3 104569 (28.0) 146106 (27.8) * In year prior to opioid index date. Open in new tab Table 1 Clinical characteristics of Veterans with new opioid use in FY 2004 and 2011 Clinical Characteristics . FY 2004 . FY 2011 . N = 373,276 . N = 526,499 . n (%) . n (%) . Demographics Age  18–34 18564 (5.0) 48803 (9.3)  35–49 65959 (17.7) 78650 (14.9)  50–64 149776 (40.1) 219431 (41.7)  65–79 95877 (25.7) 126891 (24.1)  ≥80 32617 (8.7) 46830 (8.9)  Unknown 10483 (2.8) 5894 (1.1) Sex  Male 337824 (90.5) 480273 (91.2)  Female 24979 (6.7) 44875 (8.5)  Unknown 10473 (2.8) 1351 (0.3) Race  White 246422 (66.0) 326373 (62.0)  Black 59611 (16.0) 99354 (18.9)  Other 26054 (7.0) 42134 (8.0)  Unknown 41189 (11.0) 58638 (11.1) Residence  Urban 259282 (69.5) 383740 (72.9)  Large rural 46611 (12.5) 64955 (12.3)  Small rural 27582 (7.4) 36159 (6.9)  Isolated 23694 (6.4) 30643 (5.8)  Unknown 16107 (4.3) 11002 (2.1) Medical conditions* Chronic pain 177125 (47.5) 250987 (47.7) Anxiety disorder 24835 (6.7) 49278 (9.4) Depressive disorder 35132 (9.4) 55575 (10.6) Substance abuse 37161 (10.0) 64932 (12.3) Count of comorbidities [ 27 ]  0–1 158128 (42.4) 207017 (39.3)  2–3 142503 (38.2) 201794 (38.3)  ≥4 72645 (19.5) 117688 (22.4) Health services utilization Health care setting of opioid initiation  Inpatient surgery 26685 (7.2) 26434 (5.0)  Other inpatient 13496 (3.6) 16545 (3.1)  Specialty pain clinic 2319 (0.6) 2100 (0.4)  Other outpatient 330776 (88.6) 481420 (91.4) Hospitalization count, year prior to opioid initiation  None 301595 (80.8) 449106 (85.3)  1 51829 (13.9) 59053 (11.2)  ≥2 19852 (5.3) 18340 (3.5) Mental health visit count, year prior to opioid initiation  None 282395 (75.7) 360199 (68.4)  1–2 32919 (8.8) 57964 (11.0)  ≥3 57962 (15.5) 108336 (20.6) Concurrent medications at opioid initiation [ 28 ]  0–2 104158 (27.9) 157913 (30.0)  3–4 82587 (22.1) 117025 (22.2)  5–6 68727 (18.4) 90773 (17.2)  ≥7 117804 (31.6) 160788 (30.5) Unique prescribers of concurrent medications  1 156355 (41.9) 227061 (43.1)  2 112352 (30.1) 153332 (29.1)  ≥3 104569 (28.0) 146106 (27.8) Clinical Characteristics . FY 2004 . FY 2011 . N = 373,276 . N = 526,499 . n (%) . n (%) . Demographics Age  18–34 18564 (5.0) 48803 (9.3)  35–49 65959 (17.7) 78650 (14.9)  50–64 149776 (40.1) 219431 (41.7)  65–79 95877 (25.7) 126891 (24.1)  ≥80 32617 (8.7) 46830 (8.9)  Unknown 10483 (2.8) 5894 (1.1) Sex  Male 337824 (90.5) 480273 (91.2)  Female 24979 (6.7) 44875 (8.5)  Unknown 10473 (2.8) 1351 (0.3) Race  White 246422 (66.0) 326373 (62.0)  Black 59611 (16.0) 99354 (18.9)  Other 26054 (7.0) 42134 (8.0)  Unknown 41189 (11.0) 58638 (11.1) Residence  Urban 259282 (69.5) 383740 (72.9)  Large rural 46611 (12.5) 64955 (12.3)  Small rural 27582 (7.4) 36159 (6.9)  Isolated 23694 (6.4) 30643 (5.8)  Unknown 16107 (4.3) 11002 (2.1) Medical conditions* Chronic pain 177125 (47.5) 250987 (47.7) Anxiety disorder 24835 (6.7) 49278 (9.4) Depressive disorder 35132 (9.4) 55575 (10.6) Substance abuse 37161 (10.0) 64932 (12.3) Count of comorbidities [ 27 ]  0–1 158128 (42.4) 207017 (39.3)  2–3 142503 (38.2) 201794 (38.3)  ≥4 72645 (19.5) 117688 (22.4) Health services utilization Health care setting of opioid initiation  Inpatient surgery 26685 (7.2) 26434 (5.0)  Other inpatient 13496 (3.6) 16545 (3.1)  Specialty pain clinic 2319 (0.6) 2100 (0.4)  Other outpatient 330776 (88.6) 481420 (91.4) Hospitalization count, year prior to opioid initiation  None 301595 (80.8) 449106 (85.3)  1 51829 (13.9) 59053 (11.2)  ≥2 19852 (5.3) 18340 (3.5) Mental health visit count, year prior to opioid initiation  None 282395 (75.7) 360199 (68.4)  1–2 32919 (8.8) 57964 (11.0)  ≥3 57962 (15.5) 108336 (20.6) Concurrent medications at opioid initiation [ 28 ]  0–2 104158 (27.9) 157913 (30.0)  3–4 82587 (22.1) 117025 (22.2)  5–6 68727 (18.4) 90773 (17.2)  ≥7 117804 (31.6) 160788 (30.5) Unique prescribers of concurrent medications  1 156355 (41.9) 227061 (43.1)  2 112352 (30.1) 153332 (29.1)  ≥3 104569 (28.0) 146106 (27.8) * In year prior to opioid index date. Open in new tab Opioid Course, 2004 The majority of new opioid recipients in 2004 had initial short-term therapy (60.4%; n = 225,620), where their index treatment episode lasted no longer than 30 days, and no opioids were dispensed between Days 31 and 90 ( Figure 2 ). Of these, 161,247 (71.5%, or 43.2% of all new recipients) remained opioid free through the remainder of the year following initiation. These patients received a median total days’ supply of 15 (interquartile range [IQR] 7, 30). Only 3.1% of initial short-term recipients (7,044, or 1.9% of all new recipients) went on to long-term therapy at some point during the year. Of the remaining 57,329 initial short-term recipients, 25.4% (or 15.4% of all new recipients) had subsequent opioid prescriptions and received a median total days’ supply of 59 (IQR: 35, 88) over the year following initiation. Figure 2 Open in new tabDownload slide One-year use patterns following incident opioid use in 2004 and 2011. Figure 2 Open in new tabDownload slide One-year use patterns following incident opioid use in 2004 and 2011. Initial intermediate receipt was observed in 93,782 (25.1%) new opioid recipients in 2004. Of these, 33,382 (35.6%, or 8.9% of all recipients) had no further opioid receipt beyond the first 90 days, 15,290 (16.3%, or 4.1% of all new recipients) escalated to long-term therapy, and 45,110 (48.1%, or 12.1% of all new recipients) maintained intermediate receipt. Initial long-term receipt was present in 53,874 (14.4%) veterans. The majority, 29,584 (54.9%, or 7.9% of all new recipients) had ongoing regular opioid receipt, with a median of 353 (IQR: 325, 365) total days over the subsequent year. The 24,290 remaining veterans (45.1%, or 6.5% of all new recipients) transitioned to intermediate receipt. Considering all eight course patterns together, a total of 76,280 (20.4%) new opioid recipients in 2004 had long-term receipt within the year following initiation. Opioid Course, Changes from 2004 to 2011 The median total days of opioid receipt during the year following initiation was 36 (IQR: 16, 118) in 2004 and fell to 30 (IQR: 11, 115) in 2011. Initial opioid receipt patterns observed in 2011 were similar to 2004 ( Figure 2 ). Of 526,499 new recipients, 329,463 (62.6%) had initial short-term receipt, 130,688 (24.8%) had initial intermediate receipt, and 66,348 (12.6%) initial long-term receipt. The distribution of subsequent opioid receipt patterns in 2011 was also very similar, differing within only a few percentage points of 2004 findings. Considering all patterns together, a total of 96,166 (18.3%) new opioid recipients in 2011 went on to long-term therapy within the year following initiation. Index Opioid Prescriptions Between 2004 and 2011, the proportion of patients with index prescriptions for hydrocodone increased from 49.3% to 55.4%, and for tramadol from 23.0% to 31.3% ( Table 2 ) Oxycodone decreased from 24.0% to 11.4%. Extended-release dosage forms accounted for 2.9% of initial prescriptions in 2004, and 1.4% in 2011. The mean estimated dose in morphine equivalents per day (MED) at initiation was 22.6 (SD 26.0) in 2004, and 21.6 MED (SD 20.8) in 2011; the mean index days supplied in 2004 was 22.7 (SD 15.2) and 20.0 (SD 14.6) in 2011. In 2004, 30-day prescriptions accounted for 54.8% of index prescriptions, which fell to 44.7% by 2011. The percentage of index prescriptions for 7-day supply or fewer grew from 20.9% in 2004 to 27.9% in 2011. Table 2 Characteristics of the index opioid prescription at initiation Characteristic . 2004 . 2011 . N = 373,276 . N = 526,499 . Opioid, n (%)  Hydrocodone 183,977 (49.3) 291,830 (55.4)  Oxycodone 89,560 (24.0) 59,985 (11.4)  Tramadol 85,902 (23.0) 164,660 (31.3)  Morphine 6,935 (1.9) 5,916 (1.1)  Other 6,902 (1.8) 4,108 (0.8) Extended-release dosage form, n (%) 10,662 (2.9) 7,337 (1.4) Morphine equivalents (mg/day), mean (SD) 22.6 (26.0) 21.6 (20.8) Morphine equivalents (mg/day), median (IQR) 17.9 (10, 30) 20.0 (12.5, 25) Index supply (days), mean (SD) 22.7 (15.2) 20.0 (14.6) Index supply (days), median (IQR) 30 (10, 30) 20 (7, 30) Index supply (days), n (%)  ≤ 7 77,989 (20.9) 146,984 (27.9)  8–14 52,475 (14.1) 90,175 (17.1)  15–29 26,836 (7.2) 41,612 (7.9)  30 204,495 (54.8) 235,206 (44.7)  > 30 11,481 (3.1) 12,522 (2.9) Characteristic . 2004 . 2011 . N = 373,276 . N = 526,499 . Opioid, n (%)  Hydrocodone 183,977 (49.3) 291,830 (55.4)  Oxycodone 89,560 (24.0) 59,985 (11.4)  Tramadol 85,902 (23.0) 164,660 (31.3)  Morphine 6,935 (1.9) 5,916 (1.1)  Other 6,902 (1.8) 4,108 (0.8) Extended-release dosage form, n (%) 10,662 (2.9) 7,337 (1.4) Morphine equivalents (mg/day), mean (SD) 22.6 (26.0) 21.6 (20.8) Morphine equivalents (mg/day), median (IQR) 17.9 (10, 30) 20.0 (12.5, 25) Index supply (days), mean (SD) 22.7 (15.2) 20.0 (14.6) Index supply (days), median (IQR) 30 (10, 30) 20 (7, 30) Index supply (days), n (%)  ≤ 7 77,989 (20.9) 146,984 (27.9)  8–14 52,475 (14.1) 90,175 (17.1)  15–29 26,836 (7.2) 41,612 (7.9)  30 204,495 (54.8) 235,206 (44.7)  > 30 11,481 (3.1) 12,522 (2.9) Open in new tab Table 2 Characteristics of the index opioid prescription at initiation Characteristic . 2004 . 2011 . N = 373,276 . N = 526,499 . Opioid, n (%)  Hydrocodone 183,977 (49.3) 291,830 (55.4)  Oxycodone 89,560 (24.0) 59,985 (11.4)  Tramadol 85,902 (23.0) 164,660 (31.3)  Morphine 6,935 (1.9) 5,916 (1.1)  Other 6,902 (1.8) 4,108 (0.8) Extended-release dosage form, n (%) 10,662 (2.9) 7,337 (1.4) Morphine equivalents (mg/day), mean (SD) 22.6 (26.0) 21.6 (20.8) Morphine equivalents (mg/day), median (IQR) 17.9 (10, 30) 20.0 (12.5, 25) Index supply (days), mean (SD) 22.7 (15.2) 20.0 (14.6) Index supply (days), median (IQR) 30 (10, 30) 20 (7, 30) Index supply (days), n (%)  ≤ 7 77,989 (20.9) 146,984 (27.9)  8–14 52,475 (14.1) 90,175 (17.1)  15–29 26,836 (7.2) 41,612 (7.9)  30 204,495 (54.8) 235,206 (44.7)  > 30 11,481 (3.1) 12,522 (2.9) Characteristic . 2004 . 2011 . N = 373,276 . N = 526,499 . Opioid, n (%)  Hydrocodone 183,977 (49.3) 291,830 (55.4)  Oxycodone 89,560 (24.0) 59,985 (11.4)  Tramadol 85,902 (23.0) 164,660 (31.3)  Morphine 6,935 (1.9) 5,916 (1.1)  Other 6,902 (1.8) 4,108 (0.8) Extended-release dosage form, n (%) 10,662 (2.9) 7,337 (1.4) Morphine equivalents (mg/day), mean (SD) 22.6 (26.0) 21.6 (20.8) Morphine equivalents (mg/day), median (IQR) 17.9 (10, 30) 20.0 (12.5, 25) Index supply (days), mean (SD) 22.7 (15.2) 20.0 (14.6) Index supply (days), median (IQR) 30 (10, 30) 20 (7, 30) Index supply (days), n (%)  ≤ 7 77,989 (20.9) 146,984 (27.9)  8–14 52,475 (14.1) 90,175 (17.1)  15–29 26,836 (7.2) 41,612 (7.9)  30 204,495 (54.8) 235,206 (44.7)  > 30 11,481 (3.1) 12,522 (2.9) Open in new tab Variation in the Proportion of Long-Term Opioid Use The proportion of new opioid users who were classified as receiving LTOT during the year following initiation decreased nationally from 20.4% in 2004 to 18.3% in 2011 ( Table 3 ). The proportion with LTOT varied modestly across geographic regions in 2004, and decreased over time in each region, with the largest changes observed for regions with the highest starting proportion on LTOT. Table 3 Geographic variation in the proportion of long-term opioid recipients . Proportion of long-term opioid recipients * . Analysis Level . 2004 . 2011 . % (n/N) . % (n/N) . National 20.4% (76,208/373,276) 18.3% (96,166/526,499) Region  Northeast 16.7% (8,133/48,567) 16.0% (10,955/68,372)  Midwest 22.5% (15,353/68,232) 19.3% (20,937/108,309)  South 21.1% (35,208/166,793) 18.5% (42,724/230,440)  West 19.5% (17,514/89,684) 18.1% (21,550/119,378) Medical Centers (N = 137)  Minimum 10.5% 7.7%  25 th %ile 17.5% 15.7%  Median 19.8% 18.5%  75 th %ile 24.6% 22.1%  Maximum 38.6% 35.1% . Proportion of long-term opioid recipients * . Analysis Level . 2004 . 2011 . % (n/N) . % (n/N) . National 20.4% (76,208/373,276) 18.3% (96,166/526,499) Region  Northeast 16.7% (8,133/48,567) 16.0% (10,955/68,372)  Midwest 22.5% (15,353/68,232) 19.3% (20,937/108,309)  South 21.1% (35,208/166,793) 18.5% (42,724/230,440)  West 19.5% (17,514/89,684) 18.1% (21,550/119,378) Medical Centers (N = 137)  Minimum 10.5% 7.7%  25 th %ile 17.5% 15.7%  Median 19.8% 18.5%  75 th %ile 24.6% 22.1%  Maximum 38.6% 35.1% *Proportion of new opioid recipients who were classified as long-term recipients during the year following initiation. Open in new tab Table 3 Geographic variation in the proportion of long-term opioid recipients . Proportion of long-term opioid recipients * . Analysis Level . 2004 . 2011 . % (n/N) . % (n/N) . National 20.4% (76,208/373,276) 18.3% (96,166/526,499) Region  Northeast 16.7% (8,133/48,567) 16.0% (10,955/68,372)  Midwest 22.5% (15,353/68,232) 19.3% (20,937/108,309)  South 21.1% (35,208/166,793) 18.5% (42,724/230,440)  West 19.5% (17,514/89,684) 18.1% (21,550/119,378) Medical Centers (N = 137)  Minimum 10.5% 7.7%  25 th %ile 17.5% 15.7%  Median 19.8% 18.5%  75 th %ile 24.6% 22.1%  Maximum 38.6% 35.1% . Proportion of long-term opioid recipients * . Analysis Level . 2004 . 2011 . % (n/N) . % (n/N) . National 20.4% (76,208/373,276) 18.3% (96,166/526,499) Region  Northeast 16.7% (8,133/48,567) 16.0% (10,955/68,372)  Midwest 22.5% (15,353/68,232) 19.3% (20,937/108,309)  South 21.1% (35,208/166,793) 18.5% (42,724/230,440)  West 19.5% (17,514/89,684) 18.1% (21,550/119,378) Medical Centers (N = 137)  Minimum 10.5% 7.7%  25 th %ile 17.5% 15.7%  Median 19.8% 18.5%  75 th %ile 24.6% 22.1%  Maximum 38.6% 35.1% *Proportion of new opioid recipients who were classified as long-term recipients during the year following initiation. Open in new tab The proportion initiating LTOT varied widely across medical centers in 2011, with a minimum of 7.7%, maximum of 35.1%, and an IQR of 15.7% to 22.1%. LTOT proportions by medical center were highly correlated between 2004 and 2011 (Spearman coefficient 0.739, P  < 0.001), suggesting generally stable trends in the propensity to initiate LTOT relative to other medical centers. However, the change in LTOT proportion from 2004 to 2011 at the facility level was inversely correlated with the starting 2004 LTOT proportions (Spearman coefficient -0.566, P  < 0.001). Expressed categorically, medical centers in the lowest 2004 LTOT proportion quartile had slight increases over time (0.3%), while the remaining three quartiles showed stepwise declines in LTOT proportion: -1.3% for the second quartile, -2.6% for the third quartile, and -5.3% for the highest quartile. Opioid Course, Sensitivity Analyses We conducted a number of sensitivity analyses to assess the stability of the main findings concerning LTOT proportion estimates during 2004 and 2011 ( Table 1 ). We first examined the impact of additional patient selection criteria, which restricted the sample to individuals with a demonstrated history of VHA service use in the year prior to opioid initiation, based on VHA primary care and pharmacy use. The addition of more restrictive selection criteria yielded slight reductions in absolute LTOT proportion values, but did not change the main finding this proportion declined over time even as the absolute number of opioid recipients increased. For example, the most restrictive sample required both a VHA primary care encounter and regular VHA medication use during the year prior to opioid initiation. The LTOT proportion in this patient subset was 20.1% (43,304/215,346) in 2004 and 17.4% (51,033/294,122) in 2011. We further examined two alternative operational criteria for determining long-term opioid use from pharmacy refill data ( Table 4 ). While absolute LTOT proportion estimates varied across methods, from 9.4% to 26.0% in 2004, all methods yielded similar proportional decreases over time. Table 4 Sensitivity analyses examining alternative patient selection criteria and definitions for long-term opioid exposure . Long-term opioid therapy (LTOT) proportion * . Analysis . 2004 . 2011 . n/N (%) . n/N (%) . Primary Analysis 76,280/373,276 (20.4) 96,166/526,499 (18.3) Additional Patient Selection Restrictions  ≥1 VHA primary care visit 65,525/320,867 (20.4) 83,618/462,498 (18.1)  Regular VHA medication use 44,844/223,142 (20.1) 52,592/302,918 (17.4)  Both restrictions 43,304/215,346 (20.1) 51,033/294,122 (17.4) Alternative Opioid Exposure Definitions  Method 1 [ 15 , 16 ] 35,184/373,276 (9.4) 42,946/526,499 (8.2)  Method 2 [ 13 ] 96,980/373,276 (26.0) 129,078/526,499 (24.5) . Long-term opioid therapy (LTOT) proportion * . Analysis . 2004 . 2011 . n/N (%) . n/N (%) . Primary Analysis 76,280/373,276 (20.4) 96,166/526,499 (18.3) Additional Patient Selection Restrictions  ≥1 VHA primary care visit 65,525/320,867 (20.4) 83,618/462,498 (18.1)  Regular VHA medication use 44,844/223,142 (20.1) 52,592/302,918 (17.4)  Both restrictions 43,304/215,346 (20.1) 51,033/294,122 (17.4) Alternative Opioid Exposure Definitions  Method 1 [ 15 , 16 ] 35,184/373,276 (9.4) 42,946/526,499 (8.2)  Method 2 [ 13 ] 96,980/373,276 (26.0) 129,078/526,499 (24.5) *Long-term opioid therapy proportion defined as the proportion of new opioid recipients who became long-term opioid recipients within 1 year following initiation. Open in new tab Table 4 Sensitivity analyses examining alternative patient selection criteria and definitions for long-term opioid exposure . Long-term opioid therapy (LTOT) proportion * . Analysis . 2004 . 2011 . n/N (%) . n/N (%) . Primary Analysis 76,280/373,276 (20.4) 96,166/526,499 (18.3) Additional Patient Selection Restrictions  ≥1 VHA primary care visit 65,525/320,867 (20.4) 83,618/462,498 (18.1)  Regular VHA medication use 44,844/223,142 (20.1) 52,592/302,918 (17.4)  Both restrictions 43,304/215,346 (20.1) 51,033/294,122 (17.4) Alternative Opioid Exposure Definitions  Method 1 [ 15 , 16 ] 35,184/373,276 (9.4) 42,946/526,499 (8.2)  Method 2 [ 13 ] 96,980/373,276 (26.0) 129,078/526,499 (24.5) . Long-term opioid therapy (LTOT) proportion * . Analysis . 2004 . 2011 . n/N (%) . n/N (%) . Primary Analysis 76,280/373,276 (20.4) 96,166/526,499 (18.3) Additional Patient Selection Restrictions  ≥1 VHA primary care visit 65,525/320,867 (20.4) 83,618/462,498 (18.1)  Regular VHA medication use 44,844/223,142 (20.1) 52,592/302,918 (17.4)  Both restrictions 43,304/215,346 (20.1) 51,033/294,122 (17.4) Alternative Opioid Exposure Definitions  Method 1 [ 15 , 16 ] 35,184/373,276 (9.4) 42,946/526,499 (8.2)  Method 2 [ 13 ] 96,980/373,276 (26.0) 129,078/526,499 (24.5) *Long-term opioid therapy proportion defined as the proportion of new opioid recipients who became long-term opioid recipients within 1 year following initiation. Open in new tab Discussion Our study is the first to describe the 1-year treatment course following initial opioid prescription in an inclusive cohort of Veterans. In 2011, approximately one in eight veterans who received a new opioid prescription met the definition of long-term use in the initial fill period of 90 days, and one in five veterans met this definition at some point in the year following index prescription date. A prior study of adults in two large health care plans in the years 1997 through 2005 found that long-term episodes accounted for 19.9% of all opioid use episodes [ 13 ]. While not directly comparable, both studies found that long-term use comprised approximately 20% of new opioid prescriptions. A Norwegian study including only adults with chronic non-malignant pain found that 27% of new opioid recipients had persistent opioid use at 1 year following initiation [ 17 ]. Contrary to our hypothesis, LTOT proportion decreased over time, as did the median duration of opioid use. Both trends occurred in the context of an increasing number of veterans initiating opioid therapy. We thus found no evidence that new recipients of opioids in VHA have shown any increased proclivity toward long-term therapy, at least over the time period studied. Importantly, this finding remained consistent in sensitivity analyses using two alternative definitions for long-term opioid therapy. Demographic differences in the veteran population served during the time periods may be a factor in these findings: a greater percentage of veterans with incident opioid use in 2011 were 34 years of age or younger, and the percentage of female and black veterans increased. Rates of chronic pain diagnoses were stable; it is unlikely that changes in the prevalence of chronic pain drove the observed trends in long-term prescribing. A higher proportion of veterans with incident opioid prescription in 2008 had anxiety, depression, or substance abuse diagnoses. Mental health disorders have been associated with prescription of opioids and higher risk opioid use [ 18 ]; that the proportion of incident opioid recipients with subsequent LTOT declined may reflect a change in practice based on a growing sense of caution. Decreased LTOT proportion may reflect a trend among provider and patient dyads to reduce reliance on opioids for chronic pain, either by not initiating such therapy, or by ending a short-term trial if treatment is not progressing toward therapeutic goals, consistent with 2009 clinical guidelines [ 3 ]. Veterans in this clinical scenario might be classified as initial short-term recipients or intermediate recipients with no subsequent receipt. As illustrated in Figure 2 , receipt patterns set in the initial 3 months following index prescription tend to persist at 1 year, with a substantial proportion of intermediate recipients transitioning to long-term opioid therapy. Long-term patterns of opioid therapy are persistent [ 12 ]; that they appear to be established at index or within the first few months following new prescription in approximately one in five new recipients is additionally concerning given relatively high estimated rates of opioid misuse and abuse among patients with chronic pain or on LTOT [ 19 , 20 ]. These results are clinically significant as prescribers have the resources, through the electronic health records and state prescription drug monitoring databases, to assess a patient’s prior receipt pattern when making a decision about further prescription. If a prescriber is seeking to avoid initiation of long-term opioid therapy, an understanding of characteristic use patterns as defined through a cabinet supply approach may inform prescribing decisions following prescription for acute pain, or early in the course of opioid therapy. Our findings of increases in the proportion of initial prescriptions of 7 days or less, coincident with decreased LTOT proportion, imply that limiting prescription duration initially may curtail transition to LTOT. Prescribing patterns likely result from distinct clinical scenarios and provider and facility-resource factors; previously observed facility-level variation in high-risk medication prescribing to older adults persists after adjustment for patient characteristics [ 21 ]. Additional inquiry into characteristics of facilities with low or high proportions of veterans with LTOT initiation, rather than simply rates of opioid prescribing, may reveal targets for improvement in chronic pain treatment. For example, veteran-prescriber dyads may have intentionally initiated long-term use in areas where more effective multi-modal pain treatment is challenging to access. Further, distinguishing short- and long-term use patterns should prove valuable in focusing efforts to achieve a consistent and comprehensive national policy regarding opioid therapy for non-acute or non-malignant pain conditions, while maintaining access to safe and appropriate acute pain treatment. Initiatives aimed at decreasing LTOT, specifically high-dose long-term therapy in established opioid recipients [ 22 ], may have the collateral effect of decreasing new prescriptions [ 23 ]. A less desirable consequence of renewed caution about opioids in general might result in curtailed short-term prescribing for acute or cancer-related pain or in the palliative setting. For similar reasons, changes such as reclassification of tramadol as a controlled substance (August 18, 2014) and rescheduling of hydrocodone combination products (October 6, 2014) should be assessed using the cabinet supply approach for their effects on both short- and long-term prescribing. Limitations to our work include generalizability beyond a veteran population, although prior research suggests that opioid use is similarly persistent in the private sector [ 24 ]. Although we did characterize patients by chronic pain, use of identifier codes in administrative data requires assumptions regarding accuracy and clinical relevance [ 25 ] that were not verifiable. We can make no inferences about intended versus unintended, or appropriate versus inappropriate opioid use, regardless of duration. Characterizing patient health outcomes following opioid use was outside our scope. Relating opioid use patterns and pain outcomes is challenging, as opioids treat the symptom of pain and not the underlying cause. Thus, opioid cessation may reflect ineffectiveness of therapy or spontaneous resolution of pain; likewise, opioid continuation suggests pain is ongoing, but may or may not imply that opioids are effective in reducing pain. A recent longitudinal study illustrates these challenges [ 26 ]. Finally, we could not observe opioids obtained from sources outside VHA and our results may thus underestimate long-term prescribing. Conclusion Although the number of veterans initiating any opioid therapy was greater in 2011 than in 2004, the proportion who initiated LTOT, the median duration of opioid therapy, and days of opioids prescribed at index prescription all decreased. The increase in prevalence we previously reported [ 4 ] is best explained by an accumulation of long-term opioid recipients over time, rather than by an increased proclivity for long-term therapy among patients newly prescribed opioids. As LTOT continues to come under scrutiny, providers and policy-makers may consider strategies to minimize its initiation while maintaining access to opioids for short-term therapy. Our findings that one in five patients with incident opioid exposure demonstrates a pattern of long-term opioid receipt at 1 year serves to emphasize that these medications should be prescribed and renewed with caution. Data: Available to researchers with VA accreditation. Statistical Code: Available to interested readers by contacting Dr. Lund . Protocol: Available to interested readers by contacting Dr. Mosher . Funding sources: The work reported here was supported by the Department of Veterans Affairs, Veterans Health Administration and the Health Services Research and Development (HSR&D) Service through the Comprehensive Access and Delivery Research and Evaluation (CADRE) Center (CIN 13-412) and Career Development Award CDA 10-017 (BCL). Disclosure and conflicts of interest: This manuscript is not under review elsewhere and there is no prior publication of manuscript contents. The authors had full access to and take full responsibility for the integrity of the data. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. The authors report no conflicts of interest in regards to this study. References 1 The use of opioids for the treatment of chronic pain. A consensus statement from the American Academy of Pain Medicine and the American Pain Society . Clin J Pain 1997 ; 13 ( 1 ): 6 – 8 . Crossref Search ADS PubMed WorldCat 2 The management of chronic pain in older persons. AGS panel on chronic pain in older persons. American Geriatrics Society . Geriatrics 1998 ; 53(suppl 3) : S8 – 24 . PubMed OpenURL Placeholder Text WorldCat 3 Chou R Fanciullo GJ Fine PG , et al. . Clinical guidelines for the use of chronic opioid therapy in chronic noncancer pain . J Pain 2009 ; 10 ( 2 ): 113 – 30 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Mosher HJ Krebs EE Carrel M , et al. . Trends in prevalent and incident opioid receipt: An observational study in Veterans Health Administration 2004-2012 . J Gen Intern Med 2015 ; 30 ( 5 ): 597 – 604 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Compton WM Volkow ND. Major increases in opioid analgesic abuse in the United States: Concerns and strategies . Drug Alcohol Depend 2006 ; 81 ( 2 ): 103 – 7 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Sullivan MD Ballantyne JC. What are we treating with long-term opioid therapy? Arch Intern Med 2012 ; 172 ( 5 ): 433 – 4 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Von Korff M Kolodny A Deyo RA Chou R. Long-term opioid therapy reconsidered . Ann Intern Med 2011 ; 155 ( 5 ): 325 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Gomes T Mamdani MM Dhalla IA , et al. . The burden of premature opioid-related mortality . Addiction 2014 ; 109 ( 9 ): 1482 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Alexander GC Kruszewski SP Webster DW. Rethinking opioid prescribing to protect patient safety and public health . JAMA 2012 ; 308 ( 18 ): 1865 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Chou R Turner JA Devine EB , et al. . The effectiveness and risks of long-term opioid therapy for chronic pain: A systematic review for a National Institutes of Health Pathways to Prevention Workshop . Ann Intern Med 2015 ; 162 ( 4 ): 276 – 86 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Franklin GM , American Academy of Neurology . Opioids for chronic noncancer pain: A position paper of the American Academy of Neurology . Neurology 2014 ; 83 ( 14 ): 1277 – 84 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Vanderlip ER Sullivan MD Edlund MJ , et al. . National study of discontinuation of long-term opioid therapy among veterans . Pain 2014 ; 155 ( 12 ): 2673 – 9 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Von Korff M Saunders K Thomas Ray G , et al. . De facto long-term opioid therapy for noncancer pain . Clin J Pain 2008 ; 24 ( 6 ): 521 – 7 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Sullivan MD Edlund MJ Fan MY , et al. . Trends in use of opioids for non-cancer pain conditions 2000-2005 in commercial and Medicaid insurance plans: The TROUP study . Pain 2008 ; 138 ( 2 ): 440 – 9 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Sullivan MD Edlund MJ Fan MY , et al. . Risks for possible and probable opioid misuse among recipients of chronic opioid therapy in commercial and medicaid insurance plans: The TROUP Study . Pain 2010 ; 150 ( 2 ): 332 – 9 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Edlund MJ Martin BC Russo JE , et al. . The role of opioid prescription in incident opioid abuse and dependence among individuals with chronic noncancer pain: The role of opioid prescription . Clin J Pain 2014 ; 30 ( 7 ): 557 – 64 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 17 Mellbye A Karlstad O Skurtveit S Borchgrevink PC Fredheim OM. The duration and course of opioid therapy in patients with chronic non-malignant pain . Acta Anaesthesiol Scand 2016 ; 60 ( 1 ): 128 – 37 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Seal KH Shi Y Cohen G , et al. . Association of mental health disorders with prescription opioids and high-risk opioid use in US veterans of Iraq and Afghanistan . JAMA 2012 ; 307 ( 9 ): 940 – 7 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Vowles KE McEntee ML Julnes PS , et al. . Rates of opioid misuse, abuse, and addiction in chronic pain: A systematic review and data synthesis . Pain 2015 ; 156 ( 4 ): 569 – 76 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Boscarino JA Hoffman SN Han JJ. Opioid-use disorder among patients on long-term opioid therapy: Impact of final DSM-5 diagnostic criteria on prevalence and correlates . Subst Abuse Rehabil 2015 ; 6 : 83 – 91 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Gellad WF Good CB Amuan ME , et al. . Facility-level variation in potentially inappropriate prescribing for older veterans . J Am Geriatr Soc 2012 ; 60 ( 7 ): 1222 – 9 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Garg RK Fulton-Kehoe D Turner JA , et al. . Changes in opioid prescribing for Washington workers' compensation claimants after implementation of an opioid dosing guideline for chronic noncancer pain: 2004 to 2010 . J Pain 2013 ; 14 ( 12 ): 1620 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Westanmo A Marshall P Jones E Burns K Krebs EE. Opioid dose reduction in a VA health care system-implementation of a primary care population-level initiative . Pain Med 2015 ; 16 ( 5 ): 1019 – 26 . Google Scholar Crossref Search ADS PubMed WorldCat 24 A Nation in Pain. Express Scripts; December 9 2014 . http://lab.express-scripts.com/lab/insights/drug-safety-and-abuse/americas-pain-points (accessed March 3, 2016). 25 Mosher HJ Jiang L Vaughan Sarrazin MS , et al. . Prevalence and characteristics of hospitalized adults on chronic opioid therapy . J Hosp Med 2014 ; 9 ( 2 ): 82 – 7 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Turner JA Shortreed SM Saunders KW LeResche L Von Korff M. Association of levels of opioid use with pain and activity interference among patients initiating chronic opioid therapy: A longitudinal study . Pain 2015 . Epub ahead of print. Google Scholar OpenURL Placeholder Text WorldCat 27 Quan H Li B Couris CM , et al. . Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries . Am J Epidemiol 2011 ; 173 ( 6 ): 676 – 82 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Lund BC Chrischilles EA Carter BL Ernst ME Perry PJ. Development of a computer algorithm for defining an active drug list using an automated pharmacy database . J Clin Epidemiol 2003 ; 56 ( 8 ): 802 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat Published by Oxford University Press on behalf of the American Academy of Pain Medicine. 2016. This work is written by US Government employees and is in the public domain in the US. Published by Oxford University Press on behalf of the American Academy of Pain Medicine. 2016. This work is written by US Government employees and is in the public domain in the US.
TI - The 1-Year Treatment Course of New Opioid Recipients in Veterans Health Administration
JO - Pain Medicine
DO - 10.1093/pm/pnw058
DA - 2016-07-01
UR - https://www.deepdyve.com/lp/oxford-university-press/the-1-year-treatment-course-of-new-opioid-recipients-in-veterans-EnMLSB6tmY
SP - 1282
EP - 1291
VL - 17
IS - 7
DP - DeepDyve
ER -