Discussion

We found that among patients with SUD requiring OPAT, there were no statistically significant differences in hospital readmission, emergency department visits, infection relapse, or complications for patients discharged to SNF compared to patients discharged elsewhere. However, length of hospital stay was prolonged in the group going to SNF, and patients discharged to SNF were less likely to be engaged in SUD treatment 30 days postdischarge.

Prior work has also suggested that patients with SUD requiring OPAT who were discharged to SNF experience high rates of healthcare services use, including 30-day hospital readmissions as high as 29%. ^{Reference Langenstroer, Crnich, Jolles, Furuno, Noble and Bailey21} Our work adds to this by being one of the first to directly compare patients with SUD requiring OPAT discharged to SNF versus other locations (primarily home), and further suggest that patients with SUD may have similar outcomes whether they are discharged to SNF or not. One other study similarly failed to show a difference in outcomes in patients discharged to SNF on OPAT, ^{Reference Robbins, Ard, Wakeman, Alves and Nelson12} but that study only followed 52 patients.

Other data suggest that patients discharged to SNF on OPAT have higher rates of non-receipt of laboratory monitoring and poorer follow-up than those receiving OPAT at home. ^{Reference Keller, Ciuffetelli and Bilker18,Reference Huck, Ginsberg, Gordon, Nowacki, Rehm and Shrestha19} In our OPAT Program, we focus on obtaining laboratory test results from all locations patients are discharged to, which may have made us more successful in receiving laboratory test results from SNF. A strong OPAT program may mitigate some struggles noted with laboratory test result receipt in prior studies. ^{Reference Hamad, Lane, Beekmann, Polgreen and Keller13} However, in our workflows, receiving laboratory test results from SNFs is extremely time-consuming.

In this study, patients with SUD requiring OPAT discharged to SNF versus to other locations had longer lengths of hospital stay. Some of this is likely due to the additional logistical work needed to identify an accepting SNF, although the extended length of hospital stay may also partially reflect infection severity and medical complexity. Previous work has shown that patients with SUD have lower rates of SNF acceptance and a higher likelihood of being accepted to lower quality facilities. ^{Reference Kimmel, Rosenmoss, Bearnot, Larochelle and Walley17,Reference Kuye, Prichett, Stewart, Berkowitz and Buresh22,Reference Kimmel, Rosenmoss and Bearnot23} Taken together, these data suggest that patients with SUD requiring OPAT may have similar infection-related outcomes and shorter lengths of stay if they are discharged to a location other than a SNF if they otherwise qualify for a location other than a SNF.

In addition, we noted that patients discharged to SNF were less likely to still be receiving SUD treatment at 30 days posthospital discharge. This is not surprising, as there are multiple barriers that preclude arranging long-term SUD treatment until completion of the SNF stay, leading to delayed entry to SUD treatment. Patients discharged to SNF may have also had fewer social or economic resources to support ongoing engagement in SUD treatment. In our experience, most SNFs do not provide any SUD support or treatment. ^{Reference Kuye, Prichett, Stewart, Berkowitz and Buresh22} Mean length of stay at SNFs in one study was 21 days; ^{Reference Byrd, Stark and Head24} it is possible that some patients in our study may have then engaged in SUD treatment after discharge from SNF, but most SNFs also do not provide coordination services to ensure SUD care after discharge from SNF. At our hospitals, hospital-based PRS do not have capacity to support patients after hospital discharge, and cannot make direct referrals to community-based SUD treatment for patients being discharged to SNF, other than for outpatient treatment programs for continuation of methadone. In our clinical experience, the SNF stay often is a time when SUD treatment is lacking, with no treatment besides continuation of medication started in the hospital (ie, unlikely to have dose adjustments of MOUD), and no groups or other supports as well as risks of gaps in MOUD. A previous pilot included grant funding for a SNF-embedded PRS showed promise for increased uptake of MOUD during and after SNF stay, but more sustainable solutions are required. ^{Reference Tassey, Ott, Alvanzo, Peirce, Antoine and Buresh25} We are exploring opportunities to fill this gap and better support continued SUD treatment by including outpatient PRS in care prior to discharge, and this remains an important area of study.

We also noted that one in four patients with SUD on OPAT discharged to SNF experienced a PDD. Other studies have shown that patients with SUD requiring OPAT frequently experienced PDD from SNF (20%) or otherwise left a SNF prior to the scheduled completion of therapy (16%). ^{Reference Ashraf, Hoff and Brown26} Further investigations should explore whether and why patients with SUD requiring OPAT frequently experience PDD from SNF. Of note, we may have underestimated PDD in this analysis, as we did not have access to all of the SNF records and instead relied on OPAT Program members documenting that patients had experienced PDD when reaching out to SNFs for care coordination purposes.

Our study had limitations. We performed a retrospective review of medical records at two medical centers in one health care system with an ACS and an OPAT program, and some of the patients who received OPAT at home were enrolled in a program with a focus on SUD. It is possible that these results may not be easily translated to other health care systems. Similarly, our study only included patients who required OPAT, not including patients on complex outpatient antimicrobial therapy (COpAT) or long-acting injectable antimicrobials. However, for many patients with SUD requiring OPAT, alternatives to OPAT (such as COpAT, or long-acting injectable antimicrobial) may be more appropriate. We also may have underestimated SUD treatment engagement, as well as PDD, although the OPAT Program had ongoing communication with SNFs that allowed some of this data to be captured. The study population also did not include patients who experienced PDD from the hospital prior to initiating OPAT. There were fewer patients in the non-SNF group which may have limited the power to detect differences. We lacked data on housing status and other socioeconomic factors such as access to transportation that might influence postdischarge treatment options and the ability to engage in ongoing SUD or OPAT treatment. We did not include data on ACS in the final multivariate analysis of engagement in care as we felt that this was on the pathway of treatment rather than a confounder; however, it is important to note that more patients going to SNF had an inpatient addiction consultation. We used approaches to confirm engagement in care, but may not have fully counted all patients engaged in care. Finally, there is likely a difference between patients who may require SNF versus those who may not require SNF (eg, medical complexity, severity of infection, increased needs, unstable housing or housing status), and there is likely a subset of the patients attending SNF who would not have been eligible for OPAT at other locations even in the absence of SUD. This may have led to selection bias.

Our data adds to the growing literature reflecting a lack of increased risk of poor outcomes for selected patients with SUD who may be able to receive OPAT in other settings (eg, the home, infusion centers, addiction treatment centers, or hemodialysis), and patients receiving OPAT in other settings may have decreased lengths of hospital stay. Some barriers may exist; for example, lack of housing may preclude home OPAT as a treatment option. Ongoing partnerships with SUD treatment facilities may also present an option for OPAT delivery. In addition, as patients may be more likely to be engaged in SUD treatment if they do not go to a SNF, and as a significant proportion of patients with SUD discharged to SNF on OPAT in our study experienced PDD, there may be occasions when patients with SUD requiring OPAT would benefit from not going to a SNF. We suggest that decisions about the best treatment approach should be patient centered and multidisciplinary, accounting for the risks and benefits of each treatment location for the individual patient. More health systems-based work, including embedding OPAT into residential addiction treatment facilities and expanding access to home-based OPAT for patients with SUD, is needed to expand access to patient-centered options for completion of required antimicrobial therapy for patients with SUD.