Post-acute care trends and disparities after joint replacements in the United States, 1991–2018: A systematic review

Abstract

Objective.

To review evidence on post-acute care (PAC) use and disparities related to race and ethnicity and rurality in the United States (U.S.) over the past two decades among individuals who underwent major joint replacement (MJR).

Design:

Systematic review

Setting and participants:

We included studies that examined U.S. PAC trends and racial and ethnic and/or urban versus rural differences among individuals who are >=18 years with hospitalization after MJR.

Methods:

We searched large academic databases (PubMed, CINAHL, Embase, Web of Science, and Scopus) for peer-reviewed, English language articles from January 1, 2000, and January 26, 2022.

Results:

Seventeen studies were reviewed. Studies (n=16) consistently demonstrated that discharges post-MJR to skilled nursing facilities (SNFs) or nursing homes (NHs) decreased over time, whereas evidence on discharges to inpatient rehab facilities (IRFs), home healthcare (HHC), and home without HHC services were mixed. Most studies (n=12) found that racial and ethnic minority individuals, especially Black individuals, were more frequently discharged to PAC institutions than White individuals. Demographic factors (i.e., age, sex, comorbidities) and marital status were not only independently associated with discharges to institutional PAC, but also among racial and ethnic minority individuals. Only one study found urban-rural differences in PAC use, indicating that urban-dwelling individuals were more often discharged to both SNF/NH and HHC than their rural counterparts.

Conclusions and implications:

Despite declines in institutional PAC use post-MJR over time, racial and minority individuals continue to experience higher rates of institutional PAC discharges compared to White individuals. To address these disparities, policymakers should consider measures that target multimorbidity and the lack of social and structural support among socially vulnerable individuals. Policymakers should also consider initiatives that address the economic and structural barriers experienced in rural areas by expanding access to telehealth and through improved care coordination.

Brief summary:

Despite declines in institutional post-acute care (PAC) use after major joint replacement over the past 20 years, racial/minorities persistently had higher rates of institutional PAC discharges compared to White individuals.

Introduction

Despite advancements in healthcare delivery processes, health disparities among racial and ethnic minority groups and rural residents persist. Racial and ethnic minority groups often have worse health status¹ with multimorbidity, which are often underdiagnosed^2,3 compared to White individuals. Despite limited research on rural health disparities, available studies show rural residents experience poorer health status,⁴ lower healthcare utilization,^4,5 and higher mortality⁵ compared to their urban counterparts. This gap is especially pronounced in major joint replacements (MJR) and other invasive surgeries.⁶

While postoperative outcomes after MJR are not affected by one’s racial and ethnic status alone, an environment of structural racism contributes to the association between the two. Racial and ethnic minority individuals experience worse postoperative outcomes (i.e., pain, satisfaction, function, and quality of life) and increased complications compared to White individuals post-MJR.⁷ These outcomes may be influenced by poor baseline functional status, living and working in unsafe physical environments, limited educational or employment opportunities, and restricted healthcare access.^8,9

Postoperative outcomes after MJR are influenced by place of residence. Rural-dwelling individuals are at a higher risk of complications than those in urban regions.¹⁰ These disparities may be partially explained by the medical and social complexity experienced by rural-dwelling individuals when undergoing MJR.⁶

Post-acute care (PAC) is care that is provided to individuals who need additional help recuperating from an acute illness or serious medical procedure.¹¹ PAC settings such as skilled nursing facilities/nursing homes (SNFs/NHs), inpatient rehabilitation facilities (IRFs), and home health care (HHC) offer essential alternatives to a prolonged hospital stay.¹² Often individuals with a high burden of chronic illness, baseline functional dysfunction,¹³ and lack of community/caregiver support¹⁴ are not good candidates for home discharge and are directed to PAC services. Unfortunately, the availability and quality of PAC services, particularly HHC, are often limited for racial and ethnic minority individuals and rural residents. Researchers have found that racial and ethnic minority individuals are less likely to receive HHC and more likely to receive care from low quality HHC compared to White individuals.^15,16 Similarly, rural residents access PAC services, such as HHC,⁵ less frequently than their urban counterparts.^17,18

Researchers have suggested that PAC decisions may be guided by the presence and interplay of patient demographics (i.e., number and severity of comorbid conditions), organizational characteristics (i.e., teaching status), or market characteristics (i.e., the availability of PAC services in the area).^19,20 As such, the Social Ecological Theory (SET) offers a comprehensive framework by highlighting the importance of individual, interpersonal, community, organizational, and policy-level factors in healthcare services use.²¹ This approach provides a socioecological lens that may elucidate aspects of PAC decisions post-MJR.

With the evolving healthcare landscape characterized by shorter hospital stays and an increased use of PAC, understanding general trends in PAC utilization over time among patients who underwent MJR is crucial because it provides pertinent context for clinicians and policymakers. Furthermore, this finding may highlight any existing disparities in PAC utilization among various racial and ethnic groups or rural populations. ²¹Our study had two aims. First, we sought to describe U.S. PAC use trends over time and disparities related to race and ethnic status or urban versus rural location post-MJR. Second, we aimed to evaluate existing evidence using the SET to understand the dynamics of individual and contextual factors which may impact trends and disparities regarding PAC use among individuals undergoing MJR.

Methods

We adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Fig. 1).²²

Figure 1:

PRISMA Flow Diagram

Search Strategy and Study Selection

We searched the following five electronic databases: 1) PubMed, 2) Cumulative Index to Nursing and Allied Health Literature [CINAHL], 3) Embase, 4) Web of Science and 5) Scopus. We searched for articles that focused on MJR and the following PAC services and/or discharge dispositions: 1) IRF, 2) SNF/NH, 3) HHC, or 4) home without HHC. We also hand-searched the references of the articles that met eligibility criteria for additional papers, though no other publications were found. The search terms are outlined in Table S1.

Articles were eligible if they were peer-reviewed, English-language, pertained to PAC utilization over time and/or differences in PAC use related to race and ethnicity (inclusion of different racial or ethnic groups) or urban versus rural location (at the individual or facility level) in the U.S. for the following medical/surgical diagnoses: 1) stroke, 2) sepsis, 3) acute myocardial infarction, and 4) MJR. Studies had to be published between January 1, 2000 and January 26, 2022. We excluded articles that 1) pertained to pediatric patients 2) involved discharge from a non- acute-care hospital 3) did not include discharge to home without HHC, HHC, SNF/NH, or IRF and 4) were non-empirical studies, such as commentaries, abstracts, dissertations, case studies and literature reviews. Relatedly, since this review was part of a study that included multiple medical and surgical diagnoses, team members agreed that the MJR population warranted a separate review, and we excluded studies with hospitalizations not related to MJR.

Two authors independently screened titles and abstracts and reviewed full texts. Thereafter, discrepancies were discussed/resolved through consultation with a third senior team member.

Data Extraction

Two authors extracted data on: 1) authors, 2) publication year, 3) whether the paper included discharge trends, discharge disposition by urban versus rural regions, and/or by race and ethnic status, or both, 4) study aim, 5) study design, 6) PAC type, 7) MJR type, 8) sample characteristics, 9) data sources, 10) urban versus rural variables, 11) race and ethnicity statuses, 12) relevant covariates at the individual-, interpersonal-, community-, organizational-, and policy-level, 13) statistical analysis, 14) relevant results/findings, and 15) limitations.

Study Quality Appraisal

Two authors independently assessed study quality using the Joanna Briggs Institute (JBI) quality assessment tool, assessing methodological robustness across study design, execution, and analysis.²³ We utilized the JBI checklist for Analytical Cross-Sectional Studies (eight questions) and Cohort Studies (11 questions). Each checklist item is responded to with either a “yes” or “no,” with “yes” contributing to the total score. Based on the quality assessment, studies were categorized for inclusion, exclusion, or seek further information. We included all studies to provide an exhaustive review. Again, any discrepancies during the data extraction and appraisal processes were discussed/resolved through consultation with a third senior team member.

Conceptual Framework and Data Synthesis.

We used the SET to guide our synthesis.²¹ This theory recognizes the interplay of individual-, interpersonal-, community-, organizational-, and policy-level factors and the influence of these interactions on health service utilization behavior.²¹ Since decisions on discharge disposition and utilization of PAC services may be influenced by the presence and interplay of individual (i.e., demographics) and contextual (i.e., organizational or market characteristics) factors,^19,20 this framework may help explain time trends and disparities related to racial and ethnicity status and/or rurality in PAC use post-MJR. We first describe PAC trends over time and disparities based on racial and ethnicity status and urban versus rural locations among individuals undergoing MJR to provide an overview of PAC use post-MJR. Second, we present our synthesized results in the context of the multidimensional factors from SET.²¹

Results

The initial search identified 793 studies with 109 duplicates. We screened the titles and abstracts of these 684 studies, excluding 563 studies and reviewing the full texts of 121 studies, and including 17 studies (Fig. 1). Sixteen studies^17,24–38 utilized cross-sectional designs, whereas one³⁹ used a cohort design. Regarding study quality, most cross-sectional studies (n=11)^{17,24–28,31,33,36–38} scored an 8/8 on the JBI Critical Appraisal Tool, whereas a few scored <8 points (n=5) (Table S2).^{29,30,32,34,35} The cohort study (n=1)³⁹ scored 9/11 (Table S3).

Table 1 highlights study characteristics and Table S4 presents detailed summaries. The publication dates spanned from 2009 to 2021 with data collection dates ranging from 1991 to 2018. Discharge disposition post-MJR were discharges to home (n = 11),^{17,25–27,29–34,37} HHC (n=10),^{17,26,30–33,35–37,39} SNF/NH (n=14)^{17,25,28–39} and IRF (n=14).^{25–27,29–39} Most (n=11)^{17,25–28,31,32,34,35,37,38} studies addressed differences in PAC use by either race/ethnic status or urban versus rural location,⁴⁰ a few studies (n=4)^30,33,36,39 examined both PAC use over time and PAC-related disparities. Only two articles included only PAC use over time.^24,29 As for the MJR types, most (n=9)^{17,24,26–28,31,33,35,36} articles included both total knee and total hip arthroplasty, with seven studies^{25,30,32,34,37–39} focused only on total knee arthroplasty and one study²⁹ on total hip arthroplasty. The data used in the included studies came primarily from national databases (total n = 9; Medicare: n = 5^{24–26,29,33}; American College of Surgeons National Surgical Quality Improvement Program data (ACS-NSQIP): n=2^28,38; and the National Inpatient Sample of the Healthcare Cost and Utilization Project (HCUP): n=2^40,41), followed by data from state databases (n=3)^32,37,39 and single institutions (n=6).^{26,27,30,31,34,35} Sample sizes ranged from 2,228 to 1,802,089 participants. Six studies included hospital characteristics including urban versus rural status (n=5),^{17,24,32,33,42} ownership (n=2),^17,24 admission volume (i.e., average daily census) (n=2),^17,24 and total joint replacement volume (n=2).^32,39

Table 1.

Characteristics of Included Studies (n=17)

Author (Year)	Sample Size	Time Trends/Disparities		Database			Type of Post-acute Care				Statistical Tests Used
		Time trends	Disparities	National & Medicare	State	Single Inst.	Home	HHC	SNF/NH	IRF	REG	Other
Buntin (2009)	1,787,094	★		★							★
Cram (2011)	1,802,089	★		★			★		★	★	★
Jorgenso n (2015)	129, 522		★		★		★	★	★	★	★	PSM
Inneh (2016)	7,924		★			★	★	★	★	★	★
Schwarzk opf (2016)	28,611		★		★		★	★	★	★	★
Burke (2017)	1,560,000		★	★			★	★	★		★
Lan (2017)	2869		★			★		★	★	★	★
Cram (2018)	101,409		★	★					★		★
Fang (2018)	2,276	★	★			★	★	★	★	★	★	Paired t-test
Singh (2019)	107,768	★	★		★			★	★	★	★
Cavanau gh (2020)	8,349		★	★			★		★	★	★	PSM
Kim (2020)	688,346	★	★	★			★	★	★	★		DID
Reynolds (2020)	465,702	★	★	★				★	★	★	★
Klemt (2021)	2,228		★			★	★		★	★	★
Chisari (2021a)	7,310		★	★		★	★	★		★	★
Chisari (2021b)	5,284		★			★	★			★	★	PSM
Wu (2021)	54,582		★	★					★	★	★

Notes: Single Inst= Single Institution Data, HHC = Home Health Care; SNF/NH = Skilled Nursing Facility/Nursing Home, IRF = Inpatient Rehabilitation Facility, REG= Regression, PSM= Propensity score matching, DID= Difference-in-differences.

Studies included various racial and ethnic group of Black, White, Asian, Hispanic, Native American, and other individuals. Among various pairwise comparisons, the most commonly included was Black versus White (n=11),^{17,24–28,31,33,36–38} followed by Asian versus White (n=4),^17,35,37,38 Hispanic versus White (n=3),^17,33,36 and one study each comparing Native American versus White,¹⁷ Other versus White,³⁷ and Hispanic versus non-Hispanic.³⁷

For urban versus rural status, these designations were either based on the individual’s residence (n=5) ^{24,25,31,35,40} or facility location (n=3).^32,40,43 Most studies (n=4)^24,32,40,43 dichotomized this variable.

Table 2 demonstrates individual-, interpersonal-, community-, organizational-, and policy-level variables that each study considered. All studies (n=17)^{24–28,30–35,37,38,40,41,43,44} considered individual-level variables. Seven studies^{27,28,34,37,38,41,44} focused solely on individual-level variables, while three studies^24,25,40 incorporated at least four of the five-level variables specified by the SET. Notably, no studies examined all five-level variables of the SET. The most frequently mentioned individual-level variables across all studies were the following: 1) age, 2) sex, 3) race, and 4) comorbidities. Three studies looked at interpersonal-level variables such as marital status.^25,34,40 Five studies considered community-level variables, with the most frequently mentioned variables being geographic region and neighborhood economic status (i.e., median income).^{24,25,31,35,40} Five studies focused on organization-level variables such as the hospital characteristics like teaching status, ownership, size, and MJR volume.^{24,32,33,40,43} Seven studies evaluated policies that were in effect during the time periods studied.^{24–26,30–32}

Table 2.

Study variables based on the Social Ecological Theory (n=17)

Author & Year	Individual	Interpersonal	Community	Organizational	Policy
Buntin (2009)	Age, gender, race, clinical complexity, medicaid coverage		Place of residence, rural area	Size, teaching status, medicare patient %, case mix index of hospital, low-income patient %	Post-acute care payment systems
Cram (2011)	Age, sex, race, comorbidities
Jorgenson (2015)	Age, gender, race, comorbidities, insurance		Rural-urban commuting area codes	% of Black patients at each facility; procedure volume	75% rule for inpatient rehab eligibility
Inneh (2016)	Age, gender, race/ethnicity, comorbidities		Median household income by zip code		BPCI
Schwarzkopf (2016)	Age, gender, race, comorbidities, insurance
Burke (2017)	Age, sex, race, ethnicity, comorbidities, insurance, hospitalization (urgency)	Marital status	Median income of residents	Ownership, size, geographic location
Lan (2017)	Age, gender, race, insurance, risk of mortality, illness severity		Patient residence, median income
Cram (2018)	Age, sex, race, comorbidities
Fang (2018)	Race				BPCI
Singh (2019)	Age, sex, comorbidities, insurance		MSA	MSA location, hospital annual MJR volume
Cavanaugh (2020)	Race, education, income, functional status, multimorbidity	Marital status, living arrangement	Geographic region, neighborhood socioeconomic status, levels of poverty		Medicare 60% rule
Kim (2020)	Age, sex, race, medicaid, medically complex			Teaching, safety-net, ownership, MJR volume	CJR
Reynolds (2020)	Gender, race, health insurance, number of chronic conditions
Klemt (2021)	Age, gender, BMI, insurance, race, comorbidities, preop meds	Marital status, social status
Chisari (2021a)	Age, gender, race, BMI, comorbidities				BPCI
Chisari (2021b)	Age, gender, comorbidities, BMI
Wu (2021)	Age, race BMI, comorbidities, functional status

^*Body mass index (BMI); Metropolitan statistical area (MSA); Bundled payment care initative (BPCI); Comprehensive joint replacement model (CJR)

Time trends in PAC use

Table 3 shows the six studies that examined the trends in PAC use over time and the direction of the trend by PAC setting.^{24,30,33,41,43,44} In four studies, researchers found there was decreased discharges to SNF/NH over time.^24,30,33,39 However, in one study, researchers found increased SNF/NH use.⁴⁴ Discharges to home increased over time^30,43 except in one study²⁴. Discharges to HHC and IRF were mixed with contradictory findings.

Table 3.

Trends in Post-acute Care Discharge Disposition from 2000 to 2022 (n = 6)

Author (Year)	Data Years	Home	HHC	SNF/NH	IRF
Buntin (2009)	1996–2003	−	−	−	+
Cram (2011)	1991–2008			+
Fang (2018)	2014–2017	+	+	−
Singh (2019)	2012–2015	+		−
Kim (2020)	2013–2017			−	−
Reynolds (2020)	2009–2012		−

Notes: “+” = increase discharge to over time “−” = decrease discharge to over time; SNF/NH = skilled nursing facility/nursing home, IRF = inpatient rehabilitation facility, HHC = home health care.

Disparities in PAC use for race and ethnic status or urban versus rural location

Racial and ethnic disparities in discharge dispositions are summarized in Table 4 for each discharge destination: home, HHC, SNF/NH and IRF (n=15).^{25,26,28,30–35,37,38,40,41,43} White individuals were consistently more likely to be discharged home without HHC than Black individuals (n=5).^{25–27,30,35} Evidence on the use of HHC after MJR varied across racial and ethnic groups with limited evidence for any given comparisons. For discharges to SNF/NH and IRF, more than half of the studies comparing Black and White individuals found that Black individuals were more often discharged to SNF/NH (n=12)^{25,28,30–39} or IRF (n=10).^{26,27,31–34,36–39} Though comparisons with other racial and ethnic minority individuals versus White individuals were mixed, most studies demonstrated that racial and ethnic minority individuals were more often discharged to institutional PAC settings compared to their White counterparts (n=4).^3,35,37,41

Table 4.

Evidence on Higher Likelihood of Discharge Disposition by Race and Ethnic Group Comparisons (n=15)

Author (Year)	Type of Post-acute Care
	Home	HHC	SNF/NH	IRF
Black versus White
Jorgenson (2015)		B	B	B
Inneh (2016)			B	B
Schwarzkopf (2016)		B	B	B
Lan (2017)	W		B
Cram (2018)			B
Fang (2018)	W	W	B
Singh (2019)*			B	B
Cavanaugh (2020)	W		B
Kim (2020)			B	B
Reynolds (2020)			B	B
Chisari (2021a)	W			B
Chisari (2021b)	W			B
Klemt (2021)			B	B
Wu (2021)			B	B
Hispanic versus White
Burke (2017)		W	W
Kim (2020)			H	H
Reynolds (2020)		W (2009–2010)	H (2010–2012)	H (2010–2012)
Asian versus White
Schwarzkopf (2016)		A	A	A
Burke (2017)		W	W
Lan (2017)	W		A
Wu (2021)			A	A
Native American versus White
Burke (2017)		W	W
Other Vs. White
Schwarzkopf (2016)		O
Hispanic versus Non-Hispanic (White, Asian, Black)
Schwarzkopf (2016)			H	H

Notes: HHC = home health care, SNF/NH = skilled nursing facility/nursing home, IRF = inpatient rehabilitation facility; B=Black, W=White, H=Hispanic, A=Asian, N= Native American O=Other

Only one study examined urban versus rural disparities in discharges to PAC.¹⁷ This study found that urban-dwelling individuals were more often discharged to both SNF/NH and HHC than individuals living in rural areas.¹⁷

Synthesis of Results with SET

In exploring trends in PAC utilization over time, four studies considered both individual-level factors and other contextual variables of the SET.^24,30,33,43 Three of these studies included policy-level variable(s), particularly those related to Centers for Medicare and Medicaid Services (CMS).^24,30,33 When examining discharges to PAC using the SET, ten studies considered variables across multiple levels – individual, interpersonal, community, and policy.^{25,26,30–35,40,43} Three studies incorporated interpersonal variables, such as marital status, in addition to individual-level variables.^25,34,40 In two of these studies, marital status such as being single or living alone was associated with non-home discharge.^25,34 Six studies included community-level variables such as geographic location alongside individual-level variables.^{25,31,32,35,40,43} In these studies, geographic location served as a proxy for either rurality^32,40,43 or socioeconomic (SES) status.^25,31,35 However, analyzing the impact of rurality among these studies was challenging since most studies^32,43 did not show the proportion of individuals in urban versus rural locations⁴³ or primarily included patients from urban areas.³² In contrast, studies using geographic variable as a proxy for neighborhood SES status found that individuals in racial and ethnic minority groups with low SES status were at an increased risk of discharge to institutional PAC settings.^25,31,35 Lastly, five studies considered policy factors along with other variables.^26,30–33 These studies didn’t find significant effect of policy changes on the PAC discharges among the racial and ethnic minority individuals post-MJR,^26,30–32 with one exception.³³ This study reported that implementation of a CMS bundled payments led to a decrease in institutional PAC use among Black individuals.³³

Discussion

Our review found that since 2000, there has been a notable reduction in discharge to SNF/NH for patients post-MJR. However, racial and ethnic minority individuals were more likely to be discharged to institutional PAC settings and less likely to be discharged home compared to White individuals. Guided by the SET,²¹ we sought to understand the interplay of multidimensional factors related to urban versus rural or racial and ethnic differences in discharges post-MJR. We found that more than half of the reviewed studies (n=11)^{24–26,30–35,40} considered two or more levels of variables from the SET.²¹ Overall, racial and ethnic minority groups (especially Black individuals) were more likely to be discharged to PAC settings than White individuals. Furthermore, age, sex, comorbidities, and marital status (i.e., single or living alone) were not only independently associated with discharges to institutional PAC but were also associated with such discharges among racial and ethnic minority individuals. This highlights the multifaceted nature of factors influencing PAC use, especially in the context of racial and ethnic disparities.

Our findings showing a decreasing trend in the use of SNF/NH are consistent with past evidence. Past literature regarding PAC trends over time noted a general decline in the use of institutional PAC, particularly among hospitals that participate in valued-based payment models such as Bundled Payment for Care Initiative (BPCI).⁴⁵ A narrative review regarding PAC trends and discharge disposition after MJR from 2012 to 2017 found a decrease in the proportion of hospitalized MJR patients being discharged to institutional PAC over time; notably, this decrease was more pronounced among BPCI-participating hospitals than control hospitals.⁴⁶ Additionally, the advent of outpatient joint replacements during our study period may have introduced variability in the PAC landscape, and thereby decreased PAC service use.⁴⁷ Our review also demonstrates decreases in institutional PAC use, notably in the presence of policies in place (n=3).^24,30,33 Relatedly, the periods covered by these studies varied, which may have influenced the observed differences in policy impacts. The main finding of these studies was that changes in PAC use, in general, were driven mainly by policy changes, particularly those set by CMS.^{3,24–26,30,31} Taken together, these trends imply that policy-driven financial incentives might play a role in the declining trend in institutional PAC use post-MJR.

Urban versus rural or racial/ethnicity disparities in discharges to PAC

Rural regions often have less healthcare infrastructure, including PAC availability,⁴⁸ than urban regions, potentially contributing to health disparities.⁴⁹ We found only one study⁴⁰ that specifically examined urban versus rural differences in PAC, whereas two studies^32,43 included a geographic variable as a proxy for rurality. However, synthesis was difficult among these studies since only one study⁴⁰ included rural areas, whereas the others were unclear or had low representation.^32,43 Despite this limitation, that study did find that urban-dwelling individuals were more likely to utilize institutional PAC settings and HHC compared to their rural counterparts.⁴⁰ However, this finding should be interpreted with caution since as aforementioned, the lack of PAC availability in rural regions⁴⁸ limits access to PAC in the first place and makes comparisons difficult.

Although we found a decreasing trend of institutional PAC use, this trend was different for racial and ethnic minority individuals compared to White individuals. This limited but consistent evidence aligns with previous evidence regarding racial and ethnic minority individuals undergoing orthopedic surgeries. Prior studies highlighted that Black individuals face greater risks for complications, mortality, and/or experience greater likelihood of non-home discharge after orthopedic surgeries than their White counterparts.^50,51 Also, racial and ethnic minority individuals have a high prevalence of multimorbidity,^26,38,52 and often experience challenges stemming from social determinants of health (SDoH) such as limited social support, transportation barriers, and unsafe home environments,⁵³ which may increase discharges to institutional PAC settings. Moreover, racial and ethnic minority individuals experience other social and/or structural disadvantages such as residential segregation,⁵⁴ and often have limited healthcare access, especially to high quality healthcare.^54,55 As such, PAC decisions among racial and ethnic minority individuals undergoing MJR may inevitably be influenced by not only individual, but also contextual factors. Similarly, we found that one’s race and ethnicity in conjunction with contextual factors such as marital status or neighborhood SES conditions were associated with discharges to institutional PAC settings.

Policy Implications

Existing evidence for disparities regarding racial and ethnic status in PAC use post-MJR highlights several policy implications. Disparities in PAC use may be mainly attributed to medical needs, in some cases, lack of adequate social support at home. Thus, to address these disparities, we need to consider the root causes, namely (1) increased rates of multimorbidity, and (2) lower levels of social support among racial and ethnic minority individuals compared to White individuals. Addressing multimorbidity is likely the most challenging to address since this involves the interplay of complex factors over an individual’s lifetime and would require physicians and hospital networks to take responsibility for improving population health. Utilization of institutional PAC may be appropriate if it addresses a need for more support at the time of discharge because of multimorbidity and/or lack of social support (or other issues related to SDoH). We need to ensure that the needs of individuals with fewer available social support safety net and issues with SDoH continue to be met and individualize PAC decisions based on these needs to prevent poor postoperative outcomes. Additionally, future policies should not: 1) single-mindedly aim to decrease access to institutional PAC, or 2) inadvertently incentivize decreasing the appropriate use of institutional PAC.

We found limited evidence regarding urban versus rural disparities in PAC use post-MJR. However, one study that examined urban versus rural differences in PAC use post-MJR⁴⁰ demonstrated limited access to PAC among rural residents. Hospital and provider shortages in rural areas⁵⁶ may contribute to urban versus rural differences in PAC use⁴⁸ post-MJR. Given the economic and/or structural barriers in addressing these shortages, one potential approach is to encourage urban healthcare institutions to expand their services to rural areas. Expanding remote healthcare delivery methods, such as telemedicine, may enhance rural patients’ access to specialized care. Additionally, enhancing care coordination with community health workers or nurses during the perioperative period may also address the urban versus rural gap in PAC use.

Limitations and strengths

Our review has limitations. The divergent definitions and categorizations of race and ethnicity status among the included studies complicated direct comparisons of specific groups. Similarly, inconsistent categorization/groupings regarding PAC services presented challenges for summarizing evidence consistently for each discharge disposition. Our search strategy was limited to studies conducted in the U.S., to account for nationally specific policies during the study periods and may not be generalizable to other countries. Additionally, we could not find any significant differences in discharge to PAC or related disparities depending on the type of MJR. Many studies combined the different types of MJR during their analyses and did not make direct comparisons. Lastly, our synthesis was limited since we were unable to control for the relevant contextual factors as depicted through the SET or examine the interplay of these elements. Since race and ethnicity are primarily social constructs (as opposed to biological constructs),⁵⁷ which fundamentally influences life experiences, social exposure, and health outcomes,⁵⁸ we acknowledge that each of the contextual factors in SET may not only affect discharge disposition status, but also relate to one another. However, the strengths of our review include our evaluation of study quality using a validated quality appraisal tool, following PRISMA guidelines, and using a theoretical framework to guide our synthesis.

Conclusions and Implications

Discharges to institutional PAC settings (i.e., SNF/NH) for individuals undergoing MJR have declined over the last 20 years. However, Black individuals experienced higher rates of SNF/NH discharges compared to White individuals. To further understand racial and ethnic health disparities in PAC use post-MJR, future studies should consider examining ways to address root causes such as multimorbidity and lack of social support. Future policies should avoid decreasing access to institutional PAC to minimize costs while overlooking the benefit of institutional PAC for vulnerable individuals. Policymakers should also consider measures that address the economic and structural barriers to PAC in rural areas by expanding telehealth in remote communities.