In Hospital IM Naltrexone: A Pilot Study

1. Hypotheses and Specific Aims: This proof of concept comparative effectiveness study seeks to pilot the administration of IM naltrexone among hospitalized adults with Diagnostic and Stastistical Manual-5 (DSM-5) diagnosis of alcohol use disorder at the University of Colorado Hospital and examine its acceptability and feasibility among patients and healthcare providers.

Study Aim 1: Test the feasibility and acceptability of recruiting, enrolling, and randomizing hospitalized patients, age 18 to 65 years old, with alcohol use disorder to treatment with in-hospital administration of IM naltrexone (n=30) versus oral naltrexone (n=60), both with linkage to outpatient follow-up.

Hypothesis 1: Outcomes such as recruitment rate, adverse events, and provider perception of the intervention's feasibility and acceptability in their clinical practice will help identify key lessons for a future comparative effectiveness study.

Study Aim 2: Estimate the combined 30-day readmission rate plus 30-day emergency department encounter rate for participants randomized to IM naltrexone vs. oral naltrexone.

Hypothesis 2: Combined 30-day readmission rate and emergency department encounter rate will be lower for IM naltrexone participants compared to oral naltrexone participants.

Study Aim 3: Estimate the 90-day treatment linkage rate for participants randomized to IM naltrexone vs. oral naltrexone.

Hypothesis 3: The 90-day treatment linkage rates will be higher for IM naltrexone participants compared to oral naltrexone participants.

Study Design and Research Methods Study Design: This study is pilot proof-of-concept, open-label comparative effectiveness trial of IM naltrexone vs. oral naltrexone among hospitalized adults with alcohol use disorder. Study outcomes include 1) an assessment of the feasibility to recruit, enroll, and randomize hospitalized patients to IM naltrexone (n=30) or oral naltrexone (n=60) and 2) an assessment of provider's perceived feasibility and acceptability to deliver the intervention. Comparative effectiveness outcomes include assessing the 1) combined 30-day readmission rate and 30-day emergency department encounter rate and 2) 90-day treatment linkage rates for participants randomized to IM naltrexone vs. oral naltrexone.

Study Site: The study will be conducted at the University of Colorado Hospital (UCH) where there is a weekday Addiction Consultation Service (ACS) who provides medication-assisted treatment for alcohol use disorder, including naltrexone, with facilitated linkage to post-discharge addiction care.

Recruitment: Eligible participants will be identified by an ACS team member through daily review of the ACS consultation list, focusing on consultation orders for "alcohol", e.g., alcohol withdrawal, and the sequela of alcohol use, trauma in the setting of alcohol intoxication, acute encephalopathy, and/or seizures, delirium tremens related to alcohol withdrawal. ACS attendings will complete routine addiction medicine care, including history taking, performing a physical exam, reviewing laboratory values, and a discussion of the assessment and plan, while considering the patient's alcohol use goals e.g., cut back, abstinence, no change in use, etc. If the patient meets criteria for an alcohol use disorder, the ACS attending will consider the use of naltrexone as medication-assisted treatment. If the patient expresses an interest in naltrexone, the attending will contact the one of the two addiction medicine team social workers or peer navigator, who is not involved in the patient's care, will approach the patient to inquire if the participants are interested in participating in the study. If the patient agrees, the participants will be consented and enrolled into the study and will be randomized to 2:1 to oral naltrexone or IM naltrexone. The ACS team member will alert the ACS attending who will either prescribe oral naltrexone to start during hospitalization with a 30-day prescription at discharge or who will administer IM naltrexone prior to hospital discharge. If the patient declines enrollment, the ACS attending will prescribe a 30-day supply of oral naltrexone. If a patient is interested in receiving IM naltrexone, the participants will be scheduled a follow-up appointment with a local clinic that provides this medication in the outpatient setting. All patients will be offered a follow-up appointment for ongoing treatment for their alcohol use disorder regardless of study enrollment.

Intervention: Participants will be randomized 2:1 to of oral naltrexone (n=60) oral naltrexone or IM naltrexone (n=30). Oral naltrexone is a once daily medication. IM naltrexone is dosed once monthly and is injected into the gluteal tissue. All participants will be scheduled a follow-up appointment with an addiction treatment provider or primary care provider between 30 to 90 days post hospital discharge, unless the participants decline the appointment.

The cost of oral naltrexone is covered by most insurance plans, including Colorado Medicaid, and is readily available in most pharmacies. For uninsured patients, the ACS team will work with the Atrium pharmacy and team social worker to obtain a voucher for medication coverage. This is the usual process for medication payment when a patient is unable to afford their medication co-pay or the cost of the medication at UCH. IM naltrexone is produced by Alkermes. Due to bundled insurance payments for hospitalizations, administration of IM naltrexone is prohibited for in-hospital administration at UCH (and at many hospitals) due to lack of reimbursement for services and product. In light of the potential treatment gap for hospitalized patients with alcohol use disorder, Alkermes created the "Hospital Inpatient Free Trail Program". This program allows for in-hospital administration of IM naltrexone which is provided free to the patient as a sample.

At the time of study enrollment, the ACS team member will ask a series of validated questions from the Addiction Severity Index Lite (ASI-Lite), a widely used instrument to assess the severity of problems related to substance use. The survey questions produce reliable and valid estimates of patient status in various life domains including 1) alcohol use, 2) medical health, 3) employment, and 4) illegal activity. These measure are mathematically derived and have shown reliability and validity in several settings. Composite scores will be assessed and reported for questions related to medical condition and alcohol use at baseline (see Appendix, Table 1 for ASI questions and composite questions). Assessments will be completed online using REDCap survey tools. Appendix, Table 1 lists baseline ASI-Lite questions to be asked.

Surveys to assess provider's perceived feasibility and acceptability of the intervention will be sent out prior to the study implementation, at study midpoint (when enrollment is 50% complete), and at study completion. The survey will be sent to all providers on the ACS (social workers, peer navigators, and physicians) (n=13). Surveys will be created and disseminated in REDCap and will be sent out in three waves to improve response rate. The survey will include sections where ACS team members can type in their thoughts and perceptions of the intervention and make any recommendations for improvement. These data will be qualitative, in contrast to the quantitative survey data, and will be used to inform our future comparative effectiveness trial.

Data Collection Tools: REDCap or the Electronic Helath Record (EHR) via Health Data Compass.

Data Analysis Plan:

Baseline Comparison of Treatment Groups: Participant's characteristics and baseline clinical data will be tabulated, and differences between IM vs. oral naltrexone groups will be assessed using a Fisher's Exact Test for categorical variables and Wilcoxon rank-sum tests and Chi-squared tests for continuous variables.

Propensity Score Inverse Probability of Treatment Weighting: The investigators will use inverse probability of treatment weighting (IPTW) based on the propensity score to reduce bias by ensuring that participants who are randomized to IM naltrexone (n=30) vs. oral naltrexone (n=60) will have comparable covariate distributions. Propensity scores will be calculated from a multivariable logistic regression model with a binary outcome of IM naltrexone treatment or oral naltrexone treatment. Variables shown in Appendix, Table 1 will be included if the variables are significantly associated with both IM naloxone treatment and the outcome or with the outcome only. Significance will be assessed based on clinical judgement and bivariate statistical analysis (p-value < 0.05). If highly collinear variables are included in the model, variables associated with both IM naloxone treatment and the outcome will be preferred over those only associated with the outcome. Weights will be stabilized to more accurately estimate variance if a small number of weights become especially large and influential due to particularly low or high propensity scores. After applying weights, covariate distributions will be compared between the group receiving IM naltrexone and the group receiving oral naltrexone to ensure that confounding variable distributions are approximately balanced between groups.

Once IM naltrexone and oral naltrexone groups are matched, the investigators will assess comparative effectiveness outcomes, 30-day all cause hospitalization rates+ 30-day emergency department encounter rates by IM vs. oral naltrexone, and 90-day post discharge linkage rates by IM vs. oral naltrexone group. This pilot study is not sufficiently powered for hypothesis testing on these outcomes. By way of example, the use of 30 participants compared to 60 participant controls yields just 40% power to detect a 50% relative difference in readmission using a pilot appropriate alpha of 10%.36 Regardless, the readmission risk differences will be evaluated using multiple logistic regression to assess the association of treatment on 30-day combined readmissions and emergency department encounter rates. The investigators will use multiple logistic regression to assess the association between IM naltrexone or oral naltrexone on 90-day addiction treatment linkage. These findings will provide parameter estimates that can inform a future sample size calculation in a larger effectiveness trial.