[Skip to Content]
[Skip to Content Landing]
Figure.
Hospital Variation in Adjusted Rate of New Opioid Use Within 7 Days of Hospital Discharge
Hospital Variation in Adjusted Rate of New Opioid Use Within 7 Days of Hospital Discharge

Adjusted rates of new opioid use were estimated among 2507 hospitals with at least 50 discharges in 2011. The average adjusted rate of new opioid use within 7 days of hospital discharge was 15.1%, with an interquartile range of 12.3% (25th percentile) to 17.4% (75th percentile) and an interdecile range of 10.5.% (10th percentile) to 20.0% (90th percentile).

Table 1.  
Characteristics of Medicare Beneficiaries With New Opioid Use Within 7 Days of Hospital Discharge
Characteristics of Medicare Beneficiaries With New Opioid Use Within 7 Days of Hospital Discharge
Table 2.  
Patient Characteristics Associated With New Opioid Use Within 7 Days of Hospital Dischargea
Patient Characteristics Associated With New Opioid Use Within 7 Days of Hospital Dischargea
Table 3.  
Association Between Hospital Characteristics and Adjusted Rate of New Opioid Use Within 7 Days of Hospital Discharge, Hospital-Level Analysisa
Association Between Hospital Characteristics and Adjusted Rate of New Opioid Use Within 7 Days of Hospital Discharge, Hospital-Level Analysisa
1.
Alam  A, Gomes  T, Zheng  H, Mamdani  MM, Juurlink  DN, Bell  CM.  Long-term analgesic use after low-risk surgery: a retrospective cohort study.  Arch Intern Med. 2012;172(5):425-430.PubMedGoogle ScholarCrossref
2.
Clarke  H, Soneji  N, Ko  DT, Yun  L, Wijeysundera  DN.  Rates and risk factors for prolonged opioid use after major surgery: population based cohort study.  BMJ. 2014;348:g1251.PubMedGoogle ScholarCrossref
3.
Calcaterra  SL, Yamashita  TE, Min  SJ, Keniston  A, Frank  JW, Binswanger  IA.  Opioid prescribing at hospital discharge contributes to chronic opioid use.  J Gen Intern Med. 2016;31(5):478-485.PubMedGoogle ScholarCrossref
4.
Dolin  SJ, Cashman  JN, Bland  JM.  Effectiveness of acute postoperative pain management: I. Evidence from published data.  Br J Anaesth. 2002;89(3):409-423.PubMedGoogle ScholarCrossref
5.
Kinney  MA, Hooten  WM, Cassivi  SD,  et al.  Chronic postthoracotomy pain and health-related quality of life.  Ann Thorac Surg. 2012;93(4):1242-1247.PubMedGoogle ScholarCrossref
6.
Katz  J, Seltzer  Z.  Transition from acute to chronic postsurgical pain: risk factors and protective factors.  Expert Rev Neurother. 2009;9(5):723-744.PubMedGoogle ScholarCrossref
7.
Mongardon  N, Pinton-Gonnet  C, Szekely  B, Michel-Cherqui  M, Dreyfus  JF, Fischler  M.  Assessment of chronic pain after thoracotomy: a 1-year prevalence study.  Clin J Pain. 2011;27(8):677-681.PubMedGoogle ScholarCrossref
8.
Rudd  RA, Aleshire  N, Zibbell  JE, Gladden  RM.  Increases in drug and opioid overdose deaths—United States, 2000-2014.  MMWR Morb Mortal Wkly Rep. 2016;64(50-51):1378-1382.PubMedGoogle ScholarCrossref
9.
Gupta  A, Daigle  S, Mojica  J, Hurley  RW.  Patient perception of pain care in hospitals in the United States.  J Pain Res. 2009;2:157-164.PubMedGoogle ScholarCrossref
10.
H.R. 4499: Promoting Responsible Opioid Prescribing Act of 2016.
11.
Centers for Medicare & Medicaid Services Research Data Assistance Center. Medicare managed care enrollees and the Medicare utilization files. 2011. http://www.resdac.org/resconnect/articles/114. Accessed November 9, 2015.
12.
Hospital Consumer Assessment of Healthcare Providers and Systems. Survey Instruments. 2015. http://www.hcahpsonline.org/surveyinstrument.aspx. Accessed December 8, 2015.
13.
Vila  H  Jr, Smith  RA, Augustyniak  MJ,  et al.  The efficacy and safety of pain management before and after implementation of hospital-wide pain management standards: is patient safety compromised by treatment based solely on numerical pain ratings?  Anesth Analg. 2005;101(2):474-480. table of contents.PubMedGoogle ScholarCrossref
14.
Frasco  PE, Sprung  J, Trentman  TL.  The impact of the joint commission for accreditation of healthcare organizations pain initiative on perioperative opiate consumption and recovery room length of stay.  Anesth Analg. 2005;100(1):162-168.PubMedGoogle ScholarCrossref
15.
Pletcher  MJ, Kertesz  SG, Kohn  MA, Gonzales  R.  Trends in opioid prescribing by race/ethnicity for patients seeking care in US emergency departments.  JAMA. 2008;299(1):70-78.PubMedGoogle ScholarCrossref
16.
Hanks  S.  The law of unintended consequences: when pain management leads to medication errors.  P T. 2008;33(7):420-425.PubMedGoogle Scholar
17.
Weech-Maldonado  R, Elliott  M, Pradhan  R, Schiller  C, Hall  A, Hays  RD.  Can hospital cultural competency reduce disparities in patient experiences with care?  Med Care. 2012;50(suppl):S48-S55.PubMedGoogle ScholarCrossref
18.
Saunders  KW, Dunn  KM, Merrill  JO,  et al.  Relationship of opioid use and dosage levels to fractures in older chronic pain patients.  J Gen Intern Med. 2010;25(4):310-315.PubMedGoogle ScholarCrossref
19.
Jena  AB, Goldman  D, Weaver  L, Karaca-Mandic  P.  Opioid prescribing by multiple providers in Medicare: retrospective observational study of insurance claims.  BMJ. 2014;348:g1393.PubMedGoogle ScholarCrossref
20.
Dunn  KM, Saunders  KW, Rutter  CM,  et al.  Opioid prescriptions for chronic pain and overdose: a cohort study.  Ann Intern Med. 2010;152(2):85-92.PubMedGoogle ScholarCrossref
21.
Centers for Disease Control and Prevention (CDC).  Vital signs: overdoses of prescription opioid pain relievers—United States, 1999-2008.  MMWR Morb Mortal Wkly Rep. 2011;60(43):1487-1492.PubMedGoogle Scholar
22.
Paulozzi  LJ, Budnitz  DS, Xi  Y.  Increasing deaths from opioid analgesics in the United States.  Pharmacoepidemiol Drug Saf. 2006;15(9):618-627.PubMedGoogle ScholarCrossref
23.
Solomon  DH, Rassen  JA, Glynn  RJ, Lee  J, Levin  R, Schneeweiss  S.  The comparative safety of analgesics in older adults with arthritis.  Arch Intern Med. 2010;170(22):1968-1976.PubMedGoogle ScholarCrossref
24.
Miller  M, Stürmer  T, Azrael  D, Levin  R, Solomon  DH.  Opioid analgesics and the risk of fractures in older adults with arthritis.  J Am Geriatr Soc. 2011;59(3):430-438.PubMedGoogle ScholarCrossref
Original Investigation
July 2016

Hospital Prescribing of Opioids to Medicare Beneficiaries

Author Affiliations
  • 1Department of Health Care Policy, Harvard Medical School, Boston, Massachusetts
  • 2Department of Medicine, Massachusetts General Hospital, Boston
  • 3National Bureau of Economic Research, Cambridge, Massachusetts
  • 4Leonard D. Schaeffer Center for Health Policy and Economics, University of Southern California, Los Angeles
  • 5RAND Corporation, Santa Monica, California
  • 6Division of Health Policy and Management, School of Public Health, University of Minnesota, Minneapolis
 

Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.

JAMA Intern Med. 2016;176(7):990-997. doi:10.1001/jamainternmed.2016.2737
Abstract

Importance  Use of opioids during and shortly after an acute hospitalization is warranted in some clinical settings. However, given the potential of opioids for short-term adverse events and long-term physiologic tolerance, it is important to understand the frequency of opioid prescribing at hospital discharge, hospital variation, and patient and hospital factors associated with opioid prescribing, which is currently unknown in the United States.

Objective  To estimate the frequency of opioid prescribing at hospital discharge among Medicare beneficiaries without an opioid prescription claim 60 days prior to hospitalization; to document hospital variation in prescribing; and to analyze patient and hospital factors associated with prescribing, including hospital average performance on pain-related Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) measures.

Design, Setting, and Participants  Analysis of pharmacy claims of a 20% random sample of Medicare beneficiaries hospitalized in 2011 without an opioid prescription claim in the 60 days before hospitalization.

Main Outcomes and Measures  Our main outcome was a new opioid claim within 7 days of hospital discharge. We estimated a multivariable linear probability model of patient factors associated with new opioid use and described hospital variation in adjusted rates of new opioid use. In multivariable linear regression analysis, we also analyzed hospital factors associated with average adjusted new opioid use at the hospital level, including the percentage of each hospital’s patients who reported that their pain during hospitalization was always well controlled in the 2011 HCAHPS surveys.

Results  Among 623 957 hospitalizations, 92 882 (14.9%) were associated with a new opioid claim. Among those hospitalizations with an associated opioid claim within 7 days of hospital discharge, 32 731 (42.5%) of 77 092 were associated with an opioid claim after 90 days postdischarge. Across 2512 hospitals, the average adjusted rate of new opioid use within 7 days of hospitalization was 15.1% (interquartile range, 12.3%-17.4%; interdecile range, 10.5%-20.0%). A hospital’s adjusted rate of new opioid use was modestly positively associated with the percentage of its inpatients reporting that their pain was always well managed (increase from 25th to the 75th percentile in the HCAHPS measure was associated with an absolute increase in new opioid use of 0.89 percentage points or a relative increase of 6.0%; P < .001).

Conclusions and Relevance  New opioid use after hospitalization is common among Medicare beneficiaries, with substantial variation across hospitals and a large proportion of patients using a prescription opioid 90 days after hospitalization. The degree to which observed hospital variation in short- and longer-term opioid use reflects variation in inappropriate prescribing at hospital discharge is unknown.

Introduction

Prescribing of opioids at the time of discharge from an acute hospitalization represents an important but underdescribed potential avenue through which previously opioid-naïve patients may develop long-term use.1-3 Use of opioids during and shortly after hospitalization is warranted in many clinical settings, for example, in patients with orthopedic injuries, those undergoing surgery, and those receiving palliative care. Failure to appropriately manage pain in these settings may adversely affect patient quality of life, delay discharge from the hospital, and in cases of surgery lead to interferences in postoperative rehabilitation and chronic postsurgical pain.4-7 However, prescribing of opioids is also associated with both short- and long-term risks including physical dependence. Adverse effects of opioid use are of particular concern given that deaths attributable to prescription opioids more than tripled in the United States during the 1999-2014 period.8

Limited evidence exists on prescribing of opioids at hospital discharge. Among opioid-naïve patients undergoing low-risk surgery in Ontario, Canada, 7% received opioids within 1 week of hospital discharge, and these patients were 44% more likely to use opioids long-term than patients who were not prescribed opioids in the first week postdischarge.1 In another Ontario-based study of patients undergoing major elective surgery, 49% of patients filled a prescription for an opioid after discharge and 3% continued to receive opioids after 90 days.2 In an analysis of opioid-naïve patients discharged at a US safety net hospital in 2011, 25% of patients were prescribed an opioid at discharge; the risk of opioid use 1 year later was 4-fold higher among patients with opioid receipt than among those without opioid receipt.3

The extent to which opioids are prescribed at hospital discharge, variation in prescribing across hospitals, and patient and hospital factors associated with opioid prescribing are unknown in the United States. Moreover, US hospitals are both measured and financially incentivized on the basis of patient satisfaction with hospital care through the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) initiative, an important domain of which includes patient perceptions of the quality of pain care while hospitalized.9 The possibility that this measure could incentivize both greater inpatient use of opioids and the prescribing of opioids at hospital discharge has been suggested in recently proposed legislation, the Promoting Responsible Opioid Prescribing Act of 2016,10 but the association between opioid prescribing rates at hospital discharge and hospital performance on pain-related HCAHPS measures has not been investigated.

Using data on prescription drug claims of a national random sample of Medicare beneficiaries who were hospitalized in 2011 and did not have a prescription claim for an opioid in the 60 days prior to hospitalization, we estimated the frequency with which individuals filled a prescription for an opioid within 1 week of hospital discharge, described variation across hospitals in rates of postdischarge prescription opioid use, and analyzed patient and hospital factors associated with postdischarge opioid use.

Box Section Ref ID

Key Points

  • Question How much do hospitals vary in opioid prescribing at hospital discharge, and what patient and hospital factors are associated with prescribing?

  • Findings In this observational study of Medicare beneficiaries hospitalized in 2011 without an opioid claim in the 60 days before hospitalization, 14.9% of discharges involved a new opioid claim within 7 days of discharge. A hospital’s adjusted rate of new opioid use was modestly positively associated with the hospital’s average performance on pain-related Hospital Consumer Assessment of Healthcare Providers and Systems measures.

  • Meaning New opioid use after hospitalization is common among Medicare beneficiaries, with substantial variation across hospitals.

Methods

This study was exempted from ethical review of studies with human participants by the University of Southern California.

Study Population

We identified acute short-stay hospitalizations of Medicare beneficiaries using the Centers for Medicare and Medicaid Services (CMS) MEDPAR (Medicare Provider Analysis and Review) hospital discharge data files for a 20% random sample of beneficiaries in 2011. MEDPAR files contain information on each acute hospital stay, including admission and discharge dates, International Classification of Diseases, Ninth Revision procedure and diagnosis codes, diagnosis-related group (DRG) codes, and hospital identity.

Because we were interested in measuring prescription opioid use of patients before and after hospitalization, we limited our analyses to discharges of beneficiaries who had 12 months of continuous Medicare Part D enrollment (N = 1 511 335 discharges). Because not all Medicare beneficiaries obtain drug coverage through the Part D program, our sample was representative of the Part D population, not Medicare overall. We excluded Medicare Advantage enrollees because not all hospitals submit claims to CMS for these patients.11 Our study population, therefore, included discharges of Medicare fee-for-service beneficiaries (n = 1 055 759). We required beneficiaries with a January or February 2011 discharge date to have Part D enrollment in 2010 because we needed to identify opioid use in a 60-day period before hospitalization (n = 1 042 120).

We merged discharges to American Hospital Association Annual Survey data that includes comprehensive information on US hospitals. We excluded 4866 discharges from hospitals with missing hospital identity, 1271 discharges from hospitals outside the United States, 11 478 discharges from hospitals that were not general medical or surgical hospitals, and 491 discharges from federal government hospitals. The resulting sample included 1 023 994 discharges. Finally, to identify new opioid use, we focused on hospitalized beneficiaries who did not fill a prescription for an opioid in the 60 days prior to the index hospitalization admission date (final sample, n = 623 957). A flow diagram of sample construction is provided in eFigure 1 in the Supplement.

We linked each discharge with beneficiary data from the Beneficiary Annual Summary Files, including age, sex, race (non-Hispanic white, non-Hispanic black, Hispanic, Asian, other including North American native), metropolitan residence, and 5-digit zip code of residence. Additional beneficiary data included whether a beneficiary received a low-income subsidy or received dual insurance coverage by Medicaid and Medicare at least 1 month of the year. Beneficiary files also included indicators of major comorbid conditions (Table 1).

Outcomes

Our primary outcome was whether a beneficiary filled a prescription for an opioid within 7 days of hospital discharge (“new opioid use postdischarge”). We also considered several secondary outcomes. First, to assess longer-term use, we identified discharges with at least a 90-day follow-up and identified those in which the patient filled an opioid after 90 days of discharge. Second, to analyze dosing differences across hospitals, we analyzed the proportion of discharges in which new opioid use within 7 days of hospital discharge was 20 morphine equivalents per day or higher. Third, to analyze how duration of opioid prescribing varied across hospitals, we analyzed the proportion of discharges in which new opioid use within 7 days of hospital discharge exceeded a 7-day supply. Finally, because our data did not specifically allow us to determine whether an opioid was prescribed at hospital discharge or shortly thereafter in the outpatient setting, we analyzed the proportion of discharges in which a new opioid prescription was filled within 3 days of discharge, assuming that such opioid use reflected filling of a prescription written at hospital discharge. We identified prescription opioids according to National Drug Code (complete, partial, and combination opioid agonists were included).

Statistical Analysis

We first computed the proportion of hospitalizations in which a new prescription opioid was filled within 7 days of hospital discharge and the proportion of hospitalizations in which a new prescription opioid continued to be used after 90 days postdischarge. We next estimated a multivariable, discharge-level linear probability model to study patient-level factors associated with new opioid use. The dependent variable was a binary outcome for whether a discharge involved a new prescription opioid claim within 7 days of discharge. Independent variables included patient characteristics, indicators for each of 585 DRGs, and hospital indicators. Patient characteristics included age, sex, race, Medicare-Medicaid dual eligibility and low-income subsidy for at least 1 month during the year, metropolitan residence indicator, indicators for each of 23 comorbid conditions, and patient 5-digit zip code median income. We reported changes in the adjusted probability of new prescription opioid use associated with each independent variable, with the exception of DRG given the large number of DRG variables. For categorical and binary variables, our estimates represented the difference in the probability of new opioid use for the corresponding category relative to the reference category (set at 0 for binary variables). To facilitate comparisons of how new opioid use was related to specific reasons for hospitalization, we reestimated the model replacing DRG indicators with medical and surgical major diagnostic categories (MDCs). These categories are formed by dividing all possible principal diagnoses into 25 mutually exclusive diagnosis areas. We further subdivided each MDC into surgical vs medical reason for hospitalization.

After analyzing patient-level predictors of new opioid use, we described hospital variation in adjusted rates of new opioid use. Using estimates from the multivariable model described above that included DRG, we predicted adjusted rates of new opioid use for each hospital, holding patient characteristics at their sample mean. We reported the distribution in adjusted rates of new opioid use by hospital for both primary and secondary measures of opioid use.

Finally, we analyzed hospital-level factors associated with new postdischarge opioid use by estimating a hospital-level multivariable linear regression of patient risk-adjusted rates of postdischarge opioid use as a function of hospital characteristics, including number of beds, staffing (number of attending physicians, residents, registered nurses, each per bed), for-profit status, system affiliation, teaching hospital status, rural status, and presence of an electronic health record. In addition, we included an independent variable from the 2011 HCAHPS survey on the percentage of patients reporting that their pain during hospitalization was always well controlled.12 We restricted our analysis to hospitals with at least 50 discharges in 2011 and weighted regressions by the number of observations per hospital. Analyses were performed for both primary and secondary measures of opioid use.

Analyses were conducted using Stata software, version 14 (StataCorp LP). The 95% confidence intervals (CIs) around reported means reflects 0.025 in each tail or P ≤ .05.

Results

Our sample included 623 957 hospitalizations with no opioid claim in the 60 days prior to hospital admission. Overall, 92 882 (14.9%) were associated with a new opioid claim within 7 days of discharge (Table 1). Among 77 092 of these 92 882 hospitalizations with 90-day follow-up, 32 731 (42.5%) were associated with an opioid claim after 90 days postdischarge.

While most new opioid use after hospital discharge was among beneficiaries 65 years or older, 8.2% of discharges with new opioid use involved beneficiaries younger than 45 years (Table 1). Most beneficiaries with new opioid use were female (57.9%). Non-Hispanic white and non-Hispanic black or African American beneficiaries accounted for 80.5% and 12.8% of the discharges, respectively. Overall, 41.7% of discharges with new opioid use involved beneficiaries who were dually eligible for Medicare and Medicaid. The most common comorbid conditions of beneficiaries with new opioid use postdischarge included hypertension (82.7%), hyperlipidemia (65%), anemia (55.7%), rheumatoid arthritis (49.4%), ischemic heart disease (49.3%), diabetes (40.2%), depression (31.5%), and congestive heart failure (30.4%).

Patient Factors Associated With New Postdischarge Opioid Use

After multivariable adjustment, new postdischarge opioid use was more common among younger beneficiaries (Table 2). For example, the unadjusted percentage of new opioid use for beneficiaries 85 years or older was 6.0%. Relative to these beneficiaries, the adjusted percentage of new opioid use was higher by 1.9 percentage points (95% CI, 1.7-2.2) for those aged 74 to 84 years, and by 6.8 percentage points (95% CI, 6.6-7.1) for those aged 65 to 74 years. Differences in new opioid use were not present according to patient race. New opioid use was more likely among beneficiaries who resided in rural areas, received low-income subsidies, or were dually eligible for Medicare and Medicaid. Surgical hospitalizations were more likely to be associated with new postdischarge opioid use, and among these, diseases of the hepatobiliary system and pancreas, digestive system, female reproductive system, and surgery related to burns had the largest associations (eTable 1 in the Supplement). Among medical hospitalizations, diseases of the musculoskeletal system and connective tissue (eg, nonoperative management of fractures) had among the largest associations with new opioid use after hospitalization. Similar relationships were observed in opioid use 90 days after hospital discharge (eTable 2 in the Supplement).

Hospital Variation in Opioid Use Postdischarge

New postdischarge opioid use varied across hospitals (Figure). Across 2507 hospitals, the average adjusted rate of new postdischarge opioid use was 15.1% (interquartile range, 12.3%-17.4%; interdecile range, 10.5%-20.0%). Similar variation was observed when we restricted analysis to new opioid prescriptions filled within 3 days of hospitalization (eFigure 2 in the Supplement). Among patients with new opioid use within 7 days of hospital discharge, large variation was also observed in opioid use 90 days postdischarge (eFigure 3 in the Supplement). Finally, large variation was observed across hospitals in the proportion of discharges in which new opioid use within 7 days of hospital discharge was 20 morphine equivalents per day or higher (eFigure 4 in the Supplement) and in which new opioid use exceeded a 7-day supply (eFigure 5 in the Supplement).

Hospital Factors Associated With New Opioid Use After Discharge

Hospital characteristics are described in eTable 3 in the Supplement. A hospital’s adjusted rate of new postdischarge opioid use was positively associated with its HCAHPS measure of the percentage of inpatients reporting that their pain was always well managed during hospitalization (Table 3). An increase from the 25th to the 75th percentile of the HCAHPS measure (from 66% to 71% of patients reporting their pain was always well managed, a 5 percentage-point increase) was associated with an absolute increase in new postdischarge opioid use of 0.89 percentage points (P < .001). This represents a relative increase of 6.0% over the average hospital’s percentage of new opioid use after discharge (from 15.1% to 16.0% of hospitalizations). Similar relationships were observed for our secondary opioid use measures (eTables 4 and 5 in the Supplement). For example, an increase from the 25th to the 75th percentile of the HCAHPS measure was associated with a 0.86 percentage-point absolute increase (7.1% relative increase compared with baseline of 12.2%) in the proportion of patients with an opioid claim within 7 days of discharge that was 20 morphine equivalents per day or higher (eTable 5 in the Supplement). New postdischarge opioid use was positively associated with the number of full-time registered nurses per bed and negatively associated with the number full-time resident physicians per bed. Government, rural, and system-affiliated hospitals had higher adjusted rates of new postdischarge opioid use.

Discussion

Analyzing prescription drug claims of a national random sample of Medicare beneficiaries who were hospitalized in 2011 and who were not prescribed an opioid within the 60 days prior to hospitalization, we found that approximately 15% of beneficiaries filled a new prescription for an opioid within 7 days of discharge and that the likelihood of postdischarge use varied across hospitals. For example, in the bottom decile of hospitals in terms of adjusted postdischarge opioid use, 10.5% of beneficiaries filled a new opioid prescription within 7 days of discharge, compared with 20.0% in the top decile of hospitals. Moreover, among beneficiaries with new opioid use within 7 days of discharge, 42.5% filled an opioid claim after 90 days postdischarge. We also found that hospitals that ranked more highly in terms of HCAHPS assessment of the quality of inpatient pain control had modestly higher rates of new postdischarge opioid use.

The prescribing of opioids during and after hospitalization is warranted in many clinical settings. Although we identified substantial variation across hospitals in new postdischarge opioid use, it is unknown whether this variation reflects differences in appropriate or inappropriate prescribing across hospitals. Distinguishing between these 2 possibilities is important, however, given that opioid prescribing at hospital discharge has in some settings been associated with increased risk of long-term use.1-3 Appropriate differences across hospitals in new postdischarge opioid use may stem from differences in average pain severity across hospitals or differences in the case mix of hospitalizations, which we attempted to account for through adjustment with DRG codes. In contrast, inappropriate differences across hospitals in new postdischarge opioid use may stem from systematic differences in practice patterns, which may be related to prescribing culture, other hospital characteristics, or possibly pay-for-performance measures related to the quality of inpatient pain management.13-16 Our finding that patients in the top decile of hospitals were twice as likely to fill a prescription for an opioid within 7 days of hospital discharge as patients in the bottom decile suggests that appropriate reasons for these observed differences, while important, may not fully explain the gap.

Our finding of a positive hospital-level relationship between HCAHPS pain management scores and the likelihood of postdischarge opioid use should be interpreted with caution. First, many patient, hospital, and regional factors may confound this relationship despite our attempts to account for patient characteristics and other hospital-level factors. For example, hospital performance on cultural competency measures is positively associated with performance on pain-related HCAHPS measures,17 implying that differences in cultural competency across hospitals may explain variation in opioid prescribing rather than hospital responses to pay-for-performance incentives. Second, our estimated associations were modest in size, suggesting that proposals to reduce hospital pay-for-performance incentives regarding inpatient pain management are unlikely to have large effects on opioid use.

Our findings could suggest that current outcome measurements of the adequacy of pain control should be coupled with process measurements on the appropriateness of opioid use or, put differently, that a central limitation of current pain satisfaction measures is their inability to account for appropriateness. To the extent that prescribing of opioids at discharge is aberrantly high in some hospitals, further detailed inquiry may be warranted into the appropriateness of this prescribing and whether opioid-prescribing rates could be safely reduced in those hospitals.

Our study had several limitations. First, we could not determine whether observed variation in opioid use after hospital discharge was related to appropriate or inappropriate prescribing variation. Second, given only 1 year of data, we could not study whether hospitals with higher adjusted rates of postdischarge opioid use had greater adverse events related to opioids, though many studies demonstrate links between opioid use and adverse opioid-related outcomes.18-22 Third, we focused on the Medicare population, specifically beneficiaries under fee-for-service. Medicare is an important population to consider, however, given that opioid use among the elderly and disabled is associated with higher mortality and other adverse outcomes.18,23,24 Fourth, although we found that postdischarge prescription opioid use was positively associated with hospital performance on HCAHPS pain measures, this association was not only modest in size but may have been confounded by unobserved hospital or population characteristics. Fifth, our primary outcome was opioid use within 7 days of a hospitalization. It is possible that this may reflect outpatient prescribing of opioids rather than opioids prescribed at the time of hospital discharge. Our findings, however, were unaffected by a narrower window of 3 days postdischarge.

Conclusions

Among Medicare patients without opioid use in the 60 days prior to hospitalization, prescribing of opioids at the time of hospital discharge is common, with substantial variation across hospitals and a large proportion of patients using a prescription opioid 90 days after hospitalization. The degree to which observed variation in postdischarge opioid use reflects inappropriate prescribing is unknown.

Back to top
Article Information

Accepted for Publication: April 25, 2016.

Corresponding Author: Anupam B. Jena, MD, PhD, Department of Health Care Policy, Harvard Medical School, 180 Longwood Ave, Boston, MA 02115 (jena@hcp.med.harvard.edu).

Conflict of Interest Disclosures: None reported.

Published Online: June 13, 2016. doi:10.1001/jamainternmed.2016.2737.

Author Contributions: Dr Jena had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Jena, Karaca-Mandic.

Acquisition, analysis, or interpretation of data: Jena, Goldman, Karaca-Mandic.

Drafting of the manuscript: Jena, Karaca-Mandic.

Critical revision of the manuscript for important intellectual content: Jena, Goldman.

Statistical analysis: Jena, Karaca-Mandic.

Obtained funding: Jena, Goldman, Karaca-Mandic.

Study supervision: Jena, Goldman, Karaca-Mandic.

Funding/Support: The authors report funding from the Office of the Director, National Institutes of Health (Early Independence Award, grant 1DP5OD017897-01 [Dr Jena]), National Institute on Aging (grants 5P01AG033559 [Dr Goldman] and 5K01AG036740 [Dr Karaca-Mandic]), and a University of Minnesota Academic Health Center Faculty Development Grant (Dr Karaca-Mandic).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
Alam  A, Gomes  T, Zheng  H, Mamdani  MM, Juurlink  DN, Bell  CM.  Long-term analgesic use after low-risk surgery: a retrospective cohort study.  Arch Intern Med. 2012;172(5):425-430.PubMedGoogle ScholarCrossref
2.
Clarke  H, Soneji  N, Ko  DT, Yun  L, Wijeysundera  DN.  Rates and risk factors for prolonged opioid use after major surgery: population based cohort study.  BMJ. 2014;348:g1251.PubMedGoogle ScholarCrossref
3.
Calcaterra  SL, Yamashita  TE, Min  SJ, Keniston  A, Frank  JW, Binswanger  IA.  Opioid prescribing at hospital discharge contributes to chronic opioid use.  J Gen Intern Med. 2016;31(5):478-485.PubMedGoogle ScholarCrossref
4.
Dolin  SJ, Cashman  JN, Bland  JM.  Effectiveness of acute postoperative pain management: I. Evidence from published data.  Br J Anaesth. 2002;89(3):409-423.PubMedGoogle ScholarCrossref
5.
Kinney  MA, Hooten  WM, Cassivi  SD,  et al.  Chronic postthoracotomy pain and health-related quality of life.  Ann Thorac Surg. 2012;93(4):1242-1247.PubMedGoogle ScholarCrossref
6.
Katz  J, Seltzer  Z.  Transition from acute to chronic postsurgical pain: risk factors and protective factors.  Expert Rev Neurother. 2009;9(5):723-744.PubMedGoogle ScholarCrossref
7.
Mongardon  N, Pinton-Gonnet  C, Szekely  B, Michel-Cherqui  M, Dreyfus  JF, Fischler  M.  Assessment of chronic pain after thoracotomy: a 1-year prevalence study.  Clin J Pain. 2011;27(8):677-681.PubMedGoogle ScholarCrossref
8.
Rudd  RA, Aleshire  N, Zibbell  JE, Gladden  RM.  Increases in drug and opioid overdose deaths—United States, 2000-2014.  MMWR Morb Mortal Wkly Rep. 2016;64(50-51):1378-1382.PubMedGoogle ScholarCrossref
9.
Gupta  A, Daigle  S, Mojica  J, Hurley  RW.  Patient perception of pain care in hospitals in the United States.  J Pain Res. 2009;2:157-164.PubMedGoogle ScholarCrossref
10.
H.R. 4499: Promoting Responsible Opioid Prescribing Act of 2016.
11.
Centers for Medicare & Medicaid Services Research Data Assistance Center. Medicare managed care enrollees and the Medicare utilization files. 2011. http://www.resdac.org/resconnect/articles/114. Accessed November 9, 2015.
12.
Hospital Consumer Assessment of Healthcare Providers and Systems. Survey Instruments. 2015. http://www.hcahpsonline.org/surveyinstrument.aspx. Accessed December 8, 2015.
13.
Vila  H  Jr, Smith  RA, Augustyniak  MJ,  et al.  The efficacy and safety of pain management before and after implementation of hospital-wide pain management standards: is patient safety compromised by treatment based solely on numerical pain ratings?  Anesth Analg. 2005;101(2):474-480. table of contents.PubMedGoogle ScholarCrossref
14.
Frasco  PE, Sprung  J, Trentman  TL.  The impact of the joint commission for accreditation of healthcare organizations pain initiative on perioperative opiate consumption and recovery room length of stay.  Anesth Analg. 2005;100(1):162-168.PubMedGoogle ScholarCrossref
15.
Pletcher  MJ, Kertesz  SG, Kohn  MA, Gonzales  R.  Trends in opioid prescribing by race/ethnicity for patients seeking care in US emergency departments.  JAMA. 2008;299(1):70-78.PubMedGoogle ScholarCrossref
16.
Hanks  S.  The law of unintended consequences: when pain management leads to medication errors.  P T. 2008;33(7):420-425.PubMedGoogle Scholar
17.
Weech-Maldonado  R, Elliott  M, Pradhan  R, Schiller  C, Hall  A, Hays  RD.  Can hospital cultural competency reduce disparities in patient experiences with care?  Med Care. 2012;50(suppl):S48-S55.PubMedGoogle ScholarCrossref
18.
Saunders  KW, Dunn  KM, Merrill  JO,  et al.  Relationship of opioid use and dosage levels to fractures in older chronic pain patients.  J Gen Intern Med. 2010;25(4):310-315.PubMedGoogle ScholarCrossref
19.
Jena  AB, Goldman  D, Weaver  L, Karaca-Mandic  P.  Opioid prescribing by multiple providers in Medicare: retrospective observational study of insurance claims.  BMJ. 2014;348:g1393.PubMedGoogle ScholarCrossref
20.
Dunn  KM, Saunders  KW, Rutter  CM,  et al.  Opioid prescriptions for chronic pain and overdose: a cohort study.  Ann Intern Med. 2010;152(2):85-92.PubMedGoogle ScholarCrossref
21.
Centers for Disease Control and Prevention (CDC).  Vital signs: overdoses of prescription opioid pain relievers—United States, 1999-2008.  MMWR Morb Mortal Wkly Rep. 2011;60(43):1487-1492.PubMedGoogle Scholar
22.
Paulozzi  LJ, Budnitz  DS, Xi  Y.  Increasing deaths from opioid analgesics in the United States.  Pharmacoepidemiol Drug Saf. 2006;15(9):618-627.PubMedGoogle ScholarCrossref
23.
Solomon  DH, Rassen  JA, Glynn  RJ, Lee  J, Levin  R, Schneeweiss  S.  The comparative safety of analgesics in older adults with arthritis.  Arch Intern Med. 2010;170(22):1968-1976.PubMedGoogle ScholarCrossref
24.
Miller  M, Stürmer  T, Azrael  D, Levin  R, Solomon  DH.  Opioid analgesics and the risk of fractures in older adults with arthritis.  J Am Geriatr Soc. 2011;59(3):430-438.PubMedGoogle ScholarCrossref
×