Abstract
Background:
Any opioid-related hospitalization is an indicator of opioid-related harm and should ideally trigger carefully monitored decreases in opioid prescribing following inpatient stays in many, if not most, cases. However, past studies on opioid prescribing following hospitalizations have largely been limited to overdose related visits. It is unclear whether prescribing is different for other opioid related indications such as opioid dependence and abuse and how that may compare to hospitalizations for overdose.
Objective:
To examine opioid-prescribing patterns before and after opioid-related hospitalizations for all opioid-related indications, not limited to overdose.
Research Design:
Retrospective cohort analysis of Veterans Health Administration (VHA) administrative claims from 2011-2014.
Subjects:
VHA patients who were hospitalized between fiscal years 2011-2014 and had at least one prescription opioid medication filled through the VHA pharmacy prior to their hospitalization.
Measures:
Opioid dispensing trajectories after hospitalization by opioid-related indication (i.e., opioid dependence and/or abuse versus overdose) compared to prescribing patterns for non-opioid-related hospitalizations.
Results:
Overall, opioid dosage dropped significantly (66% for dependence/abuse, 42% for overdose, and 3% for non-opioid diagnoses, p<0.001) across all three categories when comparing dose 57-63 days following admission to 57-63 days prior to hospitalization. However, 47% of patients remained on the same dose or increased their opioid dose at 60 days following an opioid-related hospitalization. After adjusting for covariates, patients with a primary diagnosis of dependence/abuse had higher odds of having their dose discontinued compared to those with overdose: OR 2.17 (1.19-3.96). Patients with admissions for opioid dependence and/or abuse had a statistically significant higher prevalence of depression, post-traumatic stress disorder, anxiety and substance use disorders compared to those with an opioid overdose hospitalization.
Conclusions:
Opioid prescribing and patient risk factors before and after opioid-related hospitalizations vary by indication for hospitalization. To reduce costs and morbidity associated with opioid-related hospitalizations, opioid de-intensification efforts need to be tailored to indication for hospitalization.
Introduction:
Recent evidence has shown that opioid use is associated with over 1 million opioid-related inpatient and emergency department hospitalizations per year in the United States, costing upward of $15 billion.1–3 Surprisingly, far more of these opioid-related hospitalizations are for opioid dependence and abuse, even though opioid-related overdoses are a persistent concern.4
Non-overdose visits recorded as dependence and abuse may be precursors to more serious adverse events, such as overdose. Furthermore, they may also include visits related to pain management such as acute pain following procedures or new medical conditions such as abdominal pain, worsening of chronic pain, or withdrawal symptoms. A recent study using a large, nationally representative commercially insured population showed that only 12% of these opioid-related hospitalizations between 2010-2014 were specifically coded as admissions for overdose, suggesting that the non-overdose opioid-related hospitalizations make up the large majority of visits.5 Moreover, these acute hospitalizations can be stressful for inpatient providers and are associated with more frequent readmissions and an increased risk of death.6 Therefore, it is critical to understand opioid prescribing patterns and predictors of changes in prescribing following all opioid-related hospitalizations, not just those related to overdoses, to get a sense of where we may intervene to prevent future opioid-related harms for both patients and providers.
Any opioid-related hospitalization is an indicator of opioid-related harm and should ideally trigger carefully monitored decreases in opioid prescribing following inpatient stays in many, if not most, cases.7 Recent evidence has shown that 91% of patients receive prescription opioids in the year following an acute hospital admission for a non-fatal overdose8 and 22% of patients receive prescription opioids at some point within thirty days of an opioid-related hospitalization.5 However, there is little knowledge to describe how prescribing patterns may change leading up to and after the hospital admission and how this may vary based upon the type of opioid diagnosis, such as overdose versus dependence and/or abuse. Therefore, using US national administrative data from the US Veterans Health Administration (VHA), we aimed to describe opioid dosage patterns following opioid-related inpatient hospitalizations for opioid dependence and/or abuse versus visits for overdose and compared to patients with non-opioid-related- hospitalizations. We also examined patient factors, including physical and mental health comorbidities, as predictors of prescribing pattern following acute hospitalizations, with a focus on dosage increases.
Methods:
Sample:
Using administrative data, we analyzed VHA patients who were hospitalized between fiscal years 2011-2014 and had at least one prescription opioid medication filled through the VA pharmacy within six months prior to the hospitalization.6 Because Veterans can receive care within and outside of the VA health system, we aimed to capture a sample of patients who were recent users of the VHA health system and pharmacy benefits to better describe opioid prescribing trajectories within the VHA. Prescription data came from the VHA’s Corporate Data Warehouse. Patients were excluded if they were in hospice, long-term care (e.g., nursing homes), inpatient psychiatric admissions, or if they had an opioid-related admission or emergency department (ED) visit in the year prior to the index hospitalization, allowing us to look at each patient’s first opioid-related hospitalization in a year. We also excluded 19 patients who did not have a diagnosis code associated with their hospitalization and a small percentage of prescription fills where dose could not be ascertained from the pharmacy record. We stratified the sample by related diagnosis during hospitalization (e.g., opioid dependence and/or abuse versus overdose versus non-opioid-related hospitalization) based upon the diagnosis billing codes from the International Statistical Classification of Diseases, Ninth Revision (ICD-9) (Supplemental Table 1). 2,5 The VHA database only allowed for specification of a primary diagnosis, defined as the diagnosis responsible for the greatest length of stay, and up to 24 secondary diagnosis codes. We limited our sample to hospitalizations with on opioid-related diagnosis in the primary (first) diagnosis position to be more specific in identifying hospitalizations that were primarily related to opioids rather than other medical indications. Visits coded as non-opioid-related hospitalizations did not list opioid dependence, abuse, or overdose codes as a reason for admission.
Opioid Dosage:
We calculated the average morphine milligram equivalents (MME) per day for 7-day intervals across diagnosis categories for the 6 months prior to and following hospitalization using established methods.9 This measurement of dosage reflects the maximum daily dose prescribed and not necessarily the actual amount consumed. Average dosage was calculated using only outpatient pharmacy claims; inpatient opioid prescriptions were not captured. We excluded buprenorphine and methadone prescribed for opioid substitution therapy due to their uniqueness in indication and inconsistent morphine equivalency.10 After attributing each prescription to specific dates based on fill dates and days supply, each patient’s total maximum daily dose for each day of the study observation period was calculated by adding the daily doses of all fills that covered that particular day. We grouped patients by their average daily dosage into categories of 0, >0 to <50 oral MME, 50-90 MME, and >90 MME at the time period of 57-63 days prior to hospitalization, based on the use of this categorization in prior studies9 and the CDC Guideline for Prescribing Opioids for Chronic Pain.11 These changes in dose were categorized as dose discontinued, decreased by at least 10% of total daily MME, initiated or increased by at least 10% of total daily MME (considered to represent a minimum meaningful dose change based upon tapering guidelines12), or stayed the same (i.e., within the +/− 10% margin). We mapped overall trajectories for 6 months before and after hospitalization. However, when conducting significance testing comparing post-hospitalization prescribing to pre-hospitalization prescribing, we focused on changes from the period of 57-63 days before to 57-63 days after the hospitalizations to allow for a reasonable length of time to attempt an opioid taper to potentially prevent future adverse events and to account for any brief increase in prescribing due to acute pain. In addition, we compared the pre-hospitalization dose at 57-63 days to 7 days before and after hospitalization to measure acute changes in prescribing surrounding the hospitalization to account for potential artifact from the way administrative data labels inpatient or outpatient prescriptions. For example, a patient may receive opioid medication the day of hospital discharge from an inpatient pharmacy that may not be visible in outpatient pharmacy dispensing data. Therefore, we aimed to compare more stable outpatient prescriptions 57-63 days before and after the hospitalization.
Other Measures:
We obtained demographic variables and ICD-9 diagnosis codes from patient records. Demographic variables included age, gender, race (white, black, other) and Hispanic ethnicity. All diagnoses reflected whether the patient had been diagnosed with each specific condition in the year prior to their hospitalization. We also calculated the Elixhauser Comorbidity index for every patient13 and created separate pain indicators for headache, neuropathy, injuries and acute pain, cancer pain and chronic bodily pains (Supplemental Table 2) .9 We did not limit to diagnoses present only during the hospitalization as we hypothesized that overall comorbidity, chronic mental health conditions, and co-morbid pain diagnosis could influence opioid prescribing in the 6 months leading up to and following the hospitalization and may not be reflected in the diagnoses during the acute hospitalization.
Data Analyses:
We examined bivariate associations of patient characteristics by whether the patient had an opioid-related hospitalization or non-opioid-related hospitalization and across opioid-related hospitalization indications using χ2 tests for categorical variables and Wilcoxon rank sum tests for continuous variables. We also plotted the average MME by diagnosis category (i.e., hospitalization related to opioid dependence and/or abuse, overdose, or non-opioid-related hospitalization) using 7-day increments for the 6 months prior to and following hospitalization. We compared average weekly dose across diagnosis categories 57-63 days prior to and following hospitalization using the Wilcoxon rank sum test and compared the change in average weekly dose during these time periods using paired t-tests. We then examined the percent of patients that discontinued opioids or decreased, increased, or stayed on the same dose in the 7 days prior to hospitalization, 7 days after hospitalization, and 57-63 days after hospitalization compared to the reference mean dose at 63-57 days prior to hospitalization. These changes in dose were stratified by whether the patient had a hospitalization for dependence/abuse versus overdose. Finally, we used a multinomial logistic model to identify predictors for changes based on these two categories between 57-63 days before and after the hospitalizations. Analyses were conducted using SAS Enterprise Guide 7.1 (SAS Institute, Cary, North Carolina). The threshold for statistical significance was p < .05. The Ann Arbor VA’s Institutional Review Board approved study protocols.
Sensitivity Analysis:
We also conducted a sensitivity analysis using the adjusted multinomial logistic model to see if there were differences in dose changes when comparing any and not just primary diagnosis of overdose to dependence/abuse. In addition, we conducted a sensitivity analysis to see if there were differences in change in dose using the adjusted multinomial logistic model comparing opioid-related hospitalizations to non-opioid-related hospitalizations.
Results:
In a cohort of VHA pharmacy users with prior opioid prescriptions, 865 unique patients had a primary diagnosis code for opioid-related hospitalizations, 592 for opioid dependence/abuse and 273 for overdose (Figure 1). In comparison, 484,648 patients had a non-opioid-related hospitalization from FY 2011-2014. Patients with hospitalizations for opioid dependence and/or abuse were younger than those with an opioid overdose (p<0.001; Table 1).
Figure 1:
Cohort Flow Diagram
* Items were excluded in sequential order
Table 1:
Characteristics of the Patient Cohort (N=485,513)
| Patient Characteristics | Opioid-Dependence or Abuse Hospitalization n=592 (0.12%) N (%) |
Opioid-Overdose Hospitalization n=273 (0.06%) N (%) |
Non- Opioid Related Hospitalization n=484,648 (99.8%) N (%) |
|---|---|---|---|
| Age, y a, d | |||
| <35 | 141 (23.9) | 20 (7.3) | 15,961 (3.3) |
| 35-55 | 227 (38.4) | 75 (27.5) | 96,994 (20.0) |
| 56-75 | 215 (36.4) | 156 (57.1) | 293,758 (60.6) |
| ≥76 | 8 (1.4) | 22 (8.1) | 77,934 (16.1) |
| Gender | |||
| Male | 547 (92.4) | 246 (90.1) | 452,103 (93.3) |
| Female | 45 (7.6) | 27 (9.9) | 32,545 (6.7) |
| Racee | |||
| White | 460 (77.7) | 216 (79.1) | 355,218 (73.3) |
| Black | 101 (17.1) | 40 (14.7) | 92,618 (19.1) |
| Other/Unknown | 31 (5.2) | 17 (6.2) | 36,812 (7.6) |
| Ethnicityf | |||
| Hispanic | 23 (3.9) | 9 (3.3) | 25,173 (5.2) |
| Not Hispanic | 569 (96.1) | 264 (96.7) | 459,475 (94.8) |
| Mean Opioid Dosage (MME) pre-hospitalization* c,d | |||
| 0 | 192 (32.4) | 70 (25.6) | 255,496 (52.7) |
| < 50 | 233 (39.4) | 100 (36.6) | 185,540 (38.3) |
| 50-90 | 60 (10.1) | 44 (16.1) | 24,694 (5.1) |
| >90 | 107 (18.1) | 59 (21.6) | 18,918 (3.9) |
| Diagnoses | |||
| Depressionb, d | 397 (67.1) | 154 (56.4) | 178,037 (36.7) |
| SMId | 121 (20.4) | 66 (24.2) | 47,028 (9.7) |
| PTSDc,d | 232 (39.2) | 83 (30.4) | 92,225 (19.0) |
| Anxietya, f | 229 (38.7) | 61 (22.3) | 73,323 (15.1) |
| SUDa, d | 503 (85.0) | 103 (37.7) | 94,250 (19.5) |
| Acute Painf | 187 (31.6) | 94 (34.4) | 139,126 (28.7) |
| Chronic Paind | 534 (90.2) | 248 (90.8) | 397,670 (82.1) |
| Headached | 87 (14.7) | 43 (15.8) | 50,326 (10.4) |
| Neuropathyc,e | 34 (5.7) | 29 (10.6) | 50,264 (10.4) |
| Cancera,d |
68 (11.5) | 62 (22.7) | 171,143 (35.3) |
| Elixhauser Comorbidity Indexa, d (mean, SD) | −5.4 (5.6) | 2.00 (8.3) | 4.7 (8.5) |
56-63 days prior to stay
MME= morphine milligram equivalents, SMI=Serious Mental Illness, PTSD= Post-traumatic Stress Disorder, SUD= Substance Use Disorder
p<.0001 across opioid-related hospitalizations
p<.01 across opioid-related hospitalizations
p<.05 across opioid-related hospitalizations
p<.0001 opioid vs non-opioid-related hospitalizations
p<.01 opioid vs non-opioid-related hospitalizations
p<.05 opioid vs non-opioid-related hospitalizations
Patients with opioid-related hospitalizations were much more likely to have been receiving 50 or more MME at 2 months prior to admission compared to patients with a non-opioid-related hospitalization (24.9% versus 9.0%, p<0.0001). Patients with admissions for opioid dependence and/or abuse had a statistically significant greater prevalence of depression, post-traumatic stress disorder, anxiety and substance use disorders) compared to those with an opioid overdose. Patients with an opioid-related diagnosis had significantly greater chronic pain compared to those with a non-opioid-related diagnosis.
Overall, average opioid dosage dropped significantly (66% for dependence/abuse, 42% for overdose, and 3% for non-opioid diagnoses, p<0.001) across all three categories when comparing dose 57-63 days following admission compared to 57-63 days prior to prescription (Figure 2). However, 47% of patients remained on the same dose or increased their opioid dose at 60 days following an opioid-related hospitalization. Pre-hospitalization opioid doses were much higher for patients with opioid-related hospitalizations compared to those with a non-opioid related diagnosis. (p<.0001). Following the hospitalization, those with dependence/abuse had similar opioid prescribing trajectories to those with a non-opioid related visit whereas those with overdoses had higher opioid doses following hospitalization compared to both those with dependence/abuse and non-opioid related visits. (p<.0001).
Figure 2: Average Weekly Opioid Dose 6 months Pre and Post Hospitalization by Primary Diagnosis*.
*Dosage in the acute hospitalization period may be artificially low due to undocumented pharmacotherapies during the inpatient period
Seventy-six percent of patients with an overdose received an opioid fill within the 6 months following hospitalization (Figure 3). In comparison, only 59% of patients with dependence/abuse received an opioid fill following hospitalization. 18% of overdose patients had a dose increase at 57-63 days following hospitalization compared to 13.3% of patients with dependence/abuse.
Figure 3: Change In Opioid Dose Compared to 63-57 Days Prior to Hospitalization*.
*A dose decrease or increase was defined by at least a 10% change in the total MME
After adjusting for covariates, patients with a primary diagnosis of dependence/abuse had higher odds of having their dose discontinued compared to those with overdose (OR 2.17 (1.19-3.96). (Table 2). For the sensitivity analysis, we found that patients with an overdose diagnosis in any recorded primary or secondary diagnosis codes were more likely to have their dose decreased but not discontinued compared to those with dependence/abuse in any order location in the diagnosis codes (Supplemental Table 3). Compared to non-opioid -related hospitalization, those with dependence and abuse and overdose were at higher odds of having their dose decreased or discontinued (Supplemental Table 4). However, those with overdose were also at higher odds of having their dose increased compared to those with a non-opioid -related hospitalization (OR 1.46 (1.02-2.09)).
Table 2:
Predictors of Opioid Dose Change Compared to Dose Staying the Same Following Hospitalization
| Patient Characteristics | Dose Discontinued (N=308) Versus stayed the Same (N=279) |
Dose Decreased (N=149) Versus stayed the Same (N=279) |
Dose Increased (N=128) Versus stayed the Same N=279 |
|---|---|---|---|
| OR (95% CI) | OR (95% CI) | OR (95% CI) | |
| Diagnosis at Hospitalization | |||
| Opioid-Related | |||
| Dependence/Abuse | 2.17 (1.19-3.96) | 0.95 (0.49-1.86) | 1.19 (0.67-2.12) |
| Overdose | ref | ref | ref |
| Age, y | |||
| <35 | 1.85 (0.82-4.17) | 1.04 (0.38-2.83) | 0.71 (0.34-1.48) |
| 35-55 | 1.10 (0.63-1.91) | 0.85 (0.45-1.60) | 1.28 (0.75-2.19) |
| 56-75 | ref | ref | ref |
| ≥76 | 1.49 (0.38-5.86) | 2.22 (0.56-8.80) | 0.31 (0.06-1.71) |
| Gender | |||
| Male | ref | ref | ref |
| Female | 0.90 (0.37-2.19) | 0.87 (0.32-2.39) | 0.82 (0.36-1.88) |
| Race | |||
| White | ref | ref | ref |
| Black | 0.98 (0.51-1.89) | 0.71 (0.32-1.57) | 1.04 (0.56-1.92) |
| Other/ Unknown | 1.25 (0.43-3.62) | 1.40 (0.44-4.45) | 1.22 (0.42-3.56) |
| Ethnicity | |||
| Hispanic/Latino | 1.79 (0.41-7.75) | 0.61 (0.09-4.29) | 1.04 (0.30-3.57) |
| Not Hispanic/Latino | ref | ref | ref |
| Daily Opioid Dosage (MME) before hospitalization* (56-63 days before) | |||
| 0 | * | * | 0.23 (0.13-0.39) |
| >0 to <50 | ref | ref | ref |
| 50 to 90 | 0.74 (0.35-1.57) | 3.26 (1.45-7.30) | 0.58 (0.22-1.49) |
| >90 | 0.49 (0.27-0.90) | 2.51 (1.28-4.89) | 0.19 (0.08-0.83) |
| Mental Health Diagnoses | |||
| Depression | 0.94 (0.56-1.58) | 0.86 (0.48-1.53) | 1.02 (0.63-1.65) |
| Serious Mental Illness | 1.39 (0.74-2.60) | 1.82 (0.90-3.69) | 0.90 (0.51-1.59) |
| PTSD | 0.93 (0.56-1.56) | 0.88 (0.49-1.58) | 0.80 (0.48-1.32) |
| Anxiety | 1.38 (0.81-2.34) | 1.14 (0.61-2.11) | 1.42 (0.86-2.34) |
| SUD | 1.83 (1.00-3.33) | 1.04 (0.54-2.00) | 0.73 (0.40-1.31) |
| Pain Diagnoses | |||
| Acute Pain | 1.40 (0.82-2.38) | 1.60 (0.89-2.89) | 1.08 (0.66-1.78) |
| Chronic Pain | 0.75 (0.30-1.87) | 1.67 (0.50-5.57) | 1.21 (0.57-2.54) |
| Headache | 0.83 (0.43-1.62) | 0.94 (0.44-2.00) | 1.42 (0.74-2.73) |
| Neuropathy | 1.63 (0.60-4.43) | 2.58 (0.92-7.24) | 2.17 (0.89-5.30) |
| Cancer | 1.19 (0.60-2.37) | 1.10 (0.52-2.31) | 0.93 (0.45-1.91) |
| Elixhauser Comorbidity Index | 1.01 (0.97-1.06) | 1.02 (0.98-1.07) | 1.04 (1.00-1.08) |
comparing dose 57-63 days post-hospitalization to 57-63 days pre-hospitalization
MME= morphine milligram equivalents, SMI=Serious Mental Illness, PTSD= Post-traumatic Stress Disorder, SUD= Substance Use Disorder
Discussion:
This study sought to characterize patterns of opioid usage before and after opioid-related hospitalizations based upon indication for hospitalization. Overall, opioid-related hospitalizations were associated with greater decreases in opioid prescriptions compared to non-opioid-related hospitalizations with visits coded as dependence and/or abuse having the greatest drops in dosage compared to overdose and non-opioid related visits. However, 14.8% of patients still had dosage increases sustained at two months following these events, and patients with overdose had increased odds of having their dose increased compared to others. In addition, patients hospitalized for dependence and/or abuse were younger and had greater prevalence of mental health comorbidities than those with overdose visits. Mental health comorbidities were also among the most prevalent primary diagnosis codes at visits also coded for dependence and/or abuse. These findings reflect the heterogeneity in patients experiencing opioid-related harms and the variety in prescribing trajectories after these visits by indication for hospitalization.
These results add to a recent study evaluating prescribing trajectories after non-fatal overdoses and show that patients hospitalized for non-overdose-related opioid hospitalizations (e.g., dependence, abuse) have different opioid prescribing trajectories from those with opioid poisoning.8 Our results demonstrate that patients with hospitalizations for dependence/abuse are much more likely to have their dose discontinued compared to those with overdose. This is surprising as overdose is often a more life threatening indication for hospitalization, and would be logically hypothesized to be more likely than other opioid hospitalizations to prompt discontinuation as a strategy to reduce risk of future overdose. Future studies should evaluate how these two cohorts may have variable interactions with the health system following hospitalizations, potentially leading to the observed differences in prescribing patterns. For example, a cohort with dependence/abuse may have more frequent interactions associated with dose changes due to their other mental health comorbidities, compared to those with overdose.
Overdose, by itself, is a rare event and much of the costs of opioid-related hospitalizations are attributed to dependence and abuse.4 In our sample, there were more than twice the number of hospitalizations for dependence/abuse compared to overdose. The indication for hospitalization was captured using ICD classifications, similar to prior studies. These classifications are based on codes that use terminology including the words “dependence,” “abuse,” and “poisoning.” Clinically, opioid poisoning is often easily visible in the form of a drug overdose, and billing codes for overdoses have been previously validated. 14–16 However, hospitalizations classified as dependence and/or abuse may be less specific and, to our knowledge, have not been validated. It is clear from our data that these two cohorts have distinct prescribing patterns before and after hospitalization. This in itself warrants further investigation with more detailed chart review to understand how these codes are assigned.
This study has some key limitations. We captured diagnosis based upon administrative codes, which may not accurately capture true diagnosis and in some systems (although not the VHA) may be captured by billing personnel after chart review rather than direct coding by clinicians. 25 In addition, we measured opioid prescriptions filled within the VHA, but not prescriptions patients received outside of the VHA.26 However, we did limit our sample to existing pharmacy users to better capture patients that likely rely primarily on the VHA pharmacy. Moreover, our results evaluated a VHA cohort, who were predominantly Veteran men. National data suggests that overdose rates are climbing rapidly in women,27 and although rates are generally higher in men. Our results may not be generalizable to this growing population of women with opioid-related harms. The VHA initiatives to reduce opioid prescribing, the integrated electronic health record, and centralized pharmacy within the VHA likely facilitate more informed prescribing due to less fragmentation in care and greater provider knowledge of hospitalizations and other opioid prescriptions.28 These results may nevertheless provide useful insights into what may be observed with more robust adherence to similar programs enhancing the transference of medical record and pharmacy data across health systems, as well as what may be observed internationally in countries with national health systems such as the UK. Finally our study focused exclusively on patients receiving opioid medications for analgesia prior to the hospitalization and evaluated the trends in their opioid prescribing in the post-hospitalization period based upon their diagnosis category. Future studies will need to evaluate whether patients with codes for opioid dependence and/or abuse warrant and appropriately receive medication-assisted treatments or other treatments for OUD following hospitalization. 29
In conclusion, this study found that opioid-prescribing trajectories following opioid-related hospitalizations varied by indication for hospitalization and patient populations across these indications have very different comorbidities. The results of this study indicate a need to distinguish between patients at risk for dependence and/or abuse-related hospitalizations versus overdose when examining ways to reduce many adverse opioid-related hospitalizations that include, but are not limited to, opioid overdose.
Supplementary Material
Acknowledgments
Funding Source: Dr. Haffajee’s work on this article was supported by funding from the National Center for Advancing Translational Sciences of the National Institutes of Health (grant #KL2TR002241). Dr. Haffajee’s and Dr. Bohnert’s work on this article was supported by the Centers for Disease Control and Prevention for the University of Michigan Injury Prevention Center (grant #3R49CE002099-05S1).
Footnotes
Competing Interests: The authors have no conflicts of interest.
References:
- 1.Ronan MV, Herzig SJ. Hospitalizations Related To Opioid Abuse/Dependence And Associated Serious Infections Increased Sharply, 2002-12. Health Aff (Millwood) 2016;35(5):832–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Weiss AJ, Elixhauser A, Barrett ML, et al. Opioid-Related Inpatient Stays and Emergency Department Visits by State, 2009-2014: Statistical Brief #219 Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Rockville (MD), 2016 [Google Scholar]
- 3.Vivolo-Kantor AM, Seth P, Gladden RM, et al. Vital Signs: Trends in Emergency Department Visits for Suspected Opioid Overdoses - United States, July 2016-September 2017. MMWR Morb Mortal Wkly Rep 2018;67(9):279–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Song Z Mortality Quadrupled Among Opioid-Driven Hospitalizations, Notably Within Lower-Income And Disabled White Populations. Health Aff (Millwood) 2017;36(12):2054–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Naeger S, Ali MM, Mutter R, et al. Prescriptions Filled Following an Opioid-Related Hospitalization. Psychiatr Serv 2016;67(11):1262–64. [DOI] [PubMed] [Google Scholar]
- 6.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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lagisetty P, Bohnert A. Web Exclusives. Annals for Hospitalists Inpatient Notes - The Opioid Epidemic-What’s a Hospitalist to Do? Ann Intern Med 2017;167(2):HO2–HO3. [DOI] [PubMed] [Google Scholar]
- 8.Larochelle MR, Liebschutz JM, Zhang F, et al. Opioid Prescribing After Nonfatal Overdose and Association With Repeated Overdose. Ann Intern Med 2016;165(5):376–7. [DOI] [PubMed] [Google Scholar]
- 9.Bohnert AS, Valenstein M, Bair MJ, et al. Association between opioid prescribing patterns and opioid overdose-related deaths. JAMA 2011;305(13):1315–21. [DOI] [PubMed] [Google Scholar]
- 10.Lawlor PG, Turner KS, Hanson J, et al. Dose ratio between morphine and methadone in patients with cancer pain: a retrospective study. Cancer 1998;82(6):1167–73. [PubMed] [Google Scholar]
- 11.Dowell D, Tamara M.Chou, Roger. CDC Guideline for Prescribing Opioids for Chronic Pain--United States, 2016. JAMA: Journal of the American Medical Association 2016;315(15):1624–45 22p. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Berna C, Kulich RJ, Rathmell JP. Tapering Long-term Opioid Therapy in Chronic Noncancer Pain: Evidence and Recommendations for Everyday Practice. Mayo Clin Proc 2015;90(6):828–42. [DOI] [PubMed] [Google Scholar]
- 13.Elixhauser A, Steiner C, Harris DR, et al. Comorbidity measures for use with administrative data. Med Care 1998;36(1):8–27. [DOI] [PubMed] [Google Scholar]
- 14.Slavova S, Bunn TL, Talbert J. Drug overdose surveillance using hospital discharge data. Public Health Rep 2014;129(5):437–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Green CA, Perrin NA, Janoff SL, et al. Assessing the accuracy of opioid overdose and poisoning codes in diagnostic information from electronic health records, claims data, and death records. Pharmacoepidemiol Drug Saf 2017;26(5):509–17. [DOI] [PubMed] [Google Scholar]
- 16.Rowe C, Vittinghoff E, Santos GM, et al. Performance Measures of Diagnostic Codes for Detecting Opioid Overdose in the Emergency Department. Acad Emerg Med 2017;24(4):475–83. [DOI] [PubMed] [Google Scholar]
- 17.Park TW, Saitz R, Ganoczy D, et al. Benzodiazepine prescribing patterns and deaths from drug overdose among US veterans receiving opioid analgesics: case-cohort study. BMJ 2015;350:h2698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jones CM, Mack KA, Paulozzi LJ. Pharmaceutical overdose deaths, United States, 2010. JAMA 2013;309(7):657–9. [DOI] [PubMed] [Google Scholar]
- 19.Bohnert AS, Logan JE, Ganoczy D, et al. A Detailed Exploration Into the Association of Prescribed Opioid Dosage and Overdose Deaths Among Patients With Chronic Pain. Med Care 2016;54(5):435–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Weiss AJ, Bailey MK, O’Malley L, et al. Patient Characteristics of Opioid-Related Inpatient Stays and Emergency Department Visits Nationally and by State, 2014: Statistical Brief #224 Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Rockville (MD), 2017. [Google Scholar]
- 21.Pletcher MJ, Kertesz SG, Kohn MA, et al. Trends in opioid prescribing by race/ethnicity for patients seeking care in US emergency departments. JAMA 2008;299(1):70–8. [DOI] [PubMed] [Google Scholar]
- 22.Harrison JM, Lagisetty P, Sites BD, et al. Trends in Prescription Pain Medication Use by Race/Ethnicity Among US Adults With Noncancer Pain, 2000–2015. Am J Public Health 2018;108(6):788–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Singhal A, Tien YY, Hsia RY. Racial-Ethnic Disparities in Opioid Prescriptions at Emergency Department Visits for Conditions Commonly Associated with Prescription Drug Abuse. PLoS One 2016;11(8):e0159224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Gaither JR, Gordon K, Crystal S, et al. Racial disparities in discontinuation of long-term opioid therapy following illicit drug use among black and white patients. Drug Alcohol Depend 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Heslin KC, Owens PL, Karaca Z, et al. Trends in Opioid-related Inpatient Stays Shifted After the US Transitioned to ICD-10-CM Diagnosis Coding in 2015. Med Care 2017;55(11):918–23. [DOI] [PubMed] [Google Scholar]
- 26.Gellad WF, Zhao X, Thorpe CT, et al. Overlapping buprenorphine, opioid, and benzodiazepine prescriptions among veterans dually enrolled in Department of Veterans Affairs and Medicare Part D. Subst Abus 2017;38(1):22–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.VanHouten JP, Rudd RA, Ballesteros MF, et al. Drug Overdose Deaths Among Women Aged 30-64 Years - United States, 1999-2017. MMWR Morb Mortal Wkly Rep 2019;68(1):1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gellad WF, Good CB, Shulkin DJ. Addressing the Opioid Epidemic in the United States: Lessons From the Department of Veterans Affairs. JAMA Intern Med 2017;177(5):611–12. [DOI] [PubMed] [Google Scholar]
- 29.Naeger S, Mutter R, Ali MM, et al. Post-Discharge Treatment Engagement Among Patients with an Opioid-Use Disorder. J Subst Abuse Treat 2016;69:64–71. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.