INTRODUCTION

Costly hospitalizations related to the consequences of substance use are rising at an unprecedented rate.1,2,3 In 2012, there were an estimated 527,000 hospitalizations related to opioid use which cost more than $15 billion in medical care.2 Amphetamine-related hospitalizations increased more than 270% from 2008 to 2015.4 Hospitalizations related to the effects of alcohol increased more than two-fold over the last decade (9.1 to 22.7 per 100,000).5 Despite this, routine treatment of substance use disorders (SUD) among hospitalized patients is largely limited to a few academic institutions with addiction consultation services.6,7,8,9 While the more traditional addiction consultation model is promising, it may not be generalizable to many hospitals in part because there are not enough Addiction Medicine specialists to provide dedicated addiction treatment. Opportunities to leverage an existing hospitalist workforce could close the SUD treatment gap in the hospital setting.10,11,12

Hospital medicine is a growing medical specialty in community and academic settings.13,14,15 Hospitalists are uniquely positioned to fill a treatment gap for hospitalized patients with SUD by diagnosing and initiating treatment for SUD with linkage to post-discharge care. High-quality evidence supports the use of buprenorphine or methadone for opioid use disorder,16 naltrexone or acamprosate for alcohol use disorder,17 and nicotine replacement therapy, bupropion, or varenicline for tobacco use disorder.18 Many hospitalized patients with SUD are insured by state Medicaid programs4,19,20 which often cover the cost of these medications and outpatient-based addiction care.21,22 In-hospital initiation of medications for substance treatment could reduce patients leaving prior to treatment completion due to uncontrolled withdrawal symptoms23 and could possibly reduce length of hospital stay24 and hospital readmissions.25 Use of methadone and buprenorphine for withdrawal prevention among hospitalized patients is legal and maintains opioid tolerance to reduce risk of overdose following discharge.26,27,28 By training hospitalists in the practice of Addiction Medicine, hospitalists could narrow an addiction treatment gap in the hospital setting.

There is an urgent need to provide SUD treatment to hospitalized patients.10,11,12 We aimed to address the addiction treatment gap within our institution by training hospitalists in the practice of Addiction Medicine with a pathway to Addiction Medicine board certification with a long-term goal to provide sustainable, dedicated addiction treatment. We describe the development and implementation of a hospitalist-directed addiction consultation service with associated outcomes including the number of consultations completed, the prevalence of substances used, and the initiation of medications for SUD treatment.

METHODS

We developed and implemented a hospitalist-directed addiction consultation service (ACS) in a 650-bed university hospital located in Aurora, CO over a 2-year period. The ACS was spearheaded by one hospitalist dual-boarded in Addiction Medicine and Internal Medicine. The ACS development was informed by the Exploration, Preparation, Implementation, and Sustainment (EPIS) framework.29,30 The EPIS framework includes four well-defined phases to describe the implementation process, to identify outer system and inner organizational factors which affect the design and implementation process, and to bridge and leverage these factors to create a sustainable intervention (Fig. 1).

Fig. 1
figure 1

EPIS framework for a hospitalist-directed addiction consultation service

Exploration Phase (Year 1)

During the exploration phase, we identified a promising funding source to support hospitalist training in Addiction Medicine and dedicated hospitalist and social work salary support. We reviewed the American Board of Preventive Medicine requirements for the Practice Pathway to Addiction Medicine board certification outside of a traditional fellowship program31 which requires documentation of time spent in dedicated addiction care provision. We created a database of local outpatient-based opioid treatment programs (OBOT) and opioid treatment programs (OTP) for the provision of buprenorphine, methadone, injectable naltrexone, or injectable buprenorphine. We identified programs that offered inpatient or residential addiction treatment, and clinics that provided dual addiction and behavioral health treatment. We met with our county-assigned emergency/involuntary commitment coordinator to clarify processes required to legally pursue commitments for addiction-related illness. We reviewed the literature describing various ACS models6,7,9 and communicated with ACS Directors across the country to learn about their ACS structure, billing, and staffing models. Within our institution, we surveyed physicians to assess whether hospital-based addiction services were desired or needed in the hospital. We met with the Division Head of Hospital Medicine and the Director of the Care Management Department to identify potential resources for hospitalist and social work support.

Adoption Decision/Preparation Phase (Year 1 and Year 2)

We secured funding through a University of Colorado School of Medicine targeted Medicaid investment grant which distributed state and federal funds to expand access to behavioral health care for Health First Colorado Medicaid members. We partnered with a local OTP to identify a physician to enroll patients with opioid use disorder into a methadone program during hospitalization. This allowed for uptitration of methadone during prolonged hospitalizations with direct OTP linkage following discharge. Within the Division of Hospital Medicine, we recruited 11 hospitalists, 20% of hospitalists in our Division, to participate in our Addiction Medicine training program. This number was determined by our ability to offer financial compensation for time spent in educational and training activities (Table 1). We identified topic experts to present a 13-part lecture series covering the neurobiology of addiction, pharmacotherapy, trauma, perinatal substance use, toxicology, drug testing and interpretation, pain management and opioid use disorder, and co-occurring depression, anxiety, and PTSD among patients with SUD. We provided hospitalists with a 1-year membership to the American Society of Addiction Medicine (ASAM) which offers online learning modules including the 40-hour ASAM Fundamentals of Addiction Medicine course, and other online lectures and training activities addressing SUD. We scheduled half-day shifts when hospitalists could shadow an addiction-trained physician in clinical care over a 9-month period. We met with our University billing department to ensure appropriate billing procedures were followed to accurately capture addiction-related care provided as a consulting team outside of Hospital Medicine. Lastly, we hired two ACS social workers for weekday coverage.

Table 1 Education and Training Cost for Hospitalists

Active Implementation Phase (Year 2)

Social workers completed eight site visits to local OTPs, OBOTs, and other treatment centers to develop partnerships for linkage to care. Our physicians completed the lecture series, all became buprenorphine waivered, and hospitalists completed half-day shadow shifts with the Addiction Medicine physician. We advertised the ACS across hospital departments by speaking at various departmental meetings to increase awareness about our service including the intensive care units, transplant services, and surgical services.

Sustainment Phase (Year 2 and beyond)

Beginning in October 2019, 11 hospitalists began independently attending on the ACS with two ACS social workers. Consultation orders for an ACS were placed by the primary teams via the electronic health record. ACS hospitalists were encouraged to discuss complicated cases with our addiction physician via weekday telephone calls as needed. ACS social workers arranged follow-up appointments or treatment referrals at hospital discharge.

Addiction Treatment: Description of Services

ACS hospitalists and social workers meet with patients and conduct a comprehensive history and physical exam, including a DSM—5 assessment for SUD, inquiry into housing and transportation status, determination of last substance use, route of use, prior overdose, past or current mental health diagnoses, history of prior SUD treatment, and HIV and hepatitis C status, among others. ACS team members utilize motivational interviewing techniques to inquire about the patient’s personal substance use goals and provide patient-centered treatment options including pharmacotherapy, behavioral treatments, and harm reduction strategies. When appropriate, hospitalists initiate methadone or buprenorphine for withdrawal prevention and opioid use disorder. ACS hospitalists initiate medications for alcohol use disorder, including naltrexone or acamprosate, assist with complicated alcohol and benzodiazepine withdrawal, and provide recommendations for pain management in patients with opioid use disorder. ACS hospitalists manage post-operative pain control in patients on buprenorphine or methadone. ACS team members assist the primary team to facilitate emergency commitments for patients with life-threatening substance use. ACS social workers link patients to post-hospital discharge care including outpatient counseling, intensive outpatient programs, opioid treatment programs, and outpatient-based opioid treatment programs. Importantly, ACS team members advocate for people with SUD by providing and modeling compassionate, patient-centered care (Fig. 2).

Fig. 2
figure 2

Services provided by a hospitalist-directed addiction consultation service

Physician Metrics: Education, Training, and Clinical Practice Data

In year two, the active implementation phase, we tracked data on hours spent on shadow shifts, buprenorphine waiver certification, completion of online training modules, and lecture series attendance. Across years two and three, the sustainment phase, we tracked data on average ACS shifts completed over 365 days and hours spent on addiction care provision in general hospital medicine practice using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) code billing data.32 We asked hospitalists to track individual addiction-related hours spent in research or teaching in accordance with the Practice Pathway requirements. We tracked billing data including charges and collections for future planning to offset service costs. Notably, billing for dedicated addiction care in the hospital setting does not require addiction medicine board certification.

Patient and Encounter Metrics: Health Care Delivery Data

We obtained demographic data on all patients who received an ACS consultation between October 2019 and November 2020. We calculated frequencies of consultations and used billing data via ICD codes to identify substance-related diagnoses and commonly reported discharge diagnoses including alcohol, opioid, tobacco, methamphetamine/amphetamine, and cocaine-related diagnoses, alcoholic liver disease, alcoholic hepatitis, alcohol withdrawal, or infections related to injection, e.g., endocarditis, cellulitis, and abscesses, hepatitis C, HIV, or trauma, with concomitant SUD diagnosis. We used pharmacy data to determine frequency of medications for substance use treatment prescribed at hospital discharge, and we tracked initial buprenorphine prescriptions at discharge, with refills at 30, 60, and 90 days post-hospital discharge using data available in the electronic health record. We used administrative data to identify discharging hospital team, length of stay, and instances of patients leaving before treatment completion. We calculated descriptive statistics using proportions for categorical variables and means and medians for continuous variables. We obtained methadone retention data in monthly check-in meetings between the ACS director, ACS social workers, and OTP staff members.

The University of Colorado Institutional Review Board determined this project to be exempt and not human subjects research (COMIRB Protocol 20–2051).

RESULTS

Physician Metrics

During the implementation phase, hospitalists completed approximately 34 hours in shadow shifts, 13 hours in lectures, 8 hours in buprenorphine training, and at least 40 hours in online learning modules, for an average of 95 hours of education and didactic training per person. Over the 12 months spanning years two and three (sustainment phase), hospitalists spent, on average, 30 days as the ACS attending. From October 2019 to November 2020, hospitalist spent approximately 9.5 hours/month on addiction care provision in general hospital medicine practice, outside of time spent on the ACS. Hours tracked in addiction-related clinical care, research, and teaching are summed toward the 1920 hour required to qualify for the Addiction Medicine board exam via the Practice Pathway.

Patient and Encounter Metrics

From October 2019 to November 2020, the ACS completed 1620 consultations on 1350 unique patients. The mean age of patients was 44.3 years old (SD 12.9). Most patients were male (n = 898; 66.5%), White (n = 767; 56.8%), and non-Hispanic (n = 1042; 77.2%). Many patients were insured with Medicaid (n = 863; 63.9%) (Table 2).

Table 2 Patient-Level and Encounter-Level Descriptors of Patients Who Received an Addiction Consultation (October 2019–November 2020)

Among ACS encounters, alcohol was the most prevalent substance (n = 1279; 79%), followed by tobacco (979; 60.4%), methamphetamine/amphetamine (n = 494; 30.5%), and opioids (n = 400; 24.7%). Frequently reported discharge diagnoses included trauma (n = 1564; 96.5%), alcohol-related liver disease (n = 537; 33.1%), and cellulitis/abscess (n = 333; 20.6%). The mean length of stay was 8.4 days (SD ± 14 days). Documentation of “AMA” or leaving before treatment completion was noted in 120 (7.4%) encounters (Table 2).

Naltrexone was the most frequently prescribed medication at hospital discharge (n = 350; 21.6%), followed by acamprosate (n = 93; 5.7%) and buprenorphine (n = 77, 4.8%) (Table 2). Sixty-five patients received a new buprenorphine prescription at hospital discharge. Thirty days after discharge, 26 patients received buprenorphine refills, 60 days after discharge, 22 patients received refills, and 90 days after discharge, 24 patients received refills (Fig. 3). We completed 47 in-hospital methadone enrollments to a local OTP. To date, 19 of these patients remain in active treatment, 23 patients were lost to follow-up, 3 patients transferred care to a different OTP, one patient died, and one patient successfully discharged from the program (Fig. 3).

Fig. 3
figure 3

Buprenorphine receipt at discharge and refill data; methadone initiation and associated outcomes

DISCUSSION

The development and implementation of a hospitalist-directed ACS supported the education and training of 11 hospitalist physicians in the practice of Addiction Medicine and expanded SUD treatment to 1350 hospitalized patients over a 14-month period. Our training model addressed a local gap in hospital-based addiction care by educating hospitalists in Addiction Medicine thereby expanding the workforce of hospital-based addiction physicians with a future goal to expand our group of board-certified Addiction Medicine physicians.

We aimed to provide Addiction Medicine education and training to develop a core group of hospitalists to attend on the ACS. Recruiting hospitalists for ACS participation was facilitated by financial incentives and the potential for Addiction Medicine board certification via the Practice Pathway. Replication of a similar model requires a project champion with knowledge and clinical experience managing SUD, fastidious tracking of time spent in training, education, and clinical practice for hospitalists-in-training, as well as support from Hospital Medicine Division leaders who are interested in expanding their group’s scope of clinical work. While our hospitalists attend on the ACS for 7 or 14-day blocks, our ACS social workers provide continuity for long-term, hospitalized patients. They work closely with our ACS hospitalists to develop and coordinate psychosocial treatment plans including direct referrals to outpatient and residential SUD treatment programs. Over the last 2 years, our ACS social workers developed relationships with an extensive network of community addiction treatment providers for care linkage. They have a detailed knowledge of the facilities that offer injectable naltrexone or buprenorphine, which facilities are opioid treatment programs, which facilities offer treatment-on-demand, and which facilities offer both medications and counseling. Many hospitalized patients with SUD are not actively engaged in addiction treatment prior to hospitalization and benefit from direct treatment linkage following discharge.33 While our grant allowed us to rapidly grow and expand our service over a 2-year time period, it is possible that a similar model can be achieved on a slower scale without separate funds. In this case, we recommend that a hospitalist champion spearhead the process with a social worker colleague who commits to developing a network of treatment resources for direct patient referrals following discharge.

The majority of our ACS encounters included patients with alcohol use (80%), as compared to amphetamine/methamphetamine (30%) and opioid use (25%). In contrast, ACS in other parts of the United States (US) reported higher percentages of opioid-related encounters.7,20 We attributed this difference to several factors. First, the 2014 to 2017 data from the Colorado Behavioral Risk Factor Surveillance System (BRFSS) demonstrated that alcohol remained the most prevalently used substance in Colorado.34 Second, in 2018, opioid-involved overdose deaths in Colorado were approximately three times lower than opioid-involved overdose deaths in northeastern regions of the US (9.5/100,000 persons vs. 15 to >27/100,000 persons).35 Third, in 2016 in Colorado, methamphetamine-involved deaths surpassed heroin-involved deaths and remains the second most commonly involved drug in drug poisoning deaths after prescription opioids.36 While rates of substance use and death are not directly related to hospitalizations, correlations can be made between local use patterns and harms leading to hospitalization related to the medical consequences of substance use, including liver disease,37,38 infections,39,40 motor vehicle accidents, and other physical trauma.41,42 Differences in substance use prevalence in our community likely affected the number of alcohol-related ACS consultations relative to other substances noted in our institution as compared to other ACS around the US. Thus, our ACS consultations appear to reflect the prevalence of substance use in our community.

Patients initiated on buprenorphine or methadone for OUD treatment were linked to care following discharge. Approximately 40% of these patients continued receiving these medications at various time points following discharge. These findings are slightly lower than those reported by other ACS.7,43 Some ACS employ peer navigators with lived experience which may improve follow-up to care post-discharge.33,44,45 Our program was not funded to support this position nor did we collect information from patients regarding reasons for lack of follow-up. Future work will include obtaining a funding source to support a peer navigator position to contact patients after discharge to understand reasons for treatment interruption and to identify actionable plans to facilitate care linkage for patients interested in ongoing treatment.

LIMITATIONS

The study has important limitations. First, this was a single-site study spearheaded by one physician with Addiction Medicine training and may not be generalizable to other institutions. The development of the ACS was funded by an external grant aimed to expand treatment to Colorado Medicaid members which supported educational training time for physicians and social worker’s salaries. In hospitals where funding is unavailable to support training and education, it is notable that no special training is necessary to initiate medications for alcohol use disorder, training opportunities to become buprenorphine waivered are available online, and telementoring models exist to expand education and support for hospital-based addiction care.46 Most hospitals employ social workers who provide general medical care linkage following hospital discharge. Hospitalists and social workers could identify local addiction treatment resources to link patients to addiction treatment after discharge. Second, this was not an effectiveness study of an ACS. We did not include data on a control group of patients who did not receive an addiction consultation. The development and implementation of the ACS intended to fill an addiction treatment gap in our hospital and to provide evidence-based care. Lastly, under the Affordable Care Act, our state expanded Medicaid coverage to many previously uninsured people. Service reimbursement, medication coverage, and ongoing treatment following discharge are dependent upon a reliable payor source for medical cost coveraget.

CONCLUSION

The hospitalist-directed ACS model is a promising clinical initiative that could be implemented across hospitals to expand addiction treatment. Future research is needed to understand the challenges to disseminating this model into other settings, including community and rural hospitals, and to evaluate the intended and unintended effects of the broad implementation of this ACS model.