Improved nursing home end‐stage renal disease patient participation in physical therapy with onsite, more frequent dialysis

For end‐stage renal disease (ESRD) patients residing in skilled nursing facilities (SNFs), the logistics and physical exhaustion of life‐saving hemodialysis therapy often conflict with rehabilitation goals. Integration of dialysis care with rehabilitation programs in a scalable and cost‐efficient manner has been a significant challenge. SNF‐resident ESRD patients receiving onsite, more frequent hemodialysis (MFD) have reported rapid post‐dialysis recovery. We examined whether such patients have improved Physical Therapy (PT) participation.

Baseline mean self-care and mobility scores were significantly lower in Onsite-MFD versus Non-ESRD or Offsite-Conventional-HD.
Discussion: SNF-resident ESRD patients receiving MFD colocated with rehabilitation had higher PT participation rates than those conventionally dialyzed offsite and equivalent PT participation rates to the non-ESRD SNFrehabilitation general population, despite being sicker, less independent, and less mobile.We report a scalable program integrating dialysis and rehabilitation care as a potential solution for ESRD patients recovering from acute hospitalization.

INTRODUCTION
Older end-stage renal disease (ESRD) patients receiving hemodialysis (HD) are often frail, 1 susceptible to falls, 2,3 are mobility-impaired, and have more functional disability relative to the age-matched general population. 4,5hese factors associate with an increased risk of mortality [6][7][8] and impede achievement of inpatient rehabilitation goals. 9,10Dialysis patients residing in skilled nursing facilities (SNFs) are predominantly older and also have considerable comorbidity, frailty, and functional dependence. 11,12Traditionally, care for SNFresident patients required "offsite dialysis," 13 with inconsistent transportation scheduling and highly draining thrice-weekly dialysis sessions challenging participation in a rehabilitation program.These factors limit access to physical therapy for SNF-resident dialysis patients and impair their outcomes. 10,13,14wo specialized acute rehabilitation facilities (Albany Medical Center 13 and Toronto Rehabilitation Institute) 14,15 demonstrated improvement of dialysis patient rehabilitation outcomes by coordinating the scheduling of physical therapy and dialysis.Both describe a solution whereby colocation of therapy and dialysis care eliminates enervating offsite transportation and avoids interference of dialysis with the physical therapy schedule.
It is a significant challenge to expand these highly tailored individual hospital programs to a national scale to broaden the number of patients who could benefit from them.Embedding a national program of in-SNF dialysis within a nationally recognized SNF chain and provider of rehabilitation therapy presents a potentially feasible model to replicate and scale up the success of these programs sustainably.We previously reported on the natural history of patients receiving five times weekly onsite dialysis within SNFs via a dedicated multiregional provider with a distributed model of care. 11These patients reported rapid post-dialysis recovery times, with 92% recovering within 2 h following treatment. 16hile patient report of recovery is assumed to parallel actual recovery with readiness to pursue normal daily activities, we asked whether this assumed relationship would be borne out by a functional test of recovery independent of patient report.
Within a SNF organization with a multistate footprint, among residents receiving rehabilitation services from a single provider (Signature Rehab), we compared physical therapy participation rates of ESRD patients receiving the onsite, more frequent hemodialysis (MFD) program against the general non-ESRD SNF population and against all other ESRD patients treated via offsite conventional dialysis programs.Could Dialyze Direct's onsite MFD restore SNF-resident ESRD patients within Signature Rehab to physical therapy participation rates in line with the general SNF Signature Rehab population and greater than that of conventional offsite dialysis ESRD Signature Rehab patients?

Setting
Signature HealthCARE is a family-based healthcare company offering post-acute care services in eight states, including skilled nursing, long-term care, and rehabilitation.Skilled nursing services are provided through an inter-disciplinary care approach covering a variety of medical conditions.Signature Rehab, an affiliate company, provides physical therapy, occupational therapy, and speech-language pathology services to the residents in Signature HealthCARE facilities.
Dialyze Direct is an onsite dialysis provider within SNFs operating in 13 states, utilizing a centralized administrative structure and distributed model of care.Pursuant to physicians' orders, Dialyze Direct's specialized dialysis nursing staff provides shorter and MFD treatment using NxStage System One, as described previously. 11ialysis care data are collected and stored by Dialyze Direct within two electronic medical records systems, Clarity (Visonex, Green Bay, WI) and Gaia (Gaia, Littleton, CO), feeding into a central data repository used for Quality Improvement and IRB-approved research.
Signature Rehab collects and stores patient data within two electronic medical records systems, Matrix-Care (MatrixCare, Bloomington, MN) and Net Health Therapy (NetHealth, Pittsburgh, PA).Each group employs a specialized data team to extract and centralize patient data from the electronic medical records.
We conducted a retrospective study of all patients receiving rehabilitation within all 105 Signature Health-CARE facilities from January 1, 2022 to June 1, 2022.This encompassed dialysis and non-dialysis patients.The facilities included five overlapping Dialyze Direct locations where dialysis care was provided on site using a MFD model of care, as well as 100 facilities where dialysis patients were sent offsite to receive conventional dialysis care in a location distinct from the site of their rehabilitation services.The technical differences among the models of dialysis care relate in large part to the number of dialysis treatments per week, with MFD typically prescribed five times per week and conventional dialysis prescribed three times per week.Additional differences in the dialysis groups' experience relate to the onsite or offsite designation.The offsite conventional dialysis group was potentially affected by the time and physical requirements of transportation, while the onsite more frequent model avoided this.
This study adhered to the principles outlined in the Declaration of Helsinki.The Institutional Review Board (WCG IRB, Puyallup, WA) ruled this study protocol exempt under 45 CFR § 46.104(d) (4) of the Common Rule with a full waiver of HIPAA authorization for use and disclosure of protected health information for this research.

Endpoint
To evaluate the extent of participation in rehabilitation services, "missed or shortened duration of physical therapy" was chosen as the prespecified composite endpoint of interest to define poor participation.Shortened therapy was defined as any daily sum of physical therapy >0 min and < 15 min.This definition was chosen by content experts at Signature Rehab prior to conducting any data analysis.Recorded reasons for missing therapy included hospitalization or being out of the SNF for a clinic visit, as well as a range of explanations including patient refusal, patient illness, and physician deferral.
All physical therapy session days were recorded in the Signature Rehab electronic medical records, including missed or shortened sessions.To construct our missed or shortened physical therapy metric, we summed all missed, shortened, or completed physical therapy session days as the denominator, with missed or shortened physical therapy session days as the numerator.
We excluded any missed physical therapy days where the reported reason was due to a hospitalization or scheduled appointment because the patient's non-participation on that day is unrelated to their physical readiness for physical therapy after dialysis, which was the focus of our study.
Active diagnoses in the Signature Rehab electronic medical records within 1 year prior or until 7 days after an admission between January 1, 2022 and June 1, 2022 were used to identify medical problems to construct a Charlson Comorbidity score using the comorbidity package in R. [17][18][19][20] Baseline levels of self-care and mobility were established from standard data collection in the SNF for Minimum Data Set reporting to CMS (Centers for Medicare & Medicaid Services) within the first 3 days of admission (Supporting Information Figures S1 and S2).We included those reports with the first seven (self-care) and six (mobility) questions completed, respectively, with scores between 1 (dependent) and 6 (independent), for maximum possible combined scores of 42 and 36, respectively.A value entry <1 or >6, indicating an unanswered question or mistaken entry, for any of these required questions excluded the subject from the baseline self-care and/or mobility analysis.
For Dialyze Direct patients, the interval of analysis for dialysis treatment frequency was the dialytic episode, defined as the patient's first in-SNF dialysis session, inclusive of all subsequent dialysis sessions, until treatment cessation at that facility due to hospitalization, death, transfer to another facility, transfer home, or withdrawal from dialysis.In a readmission to the SNF, the first dialysis session upon readmission initiates a new dialytic episode.Weekly dialysis frequency is established by counting the number of dialysis treatments each week but excluding the first and last week of a dialytic episode, since these weeks may be abbreviated by late entry or early discharge.
Post-dialysis recovery time information was collected by Dialyze Direct nursing staff at each dialysis session by asking the patient to report on their duration of recovery time to baseline following their previous dialysis session within discrete categories, as described previously. 16

Statistical analysis
Categorical variable proportions are compared with Chi-square test.Continuous comparisons of variables with two groups were compared with a t test, and for multiple groups, with linear regression.As the outcome of missed or shortened therapy sessions involved multiple physical therapy opportunities per patient, a mixed model logistic regression in Stata version 17 (College Station, TX) allowed control for multiple explanatory variables as well as the multiple observations per patient.This study therefore consisted of three groups for analysis: two dialysis groups with ESRD diagnosis, designated Offsite-Conventional-HD and Onsite-MFD, respectively, as well as the general Signature Rehab population without ESRD, designated Non-ESRD.Included patients are described in Figure 1 with the three analytical groups highlighted in gray.The two ESRD patient populations included both prevalent and incident patients.
Using a mixed model logistic regression with Onsite-MFD as the baseline population, we find that Offsite-Conventional-HD patients have greater odds of missed or shortened physical therapy compared to the Onsite-MFD patients (odds ratio: 1.8, CI: 1.1-3.0;p = 0.024).
Further mixed model logistic regression with covariates demonstrated the same finding and in addition demonstrated no statistical difference in odds of poor physical therapy participation between the Onsite-MFD population and the Non-ESRD population, where Onsite-MFD was the baseline group (odds ratio: 1.2, CI: 0.30-1.94;p = 0.46; Table 6).
As expected, worsening comorbidity with rising Charlson score resulted in progressively greater odds of poor physical therapy participation, but surprisingly, older age (compared to reference category age 0-50), non-White race, and Hispanic ethnicity each had statistically significantly lower odds of poor physical therapy participation.Even with removal of Charlson comorbidity adjustment from the model, the paradoxical observation of increasing age as protective against poor physical therapy participation persisted (Supporting Information Table S1).The self-care (Table 3) and mobility (Table 4) measures recorded in the first 3 days of SNF admission reveal that Onsite-MFD patients receiving rehabilitation were more frail (lower self-care) and had poorer mobility than both the general non-ESRD rehabilitation population (Non-ESRD) as well as the ESRD Offsite-Conventional-HD rehabilitation population, yet Onsite-MFD patients in rehabilitation had the highest physical therapy participation rates.
Among non-missed physical therapy sessions only, duration of physical therapy for Offsite-Conventional-HD was not significantly different from NonESRD via t test (p = 0.22).Onsite-MFD had longer physical therapy duration in non-missed sessions compared to Offsite-Conventional-HD (1.4 min, CI: 0.7-2.1;p = 0.0002) and compared to Non-ESRD (0.97 min, CI: 0.6-1.3;p < 0.00001), but mixed model regression controlling for covariates of age category, race/ethnicity, payor, and Charlson score failed to show any significant difference in session duration between the three groups.
The Signature Rehab electronic medical records do not record physical therapy start times, and dialysis start time was only available for 73/92 patients (80%) in 1334 dialyses recorded in the Dialyze Direct EMR, so we cannot say with certainty what percent of Onsite-MFD physical therapy sessions occurred, or were "missed or shortened," following completion of a scheduled dialysis session on the same day.The cumulative percent of dialyses started before 8, 9, 10, and 11 a.m. were 49%, 66%, 79%, 86%, respectively, suggesting a substantial proportion likely received physical therapy after a day's scheduled dialysis.
In the 73 Onsite-MFD patients for whom reports on post-dialysis recovery time were available (spanning 1207 dialyses), 1200/1207 (99.4%) reported recovery to baseline in 2 h or less.2][23] The Onsite-MFD group received on average 4.5 dialyses per week, while those in the Offsite-Conventional-HD group were prescribed dialysis three times per week.The finding of improved physical therapy participation rates within the onsite dialysis program is strengthened by an inherent bias in the study design favoring the offsite conventional dialysis group, against our observation.
Missed physical therapy sessions caused by scheduling conflict with offsite dialysis treatment are often recorded with the reason of "outside clinic visit."Our exclusion criteria eliminated missed sessions due to the reason of "outside clinic visit," artificially lowering the number of missed sessions for the offsite dialysis group.Had the same event occurred in the onsite MFD group, it would have been classified as a missed session.This suggests that the true success at increasing physical therapy participation by onsite MFD was likely greater than described.When patients repeatedly miss physical therapy sessions due to exhaustion following their offsite dialysis care, providers often respond by discontinuing orders for physical therapy on dialysis days, thereby reducing the number of missed or shortened physical therapy sessions.While we could not easily ascertain the physical therapy ordering pattern to assess this question thoroughly, this additional bias would likely favor the offsite conventional dialysis group and operated against the onsite MFD group, further reinforcing our findings.
By the metrics of self-care and mobility, Onsite-MFD patients were significantly frailer than the general non-ESRD skilled nursing facility rehabilitation residents and the Offsite-Conventional-HD patients.Increased frailty in the onsite MFD group would also bias against their better performance on this participation metric.
Farragher et al. 15 cites short daily dialysis as one of the key ingredients to the success of its integrated onsite dialysis and rehabilitation program for seniors.Previous studies also showed a wide array of benefits of more frequent hemodialysis [24][25][26] including reduced post-dialysis recovery time, reflecting reduced post-dialysis fatigue, which should play a significant role in the ability to attend and fully participate in physical therapy sessions.We previously showed 92% recovery within 2 h of dialysis in the Dialyze Direct onsite MFD population, 16 and in the present study, we find a similarly high 99% reported recovery to baseline within 2 h.Some rehabilitation programs utilizing conventional thrice weekly dialysis even build in a diminished therapy time on dialysis treatment days by default. 27While our study showed a clear benefit on physical therapy participation with onsite MFD versus offsite conventional dialysis, we cannot determine the individual contributions of dialysis frequency and colocalization/schedule coordination to this improvement.
As expected, greater comorbidities were associated with worsened physical therapy participation.Surprisingly, increasing age was protective against this metric.We tested the persistence of this finding by removing from the model the Charlson comorbidity score, which we suspected might have been absorbing much of the increased risk that should have been apparent with increasing age.However, age remained protective.Physical therapy goals are tailored to the individual patient, irrespective of age, with a focus on restoring the individual's baseline function, which existed prior to a deteriorating event.We look forward to additional research to further explore this unexpected finding of a relationship between older age and greater physical therapy participation rates.
The observed protective effect of Black race or Hispanic ethnicity on physical therapy participation rates is also unexplained.A prior study 28 in a different setting found a seemingly opposite relationship, showing an association of Black and Hispanic race with fewer physical and occupational therapy minutes per week at a single safety net hospital.That relationship persisted after the authors adjusted for missed therapy days as a covariate, but the study did not elaborate on participation rates by race/ethnicity or the relationship between missed sessions and total therapy minutes received.Inpatient hospital rehabilitation could also differ in important ways from a SNF rehabilitation program.Further research is needed to elaborate on our finding of association between fewer missed or shortened physical therapy sessions and Black race or Hispanic ethnicity.An implication of our findings is that SNFs can accept frailer, less mobile patients by embedding dialysis care on site with schedule-coordinated physical therapy, helping hospitals and patients.Hospitals may be able to discharge patients earlier, reducing length of stay, and patients benefit from leaving the acute hospital environment with its infections and other iatrogenic complications.The clinical metrics and first laboratory values of Onsite-MFD patients shown in Table 5 confirm that a considerable percentage of these patients are highly clinically compromised with high disease burden upon hospital discharge.Our centrally supervised care model enables early discharge of these patients from the hospital to begin their rehabilitation sooner in a more appropriate, less acute setting.
The major limitation of our study is its observational nature.Assignment of dialysis program was not randomized, and better participation rates in facilities incorporating an onsite dialysis program may have been related to rehabilitation program or leadership effectiveness differences in those facilities.While previous studies 28,29 have linked participation levels and therapy duration with rehabilitation outcomes, patient physical therapy outcomes were not assessed in this study.
The strength of the study is its "real world" report in a multistate chain of nursing homes with dialysis patients distributed across the network and onsite more frequent hemodialysis patients distributed within five of those nursing homes.

CONCLUSION
Within a national nursing home chain, ESRD rehabilitation patients provided onsite MFD can engage in physical therapy, without shortening or missing sessions, equally to the general non-ESRD nursing home rehabilitation population and better than ESRD rehabilitation patients provided conventional 3Â/week dialysis offsite.What was suggested over a decade ago under specialized conditions can be replicated in the real world on a large scale by a distributed dialysis system within a skilled nursing facility rehabilitation chain.
January 1, 2022 to June 1, 2022, 105 Dialyze Direct dialysis patients (denoted Sig-DD) and 13,478 non-Dialyze Direct (non-DD) patients (denoted Sig-Non-DD) (Figure 1) were scheduled for 2326 and 234,578 physical therapy sessions, respectively, within Signature HealthCARE SNFs.To identify those within Signature Rehab receiving conventional offsite dialysis (not under Dialyze Direct's care), we used ESRD ICD10 code N18.6 as the selection criterion.Of 13,478 non-DD patients, 562 (4.2%) had ESRD diagnosis (Offsite-Conventional-HD). To assure comparability between these non-DD Signature Rehab patients with ESRD (Offsite-Conventional-HD) and the Dialyze Direct patients receiving onsite MFD, we restricted the analysis of Dialyze Direct patients to only those who were identified in the Signature Rehab electronic medical records with ESRD diagnosis and designated them Onsite-MFD.Diagnosis code entry within 7 days is an imperfect measurement since only 92/105 (88%) Dialyze Direct patients had the ESRD ICD10 code (Onsite-MFD).
The demographics and key variables describing the three populations of interest, Non-ESRD, Offsite-Conventional-HD, and Onsite-MFD, are provided in Table2.Baseline mean ± SD Charlson Comorbidity score was significantly F I G U R E 1 Included patients.NonESRD, Signature Rehab patients without ESRD (No. ICD10 N18.6; not under Dialyze Direct's care); Offsite-Conventional-HD, Signature Rehab ESRD (ICD10 N18.6) patients receiving conventional hemodialysis offsite (not under Dialyze Direct's care); Onsite-MFD, Dialyze Direct patients in Signature Rehab with ESRD (ICD10 N18.6) receiving onsite, more frequent hemodialysis; Sig-DD, Signature Rehab patients under the care of Dialyze Direct; Signature Rehab, entire Signature Rehab population; Sig-Non-DD, Signature Rehab patients not under Dialyze Direct's care.
Patient and Scheduled Physical Therapy SessionCounts in Three Signature Rehab Populations.Baseline clinical and demographic characteristics for each of the three groups of rehabilitation participants.Data are presented as n (%) unless otherwise indicated.Bold denotes statistically significant, p-value < 0.05.p-values obtained by linear regression for continuous variables and chi-square test for binary/categorical variables.
T A B L E 1DISCUSSIONESRD patients residing in SNFs treated with onsite MFD experienced a significantly decreased percentage of missed or shortened physical therapy sessions compared to those receiving conventional offsite hemodialysis (HD).Patients receiving onsite MFD were also indistinguishable on this predetermined physical therapy participation metric from the general (non-ESRD) SNF rehabilitation population, with or without controlling for comorbidities.The significantly improved physical therapy participation rates observed for onsite MFD patients compared to offsite conventional HD patients was not surprising given T A B L E 2 Note: T A B L E 3 Baseline self-care score (sum) by Rehabilitation Participant Group.Maximum potential score from seven items was 42.Median 20, interquartile range: 15-26 across entire population.Abbreviations: CI, 95% confidence interval; ESRD, end-stage renal disease; HD, hemodialysis; MFD, more frequent hemodialysis; non-ESRD, no ESRD diagnosis; Offsite-Conventional-HD, ESRD patient provided conventional hemodialysis offsite; Onsite-MFD, ESRD patient provided more frequent hemodialysis onsite (by Dialyze Direct); SD, standard deviation.Maximum potential score from six items was 36.Median 21, interquartile range: 15-29 across entire population.Abbreviations: CI, 95% confidence interval; ESRD, end-stage renal disease; HD, hemodialysis; MFD, more frequent hemodialysis; non-ESRD, no ESRD diagnosis; Offsite-Conventional-HD, ESRD patient provided conventional hemodialysis offsite; Onsite-MFD, ESRD patient provided more frequent hemodialysis onsite (by Dialyze Direct); SD, standard deviation.First laboratory values and clinical metrics for end-stage renal disease patients receiving onsite more frequent hemodialysis (Onsite-MFD).Note: Values obtained from first recorded measurement during study period.Dialysis Time in treatment minutes.Abbreviations: FeSat, iron saturation; Hgb, hemoglobin; SD, standard deviation; stdKt/V, standard Kt/V; UFR, ultrafiltration rate.
Note:Note:T A B L E 5 T A B L E 6 Odds of missed or shortened physical therapy by mixed model logistic regression.Bold denotes statistically significant, p-value < 0.05.Abbreviations: ESRD, end-stage renal disease; Non-ESRD, Skilled nursing facility resident in rehabilitation without ESRD diagnosis (non-Dialyze Direct); Offsite-Conventional-HD, Skilled nursing facility-resident ESRD patients in rehabilitation provided offsite, conventional dialysis by other dialysis providers; Onsite-MFD, Skilled nursing facility-resident ESRD patients in rehabilitation provided onsite, more frequent dialysis by Dialyze Direct. Note: