Background

Chronic pain, opioid prescribing, opioid use disorder (OUD), and opioid-related adverse events remain prevalent and problematic. Opioid agonist therapies, namely buprenorphine and methadone, are the best evidence-based strategies for treating and maintaining patients with OUD.1 Efforts to increase access to medication therapy for opioid use disorder (MOUD) are ongoing in Veterans Administration (VA).2 Moreover, both agents are therapies for chronic pain and may represent safer alternatives than other opioids that lack an indication for OUD. Within VA, there have been substantial efforts to increase access to and prescribing of buprenorphine for patients with OUD, and some buprenorphine products are formulary options for treatment of chronic pain.2,3,4 As such, the population of Veterans treated with buprenorphine will continue to increase. Hospital-based clinicians can anticipate seeing greater numbers of patients in hospital who are currently prescribed buprenorphine.

Few studies have explored current patterns of buprenorphine continuation among hospitalized veterans. One recent study found that 66% of patients experienced a perioperative hold on buprenorphine.5 This suggests variable practice, perhaps reflecting evolving evidence and recommendations for treatment of acute on chronic or chronic pain alone in the acutely ill hospitalized patient.6 More recently, consensus is growing that buprenorphine should be continued during the hospitalization.6 This may be particularly important as it affects retention in treatment 7, especially given evidence that stopping buprenorphine is associated with increased mortality.8, 9 Implementation of practice change can be slow, and an additional barrier to retention in therapy for buprenorphine in the context of hospitalization is the medication reconciliation process in general between inpatient and outpatient transitions of care, in which both intentional and inadvertent medication discontinuations can occur.10

We therefore sought to examine the current rate of in-hospital discontinuation of buprenorphine prescribed with the presumed indication of OUD (sublingual formulations) and for chronic pain (buccal and transdermal formulations) during medical and surgical hospitalizations in VA hospitals. As part of this analysis, we specifically hypothesized that administration of a full agonist opioid during the first 36 h of hospitalization would increase the likelihood of in-hospital buprenorphine discontinuation. In addition, we examined the association between in-hospital discontinuation of buprenorphine and subsequent retention in therapy. An understanding of the current approaches to inpatient buprenorphine use informs the need for and potential design of efforts to improve and monitor hospital-based buprenorphine use to address OUD and chronic pain.

Methods

Data Sources

National administrative data from the VA Corporate Data Warehouse were accessed via the VA Informatics and Computing Infrastructure. Inpatient admissions to a Veterans Health Administration (VHA) hospital were identified using inpatient encounter data. Inpatient administration of buprenorphine and full agonist opioids was ascertained using the inpatient bar code administration data. Outpatient pharmacy data was used to identify outpatient buprenorphine prescriptions to establish use status prior to admission and continuing use after discharge, as well as to determine which buprenorphine dosage form was prescribed. This study was approved by the University of Iowa Institutional Review Board and the Iowa City Veterans Administration Research and Development Committee.

Patients

Patients admitted to a VHA hospital between 10/1/2018 and 3/31/2020 were identified and the admitting service was classified as medical, surgical, or other. Active buprenorphine use at admission was defined as an outpatient prescription dispensed in the 60 days prior to admission, including sublingual formulations with or without naloxone, topical patches, buccal film, and long-acting injections. Unique patients were allowed to have more than one admission included in the analysis and were analyzed as independent events.

Buprenorphine Metrics

Buprenorphine use was assessed at several time periods following admission. Buprenorphine administration was recorded as a dichotomous indicator for three periods during hospitalization, the first 36 h of hospitalization, the final 36 h of hospitalization, and at any point during hospitalization. Administration of full agonist opioids, such as hydrocodone and oxycodone, was also recorded as dichotomous indicators during these timeframes. The primary measure of post-discharge buprenorphine receipt was any outpatient buprenorphine prescription dispensed between 1 day prior to discharge and 60 days following discharge. This 60-day time frame was selected to conservatively represent the intention for continued buprenorphine prescribing after discharge and allow patients sufficient time to reengage with their outpatient treatment providers, even if the inpatient admission represented a brief disruption in the continuity of buprenorphine therapy. The secondary measure of post-discharge buprenorphine receipt was the proportion of days covered over the 180 days following discharge. This was estimated using a cabinet supply approach for outpatient buprenorphine prescriptions dispensed within 1 day prior to discharge and 180 days following discharge, where any estimated remaining cabinet supply at the time of admission was carried forward.11

Analysis

Figure 1 illustrates the initial patient selection process and additional exclusions applied at progressive phases of the analysis. Patient characteristics were tabulated for patient admissions meeting initial selection criteria (N = 852). The frequency of buprenorphine administration during the inpatient admission was examined in the inpatient administration analysis (N = 830), after excluding patients who were receiving the long-acting injectable form buprenorphine prior to admission, as most individuals would not be expected to require a monthly injection during admission. Post-discharge continuation analyses excluded patients who were transferred to a non-medical or non-surgical service prior to discharge, as the focus of this analysis was on inpatient management by these services (N = 766). For both the inpatient administration and post-discharge continuation phases of analysis, primary analyses considered patients receiving all buprenorphine dosage forms together, along with stratified analyses for sublingual versus topical/buccal formulations due to differences in presumed indication. Sublingual dosage forms are approved by the US Food and Drug Administration for the treatment of OUD, whereas topical and buccal formulations are approved for the management of severe pain where daily, long-term opioid treatment is required.

Figure 1
figure 1

Patient selection and analysis flowchart.

Full size image

Contrasts in dichotomous buprenorphine metrics were performed using chi-squared tests and proportion of days covered metrics were contrasted using the Wilcoxon rank sum test. To account for the possibility of patients being intentionally switched from buprenorphine to an alternative OUD maintenance therapy during hospitalization, a secondary analysis was conducted excluding patients who received outpatient methadone or naltrexone at discharge or within the subsequent 60 days following discharge. A diagnosis of OUD was determined by the presence of at least one outpatient or inpatient encounter in the two years prior to admission with an International Classification of Diseases, Tenth Revision code of F11.x. All statistical tests were two tailed at the α = 0.05 level of significance.

Results

A total of 1278 unique inpatient hospitalizations among Veterans receiving buprenorphine prior to admission occurred during the study observation period (10/1/2018 to 3/31/2020). Among these, 852 were admissions to a medical (n = 757) or surgical (n = 95) service. The majority of the 426 excluded admissions were admissions to psychiatric (n = 288, 67.6%) and substance use disorder units (n = 45, 10.6%). The mean age of included patients was 61.6 years at admission and the majority were male (92.0%) and White (75.8%) (Table 1). Approximately one-quarter of admissions were rural-dwelling residents. The distribution of preadmission buprenorphine dosage forms was 401 (47.1%) receiving sublingual, 429 (50.4%) receiving patch or buccal formulations, and 22 (2.6%) receiving a long-acting injectable. In contrasting patients receiving sublingual buprenorphine versus patch or buccal dosage forms, the most notable differences were that the former were substantially younger on average (57.0 vs. 66.1 years) and more likely to have a documented OUD diagnosis (91.8% vs. 31.7%).

Table 1 Characteristics of Outpatient Buprenorphine Recipients Admitted to a VHA Medicine or Surgical Inpatient Service, Stratified by Dosage Form Received as an Outpatient Prior to Admission*
Full size table

After excluding patients receiving long-acting injectable buprenorphine prior to admission (n = 22), 364 (43.9%) of the remaining 830 patients received buprenorphine at some point during the medical or surgical portion of their hospital stay. This proportion varied by pre-admission formulation, from 60.6% (243/401) for sublingual buprenorphine, 51.0% (26/51) for buccal formulations, and 25.1% (95/378) for topical patches. Of those who received buprenorphine as inpatients, the majority (94.0%; 342/364) received the same formulation as they were receiving prior to admission.

As we hypothesized buprenorphine use during the inpatient hospitalization would be related to receipt of a full agonist opioid, we further examined administration of these medications during the first 36 h of hospital admission (Table 2). Across all dosage forms and services, 306 (36.8%) of 830 patients received buprenorphine during the first 36 h of admission. Among the 273 patients who received a full agonist opioid during this period, only 18.3% (n = 50) received buprenorphine. In contrast, 46.0% (n = 256) received buprenorphine among the 557 patients who did not receive a full agonist opioid (χ2 = 60.2; p < 0.001). This result was further stratified by medicine and surgical service and buprenorphine dosage form. In the case of sublingual formulations, administration during the first 36 h of hospitalization was 25.5% among patients who received a full opioid agonist and 67.1% among those who did not (Table 2). This difference was similar for both medicine and surgical services. In contrast, administration of buccal and topical formulations was 13.8% among those who received a full agonist opioid and just 22.1% among who did not. Administration of buccal and patch dosage forms was rare in the surgical setting (3.8%; 2/52), regardless of full agonist opioid receipt.

Table 2 Frequency of Inpatient Buprenorphine Administration Within the First 36 H of Admission, in Relation to Receipt of a Full Agonist Opioid and Stratified by Buprenorphine Dosage Form Prior to Admission and Hospital Service
Full size table

In the next phase of analysis, we examined buprenorphine administration during the final 36 h of hospitalization, among patients who did not receive buprenorphine during the first 36 h and whose total hospitalization lasted at least 72 h (n = 216). The latter part of the hospitalization represents a time when patients may be transitioned to their homegoing regimen, including cases where buprenorphine had previously been held due to concerns with coadministration with full agonist opioids or for other medical reasons. In total, only 35 of these patients (16.2%) were administered buprenorphine during the final 36 h of hospitalization. Stratified by opioid exposure during this period, buprenorphine administration rates were 19.1% (25/131) among patients who did not receive a full agonist opioid and 11.8% (10/85) among those who did.

The final phase of analysis examined outpatient buprenorphine dispensing following discharge, among the 766 patients discharged from a medical or surgical unit (Table 3). Overall, 569 (74.3%) patients received an outpatient buprenorphine prescription within the 60 days following discharge. Importantly, this proportion was significantly associated with inpatient buprenorphine receipt prior to discharge (χ2 = 83.5, p < 0.001). Among patients who had received buprenorphine during the final 36 h prior to discharge, subsequent outpatient buprenorphine receipt was observed in 94.0% (251/267), compared to 63.7% (318/499) among those not receiving buprenorphine during the final 36 h. These findings were similar when stratified by pre-admission buprenorphine formulation. Similar findings were also noted for the secondary post-discharge buprenorphine metric of proportion of days covered within 180 days of discharge. Overall, this metric was significantly higher among inpatient buprenorphine recipients than non-recipients (0.71 vs 0.47; p < 0.001) and observed for both sublingual and other formulations. Six patients who did not receive buprenorphine during the 60-day post-discharge period received either methadone (n = 5) or naltrexone (n = 1), which may have represented an intentional discontinuation of buprenorphine. After excluding these individuals in a secondary analysis, the primary comparison of post-discharge buprenorphine receipt between patients with buprenorphine during the final 36 h of hospitalization (251/266; 94.4%) versus those who did not (318/494; 64.4%) remained statistically significant (χ2 = 82.6, p < 0.001), as did all other comparisons examined in Table 3.

Table 3 Outpatient Buprenorphine Receipt Following Discharge and the Association with Inpatient Receipt During the Final 36 hours of Hospitalization
Full size table

Discussion

This retrospective observational study of buprenorphine administration in the hospital setting revealed that buprenorphine was frequently not continued, with 56.1% of patients receiving buprenorphine during their hospital stay. This finding compares to the 66% of VHA patients with MOUD undergoing inpatient surgery who had buprenorphine held around the time of surgery.5 Our study expands on this work in several important ways, as we include buprenorphine formulations indicated for OUD as well as for chronic pain, and we consider both medical and surgical patients. We found higher rates of inpatient continuation of sublingual buprenorphine with and without naloxone (59.5%), indicated for OUD treatment, than for buccal and topical formulations. This may reflect an awareness in some centers or among some clinicians of the importance of continuing MOUD treatment, as well as discomfort among inpatient clinicians with the role of buprenorphine in pain treatment.

There has been a growing emphasis on continuing buprenorphine during hospitalization and co-administering buprenorphine formulations and full agonist opioids to address pain acutely.6, 12 In practice, during the study period co-administration of buprenorphine and full agonist opioids was uncommon: just over one-third of patients in the study received full agonist opioids during the first 36 h of admission, and receipt of buprenorphine was less common among those who did. However, among patients not receiving full agonist opioids, a substantial proportion did not receive buprenorphine, implying that no full or partial agonist opioids were employed for treatment in a setting where patients had a high likelihood for new or ongoing pain. This finding raises concerns that uncertainty about managing these medications may be contributing to under-treatment of pain in the inpatient setting. It is possible that during this time (2018–2020) hospitalists may have lacked clarity on their ability to continue buprenorphine prescriptions without an X-waiver. Of note, this perceived barrier to care has now been removed as any prescriber with a drug enforcement administration (DEA) license can now prescribe or order buprenorphine.

The potential for hospitalization to lead to disruptions in medication therapy is well-recognized.13,14,15 In the current study, 1 in 5 patients admitted on sublingual buprenorphine, indicated for OUD, did not have a prescription for buprenorphine in the 60 days following discharge. While we are unable to determine from these data whether or not the discontinuation was purposeful and medically indicated, such discontinuation can be a risk factor for increased mortality.8 Further, we found that receipt of buprenorphine within the last 36 h of hospitalization, a time frame during which inpatient clinicians may be resuming home medications in anticipation of discharge, was strongly associated with greater likelihood of retention in treatment in the 60 days following discharge. Buprenorphine retention may be improved by emphasizing resumption of the pre-hospitalization regimen prior to leaving the hospital.

Our findings also suggest the potential utility of additional transitional care services for this patient population.16, 17 Within the VA, medication reconciliation services aim to prevent such gaps in care.18 However, the post-discharge period may represent a particularly vulnerable time for this patient population. Those in active treatment for OUD may benefit from additional early involvement and check-in regarding both pain management and OUD specific concerns.

The reasons to alter medications in the inpatient setting are many, ranging from patient factors such as the development of contraindications during acute illness (e.g., altered mental status), to perceived therapeutic failure in the outpatient setting, to determination that medication is no longer needed. A major limitation of our study is the inability to elucidate whether interruption of buprenorphine formulations during hospitalization is warranted or unwarranted. Moreover, our method of defining active buprenorphine at the time of admission may be imperfect: it is possible that buprenorphine therapy was intentionally discontinued prior to admission despite the presence of a recently dispensed prescription prior to admission. However, in cases where buprenorphine prescriptions were not intentionally discontinued, there may be an opportunity to shift inpatient treatment to be more concordant with current recommendations and emerging evidence about continuing buprenorphine during inpatient admissions.6 Hospitalists may benefit from additional training and resources to optimally care for patients on active treatments for OUD.

Additional limitations to our study include that, due to the nature of administrative data analyses, the indication for each buprenorphine prescription cannot be definitively determined, but rather is inferred based on route of administration. Further, these data only capture within-VA care, so it is possible that some patients who received buprenorphine from a VA provider prior to hospitalization may have received it from a non-VA source following discharge. Finally, as mentioned above, reasons for continuing or discontinuing buprenorphine prescriptions during hospitalization cannot be ascertained from administrative data; future qualitative work to elucidate provider decision-making is warranted.

Conclusions

Our findings point to relatively low rates of buprenorphine continuation during hospitalization and a significant subset who do not continue or reinitiate buprenorphine therapy following discharge. As recommendations for perioperative and inpatient management of buprenorphine coalescence around continuation, an understanding of the current approaches to inpatient buprenorphine use informs the need for and potential design of efforts to improve and monitor hospital-based buprenorphine use to address OUD and chronic pain. Patients in active treatment for OUD may benefit from additional follow-up to coordinate OUD and pain-related care on discharge.