Ultimately, 65.1% of survivors in our cohort completed antibiotic therapy, 87.2% were seen by addiction specialists, and 89.4% received MOUD. This MOUD initiation rate differs greatly from the 5.7% initiation rate reported in a national study of IDU-IE patients. . Completion of antibiotic therapy in IDU-IE has not been well studied; a study of 26 patients reported a completion rate of 92.3% . In a study of hospitalized patients with infectious complications of IDU, including but not limited to IDU-IE, 52.0% completed antibiotic therapy and 30.4% received consultation in drug addiction . In particular, 38.4% of patients in this study received MOUD, while 48.9% of patients in our study received it throughout their hospitalization. Thus, a significant fraction of patients in both cohorts did not complete antibiotic therapy despite relatively higher use of SUD-oriented interventions at our institution.
In the cohort of Marks et al. , 113 IDU-IE patients were included. An overall 90-day readmission rate of 36.3% was found; among those with and without a substance abuse assessment, this rate was 28.6% and 54.5%, respectively (LR Marks, personal communication, June 5, 2020). In our study, a longer follow-up period was considered and 88.4% of patients were rehospitalized at least once. Thus, at our center with its widespread use of SUD-directed interventions, readmissions were common. The MOUD may be only one critical component to improving outcomes in this population; others have called for greater awareness of the social determinants of health, saying medical interventions are only one aspect of optimizing care for this vulnerable patient population. .
Care transitions are a key consideration in evaluating the outcomes observed in our study, since most survivors in our cohort were discharged to subacute care settings. Federal policy, specifically Title 21 of the Code of Federal Regulations, makes it difficult for patients to receive MOUD in these facilities, preventing continuation of MOUD started in a hospital setting unless the patient is already enrolled in a treatment program OUD . While facilities were expected to continue appropriate therapies after discharge, we were unable to independently verify continuation of MOUD after discharge. Strategies to improve outcomes for people transitioning to non-institutional settings were also explored. In a randomized trial of patients with IDU-related infection, patients in the experimental group underwent frequent outpatient visits after an inpatient stay rather than remaining hospitalized for antibiotic therapy alone; any antibiotic therapy completed , suggesting that careful planning after discharge may improve completion of antibiotic therapy. Support teams specific to the SOUTH have been proposed and can provide key support during the transition of patients to the ambulatory setting.
The main limitation of our study is its sample size, which limits the ability to perform analytical statistics beyond descriptive calculations. Although not all patients were seen by the addiction medicine service, the dichotomy of the sample by this factor created subgroups that were themselves too small for an analytical approach. Similarly, the absence of a control group also prevented us from adopting an analytical approach. The eligibility window boundaries, which determined the size of the study, were chosen to allow exclusive use of ICD-10 billing codes while allowing sufficient time for 12 months of follow-up. ICD-10 codes were only used to capture infection-related diagnoses, not to identify SUDs, for which they are unreliable [14, 15]. Additionally, the study eligibility window took into account the availability of addiction medicine consultation and allowed the study results to reflect contemporary issues in the treatment of IDU-IE. This is particularly relevant as fentanyl has increasingly replaced heroin to become the dominant opioid in our region over the study period. . Other key limitations were the inability to assess engagement in SUD-specific care and receipt of MOUD after discharge from hospital. While the availability of reception data from the MOUD after release would have greatly contributed to the impact of the study, this data was not available as described above. In addition, follow-up data was not available for two people. A final limitation of our study is that it was performed in a single institution that largely cares for underserved and low-income patients, and therefore its generalizability is limited.