INTRODUCTION
Drug and alcohol-related hospitalizations with need for subsequent skilled nursing facility (SNF) placement are increasing.1 Additionally, current literature illustrates an urgent need to address discharge barriers such as lack of post-acute care options that contribute to unnecessarily long hospitalizations.2 Previous studies have documented discriminatory refusals from post-acute facilities related to opioid use disorder or opioid agonist therapy3,4; however, the inability to secure SNF placement for patients with any substance use disorder (SUD) has not been fully explored.
OBJECTIVE
The primary objective of this study was to measure the increased odds of SNF referral failure for patients with SUD versus those without SUD and the secondary objective was to compare the time spent inpatient between SNF referral and discharge for both groups.
METHODS
We reviewed administrative data for all patients discharged from a 400-bed, urban, level 1 trauma-designated US hospital between January 1, 2019, and January 1, 2021 who were also referred to SNF during hospitalization. Patients who died, left against medical advice, or were transferred to another inpatient hospital were excluded. The primary outcome of interest, SNF referral failure, was defined as discharge to home/self-care despite SNF referral. SUD in our administrative data was defined as the presence of a dedicated flag in the patient’s electronic medical record indicating active drug or alcohol use disorder within the previous year. Tobacco use is not included in the SUD flag. The SUD flag is entered by a member of the patient’s care team as part of their standard screening process and is readdressed at each hospitalization. We chose this method of SUD detection given the reported low sensitivity of other methods such as ICD-10 codes.5
Independent variables related to patient demographics, sociodemographic markers, hospitalization details, and mental health comorbidity were also captured. Univariable analysis using chi-square test was conducted and variables with p ≤ 0.25 were then used to create a stepwise binary logistic regression model of the primary outcome. Secondary outcome data was summarized using median with interquartile range and differences between groups were tested using Wilcoxon rank-sum or one-way ANOVA.
RESULTS
Our sample included 3926 hospitalizations with 732 (18.6%) SNF referral failures. Univariable analysis of primary and secondary outcome data is presented in Table 1. Patients with SUD experienced a high proportion of SNF referral failure (278/800; 34.8%) and remained inpatient longer between SNF referral and discharge than those without SUD (median 7.5 days [IQR 3–20] vs. median 4 days [IQR 2–8], p<0.0001). In our fully adjusted model (Table 2), SUD was an independent predictor of SNF referral failure with a 94% increase in odds as compared to patients without SUD (aOR 1.94; 95%CI 1.58–2.38). Other factors related to increased odds of SNF referral failure included homelessness, primary insurance, and race/ethnicity. Older age and ICU stay were associated with reduced odds of SNF referral failures.
Table 1.
Univariable Analysis of Failed SNF Referral and Days Between SNF Referral and Discharge
| Failed SNF referral, n (%) | p value | Days from referral to discharge, median (IQR) | p value | |
|---|---|---|---|---|
| All patients (n = 3926) | 732 (18.6%) | 4 (2–10) | ||
| Substance use disorder | <0.0001 | <0.0001 | ||
| Yes (n = 800) | 278 (34.8%) | 7.5 (3–20) | ||
| No (n = 3126) | 454 (14.5%) | 4 (2–8) | ||
| Age | < 0.0001 | <0.0001 | ||
| < 50 years (n = 590) | 206 (34.9%) | 9 (4–22) | ||
| 50–64 years (n = 1098) | 297 (27.1%) | 6 (3–14) | ||
| ≥ 65 years (n = 2238) | 229 (10.2%) | 3 (1–7) | ||
| Sex | 0.001 | <0.0001 | ||
| Male (n = 2197) | 450 (20.5%) | 5 (2–12) | ||
| Female (n = 1729) | 282 (16.3%) | 4 (2–8) | ||
| Race/ethnicity, 24 missing entries | < 0.0001 | <0.0001 | ||
| American Indian/Alaska Native (n = 118) | 39 (33.1%) | 6 (2–16) | ||
| Asian (n = 319) | 50 (15.7%) | 4 (2–8) | ||
| Black/African American (n = 417) | 95 (22.8%) | 6 (2–14) | ||
| Hispanic/Latino (n = 219) | 47 (21.5%) | 6 (2–16) | ||
| Native Hawaiian/Pacific Islander (n = 28) | 6 (21.4%) | 8.5 (2.5–31.5) | ||
| White, non-Hispanic (n = 2801) | 491 (17.5%) | 4 (2–9) | ||
| Preferred language, 6 missing entries | 0.79 | 0.70 | ||
| English (n = 3503) | 655 (18.7%) | 4 (2–10) | ||
| Non-English (n = 417) | 75 (18.0%) | 4 (2–11) | ||
| Primary insurance | < 0.0001 | <0.0001 | ||
| Medicare (n = 2402) | 285 (11.9%) | 3 (1–7) | ||
| Medicaid (n = 800) | 283 (35.4%) | 9 (4–21) | ||
| Private/commercial (n = 543) | 127 (23.4%) | 6 (3–12) | ||
| Self-pay (n = 11) | 6 (54.6%) | 2 (1–6) | ||
| Other (n = 170) | 31 (18.2%) | 4.5 (2–11) | ||
| Experiencing homelessness | <0.0001 | <0.0001 | ||
| Yes (n = 349) | 143 (41.0%) | 12 (4–28) | ||
| No (n = 3577) | 589 (16.5%) | 4 (2–9) | ||
| ICU during hospitalization | 0.18 | <0.0001 | ||
| Yes (n = 1989) | 354 (17.8%) | 6 (2–14) | ||
| No (n = 1937) | 378 (19.5%) | 3 (2–7) | ||
| Discharging service type | < 0.0001 | 0.0003 | ||
| Medical (n = 1603) | 247 (15.4%) | 5 (2–10) | ||
| Surgical (n = 2234) | 479 (21.4%) | 4 (2–10) | ||
| ICU (n = 89) | 6 (6.7%) | 8 (3–15) | ||
| Mental health disorder | 0.003 | <0.0001 | ||
| Yes (n = 1144) | 247 (21.6%) | 6 (2–13) | ||
| No (n = 2782) | 485 (17.4%) | 4 (2–9) |
Table 2.
Logistic Regression Model Predicting Odds of SNF Referral Failure
| aOR (95%CI) | p value | |
|---|---|---|
| Substance use disorder | 1.94 (1.58–2.38) | <0.0001 |
| Age | ||
| <50 years | Ref | |
| 50–64 years | 0.82 (0.65–1.04) | 0.10 |
| 65+ years | 0.40 (0.30–0.53) | <0.0001 |
| Male sex | 1.07 (0.89–1.29) | 0.45 |
| Race/ethnicity | ||
| White, non-Hispanic | Ref | |
| American Indian/Alaska Native | 1.59 (1.03–2.45) | 0.04 |
| Asian | 1.17 (0.83–1.64) | 0.37 |
| Black/African American | 1.08 (0.82–1.42) | 0.59 |
| Hispanic/Latino | 0.78 (0.54–1.13) | 0.19 |
| Native Hawaiian/Pacific Islander | 1.34 (0.52–3.44) | 0.54 |
| Primary insurance | ||
| Medicare | Ref | |
| Medicaid | 1.78 (1.37–2.30) | <0.0001 |
| Private/commercial | 1.48 (1.12–1.97) | 0.007 |
| Self-pay | 8.58 (2.41–30.53) | 0.001 |
| Other | 1.31 (0.86–2.03) | 0.21 |
| Experiencing homelessness | 1.84 (1.41–2.42) | <0.0001 |
| ICU during hospitalization | 0.72 (0.61–0.87) | <0.0001 |
| Discharging service type | ||
| Medical | Ref | |
| Surgical | 1.44 (1.20–1.74) | <0.0001 |
| ICU | 0.37 (0.16–0.88) | 0.02 |
| Mental health disorder | 1.18 (0.98–1.43) | 0.08 |
DISCUSSION
In this analysis of all discharges from a public hospital over a 2-year period, patients with SUD referred to SNF were more likely to discharge to self-care and remained longer in the hospital awaiting disposition than those without SUD. These SNF referral failures are important markers of inequitable access to needed resources as well as ongoing marginalization of individuals with SUD. Patients with SUD had longer inpatient length of stay awaiting disposition in this study, suggesting that additional care coordination activities were required (e.g., SNF referral expansion until all options were exhausted then referrals to outpatient support resources) and/or that patients received outpatient level care in the hospital until appropriate clinical progression had occurred. Further research delineating individual substance use patterns, type of substance used, and primary vs. secondary diagnoses would provide useful insights to better understand SNF referral failures. Furthermore, centering patient-related data such as experiences of stigma or mistreatment6 as well as outcomes such as adverse events, discharge against medical advice, etc., is an important area of future work. This study builds upon a growing body of literature highlighting that systemic, targeted intervention by health systems and payors as well as legislative and regulatory bodies is necessary to ensure equitable access and to mitigate the burden borne by hospitals and individual patients in need of SNF-level care.
Declarations
Conflict of Interest
The authors declare that they do not have a conflict of interest.
Footnotes
Prior Presentations:
This material was presented at the 2022 Society of Hospital Medicine national conference.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Han BH, Tuazon E, Kunins HV, Paone D. Trends in inpatient discharges with drug or alcohol admission diagnoses to a skilled nursing facility among older adults, New York City 2008-2014. Harm Reduct J. 2020;17(1):99. doi: 10.1186/s12954-020-00450-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Meo N, Bann M, Sanchez M, Reddy A, Cornia PB. Getting unstuck: challenges and opportunities in caring for patients experiencing prolonged hospitalization while stable for discharge. Am J Med. 2020;133(12):1406–10. doi: 10.1016/j.amjmed.2020.05.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wakeman SE, Rich JD. Barriers to post-acute care for patients on opioid agonist therapy; an example of systematic stigmatization of addiction. J Gen Intern Med. 2017;32(1):17–9. doi: 10.1007/s11606-016-3799-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kimmel SD, Rosenmoss S, Bearnot B, Larochelle M, Walley AY. Rejection of patients with opioid use disorder referred for post-acute medical care before and after an anti-discrimination settlement in Massachusetts. J Addict Med. 2021;15(1):20–6. doi: 10.1097/ADM.0000000000000693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Rowe CL, Santos G-M, Kornbluh W, Bhardwaj S, Faul M, Coffin PO. Using ICD-10-CM codes to detect illicit substance use: A comparison with retrospective self-report. Drug Alcohol Depend. 2021;221:108537. doi: 10.1016/j.drugalcdep.2021.108537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Biancarelli DL, Biello KB, Childs E, et al. Strategies used by people who inject drugs to avoid stigma in healthcare settings. Drug Alcohol Depend. 2019;198:80–6. doi: 10.1016/j.drugalcdep.2019.01.037. [DOI] [PMC free article] [PubMed] [Google Scholar]