1 712 ORIGINAL ARTICLE A Comparison of Discharge Functional Status After Rehabilitation in Skilled Nursing, Home Health, and Medical Rehabilitation Settings for Patients After Lower-Extremity Joint Replacement Surgery Trudy R. Mallinson, PhD, OTR/L, NZROT, Jillian Bateman, OTD, OTR/L, Hsiang-Yi Tseng, MA, OTR/L, Larry Manheim, PhD, Orit Almagor, MA, Anne Deutsch, PhD, CRRN, Allen W. Heinemann, PhD, ABPP ABSTRACT. Mallinson TR, Bateman J, Tseng H-Y, Manheim L, Almagor O, Deutsch A, Heinemann AW. A comparison of discharge functional status after rehabilitation in skilled nursing, home health, and medical rehabilitation settings for patients after lower-extremity joint replacement surgery. Arch Phys Med Rehabil 2011;92: Objective: To examine differences in outcomes of patients after lower-extremity joint replacement across 3 post acute care (PAC) rehabilitation settings. Design: Prospective observational cohort study. Setting: Skilled nursing facilities (SNFs; n 5), inpatient rehabilitation facilities (IRFs; n 4), and home health agencies (HHAs; n 6) from 11 states. Participants: Patients with total knee (n 146) or total hip replacement (n 84) not related to traumatic injury. Interventions: None. Main Outcome Measure: Self-care and mobility status at PAC discharge measured by using the Inpatient Rehabilitation Facility Patient Assessment Instrument. Results: Based on our study sample, HHA patients were significantly less dependent than SNF and IRF patients at admission and discharge in self-care and mobility. IRF and SNF patients had similar mobility levels at admission and discharge and similar self-care at admission, but SNF patients were more independent in self-care at discharge. After controlling for differences in patient severity and length of stay in multivariate analyses, HHA setting was not a significant predictor of self-care discharge status, suggesting that HHA patients were less medically complex than SNF and IRF patients. IRF patients were more dependent in discharge self-care even after controlling for severity. For the full discharge mobility regression model, urinary incontinence was the only significant covariate. Conclusions: For the patients in our U.S.-based study, direct discharge to home with home care was the optimal strategy for patients after total joint replacement surgery who were healthy and had social support. For sicker patients, availability of 24-hour medical and nursing care may be needed, but intensive therapy services did not seem to provide additional improvement in functional recovery in these patients. Key Words: Arthroplasty, replacement, hip; Arthroplasty, replacement, knee; Recovery of function; Rehabilitation; Skilled nursing facilities by the American Congress of Rehabilitation Medicine LOWER-EXTREMITY JOINT replacement is a rapidly growing surgical procedure in the United States. 1 Between 1990 and 2002, the rate of primary THRs per 100,000 persons increased by approximately 50%, while the rate of primary TKRs increased almost 3-fold. 2 In 2003, a total of 205,500 primary THRs and 402,100 primary TKRs were performed in the United States. During the same year, 36,000 revision THRs and 32,700 revision TKRs were performed. 3 The aging of the population, increased rates of obesity, and increased prevalence of arthritis is are creating an increased demand for lower-limb joint replacements. 4 Immediately after acute hospital discharge, 87% of Medicare beneficiaries receiving lower-extremity joint replacement surgery use postacute rehabilitation services. Approximately 37% of these patients use SNF care, 19% use IRF care, 36% return home with home health care, and 7% receive rehabilitation in List of Abbreviations From the Department of Occupational Science and Occupational Therapy, the Herman Ostrow School of Dentistry, University of Southern California, Los Angeles, CA (Mallinson); Rehabilitation Institute of Chicago, Chicago, IL (Bateman, Tseng, Deutsch, Heinemann); Department of Physical Medicine and Rehabilitation (Mallinson, Manheim, Deutsch, Heinemann), Department of Rheumatology (Almagor), and Institute for Healthcare Studies (Manheim, Almagor, Deutsch, Heinemann), Feinberg School of Medicine, Northwestern University, Chicago, IL. Supported by the National Institute on Disability and Rehabilitation Research Rehabilitation Research and Training Center on Measuring Rehabilitation Outcomes and Effectiveness (grant no. H133B040032). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Reprint requests to Trudy Mallinson, PhD, OTR/L, NZROT, University of Southern California, Dept of Occupational Science and Occupational Therapy, 1540 Alcazar St, CHP 133, Los Angeles, CA 90089, usc.edu /11/ $36.00/0 doi: /j.apmr ANOVA HHA HSD IRF IRF-PAI LOS MedPAC MDS OASIS OT PAC PT SNF THR TJR TKR analysis of variance home health agency honestly significantly difference inpatient rehabilitation facility Inpatient Rehabilitation Facility Patient Assessment Instrument length of stay Medicare Payment Advisory Commission Minimum Data Set Outcomes and Assessment Information Set occupational therapy postacute care physical therapy skilled nursing facility total hip replacement total joint replacement total knee replacement
2 COMPARING POSTACUTE CARE SETTINGS, Mallinson 713 outpatient settings. 5 It is estimated that $3.2 billion is spent annually on rehabilitation after lower-extremity arthroplasty. 6 Medicare is the primary payer of lower-extremity joint replacements, covering 60% of these surgeries 7 and up to 85% of subsequent rehabilitation services. 8 Each type of PAC provider has unique regulatory requirements, including rules regarding the availability and provision of rehabilitation therapy services. To be Medicare certified, IRFs must provide a minimum of 3 hours of therapy at least 5 days a week. SNFs are not mandated to provide a specific intensity or type of therapy program, but are reimbursed for providing patients with a given number of therapy minutes based on the resource utilization group classification system. 9,10 Some SNFs provide average amounts of total therapy similar to those of IRFs. 11 The amount of therapy a patient receives in home health care is based on discipline-specific evaluations. In 2001, patients with lower-extremity joint replacement received on average 14.1 home health visits (including therapy and nursing visits) for a 120-day episode, although this mean value varies by region from 9.8 to 17.1 visits an episode. 12 Characteristics of patients with lower-extremity joint replacement treated in the 3 PAC settings vary, although other studies have suggested that there may be some substitution across PAC settings. 5 SNF patients typically are older than patients treated in IRFs 13,14 and HHAs. 15,16 In general, patients with lower functional independence are more likely to be admitted to IRFs than SNFs. 14 The literature is mixed regarding the prevalence of comorbid conditions by type of PAC. For example, DeJong et al 14 found that IRF patients had more comorbid conditions, whereas Buntin 13 and Gage 17 and colleagues found this to be the case for patients with joint replacement admitted to SNFs. These differences may reflect how comorbid conditions are defined within each study. Facilitylevel and geographic factors also influence PAC use. Patients in acute-care hospitals with a PAC subprovider or colocated PAC are more likely to be discharged to that subprovider In addition, PAC settings are not distributed equally across geographic regions. 17,18 Studies that compared outcomes for patients with lowerextremity joint replacement across rehabilitation settings generally have focused on comparing only 2 of the available PAC providers (IRF vs SNF, IRF vs HHA, SNF vs HHA). Walsh and Herbold 20 reported that patients admitted in 2004 with a TJR and treated in a large free-standing IRF experienced shorter LOSs and achieved functional outcomes superior to those treated in subacute units in 5 SNFs in the same region. DeJong et al 21 found that patients with TJR experienced better outcomes in IRFs compared with SNFs. Other studies have reported that discharging a patient with TJR to an IRF versus home with HHA services did not result in significant improvement in functional outcomes No studies have compared outcomes of patients with TJR across the 3 primary PAC settings (IRF, SNF, HHA) to permit a side-by-side comparison of benefits. Therefore, the purpose of this study was to examine differences in outcomes for patients with TJR across the 3 most commonly used PAC rehabilitation providers. Study questions were as follows. (1) In what ways do patients vary on key demographic and clinical factors at admission across PAC settings? (2) To what extent does the type of PAC provider influence selfcare and mobility function at discharge, after controlling for patient characteristics and LOS? METHODS Study Design This was a multicenter prospective observational cohort study including 4 IRFs, 5 SNFs, and 6 HHAs from 11 states. A nurse from each PAC provider served as the data collector. Data were collected by nurses to avoid differences in rater severity between nurses and therapists 25 and because they often were responsible for completing patient assessments. Each nurse had at least 3 years of experience and participated in a training session covering the data collection instruments and protocol, research ethics, and consent procedures. Facilities Each provider volunteered to participate from solicitations posted on listservs and recommendations from colleagues. We selected facilities in several regions and did not include facilities from states known to have practice patterns distinctly different from the rest of the county. Participation depended in part on the providers having sufficient numbers of eligible patients to complete data collection in a timely manner. In addition, because data collection was extensive (information from all 3 PAC federally mandated assessment tools for each patient), the facility needed to be able to dedicate nursing time to the project. Facilities were reimbursed for each completed data collection packet. Patient Selection Eligibility criteria included (1) primary diagnosis of hip or knee replacement or revision, (2) aged 65 years or older, (3) admitted to the PAC setting directly from an acute-care hospital, and (4) receiving rehabilitative therapy, including PT and/or OT. A fifth criterion, that Medicare fee-for-service be the primary payer for the patient, was revised toward the end of data collection because several HHAs and SNFs experienced an increase in their managed care Medicare population. Data for patients with joint replacement used in this study were collected almost entirely before eligibility criteria were revised. Patients receiving a TJR after hip fracture were excluded. Patients became ineligible if they were readmitted to an acutecare setting for more than 48 hours during their PAC stay. A total of 241 patients with TJR were enrolled; 11 were ineligible because of readmission to an acute-care hospital (n 5), admission from another PAC setting (n 2), not having Medicare fee-for-service as the primary insurance (n 1), incorrect data collection procedures (n 1), and patient withdrawal of consent (n 2). Analyses were completed for 230 patients. Data Collection/Instrumentation The study protocol was reviewed and approved by Northwestern University s Institutional Review Board and each of the participating providers review boards. Patient recruitment occurred from December 2005 through December Patients who met study inclusion criteria were approached and consented by a study nurse. For patients with cognitive impairment, informed consent was obtained from a legally authorized representative or next of kin. Data were collected for each patient within 48 hours after admission and 48 hours before discharge. Data collection consisted of an interview, observation, medical chart review, and correspondence with the patient s health care team. Data elements included demographic, social support, and medical information, including comorbid conditions, type of replacement, and revision status. All patients were scored on medical and functional items from the IRF-PAI, MDS 2.0, and OASIS. We
3 714 COMPARING POSTACUTE CARE SETTINGS, Mallinson used MDS 2.0 items for patients vision status, cognitive function, and bowel and bladder function. For this report, functional status was reported from IRF-PAI items. Demographic characteristics included sex, race, and age. Social support included marital status, living location, and living situation. Health conditions included time from surgery, type of joint replacement, revision status, number of comorbid conditions, obesity, visual impairment, cognition, and bowel and bladder continence. Cognitive skills included short- and long-term memory and daily decision making. These items were collected by using items from the MDS 2.0. Memory items were dichotomous, indicating the presence or absence of a problem. Although decision making was rated on a 4-point scale reflecting the degree of difficulty making decisions, we dichotomized it as either independent (score of 0) or problem (score of 1, 2, or 3). Bowel and bladder continence items were scored on a 4-point categorical scale reflecting level of incontinence; we dichotomized these items to either continent (score of 0) or incontinent (score of 1, 2, or 3). International Classification of Diseases-9th Revision-Clinical Modification codes were used to report the number of obese patients and number of other comorbid conditions. A maximum of 10 comorbid conditions could be recorded on the data collection forms. Nurses reported comorbid conditions in the same order that they appeared in the medical record or assessment for that setting. After discharge, the study nurse documented discharge destination, LOS, and number of days, hours, and minutes patients received of each therapy discipline during their PAC stays. Data for therapy minutes were abstracted from the medical record. Time from surgery was calculated as number of days from the date of surgery to date of admission to the PAC setting. Therapy Measures Therapy data were collected for OT, PT, speech language pathology, psychology, and other therapy (eg, recreational). Therapy was reported in several ways. Number of therapy days was the number of days a patient received therapy from at least 1 discipline. Total therapy minutes were the total number of rehabilitation minutes received from a particular discipline during the admission. Discipline intensity was the total number of therapy minutes divided by the number of days on which therapy occurred; it is an indicator of how much therapy a patient received from a discipline a day. Because the specific days on which therapy occurred were unavailable, we did not report therapy per day across all disciplines. Overall therapy intensity was calculated as total therapy minutes divided by LOS; it is an indicator of the intensity of therapy averaged across all disciplines and the entire stay. Functional Status Nurses were instructed to record the most dependent functional performance in the first 48 hours of admission and last 48 hours before discharge using the IRF-PAI. Facilities typically assess functional status multiple times during the admission and discharge assessment periods; they report the lowest scores, reflecting the most dependent performance during that period, consistent with IRF-PAI instructions. We distinguished self-care and mobility components of functional status. The IRF-PAI includes the 18 FIM items that cover self-care, sphincter control, transfers, locomotion, communication, and social cognition; in addition, there are 10 function modifiers. We analyzed 6 self-care items, including eating, grooming, bathing, dressing upper body, dressing lower body, and toileting, and 7 mobility items, including tub transfer (modifier), shower transfer (modifier), bed-chair transfer, toilet transfer, walking (modifier), wheelchair (modifier) and climbing stairs. FIM items and function modifiers are scored on a 7-point rating scale from 1 (total dependence) to 7 (complete independence). Because of differences in levels of difficulty, we used the modifiers, which separated tub/shower transfer and walk/wheel items. We used Rasch analysis to convert raw self-care and mobility scores to interval-level measures using the Winsteps program, Version a Admission and discharge ratings were available for each patient. To avoid violation of the assumption of local independence, 26 estimates of item difficulty locations were established by randomly selecting half the admission records and combining these with discharge records from the other half of the sample. Thus, estimates of item difficulty were obtained across the full range of function from admission to discharge while no patient contributed more than 1 data point to the estimates. We used estimates of item difficulty from this initial analysis to anchor item locations 27 in the subsequent analysis, from which we obtained estimates of function for all patients at both admission and discharge. Rasch measures were reported as logits 28 ; we transformed the logits to match the raw score range for each scale to enhance interpretability. Thus, self-care logit measures ranged from 6 to 42, and mobility logit measures ranged from 7 to 49. Analysis Data were analyzed using Stata, Version b After an overall description of the sample, we compared patients admitted to each PAC settings, testing for significant differences by using chi-square test of proportions or 1-way ANOVA for categorical and continuous variables, respectively. Differences among settings for ANOVAs were tested by using Tukey HSD test. Because we were interested in the effect of setting on functional outcomes, we conducted a series of hierarchical linear regression models, progressively examining the effect of setting after controlling for patient demographics (age at admission, sex), impairment severity (procedure, revision status, number of comorbid conditions, urinary and fecal continence status, mobility at admission, self-care at admission), and LOS. We computed models separately by level of functional status because factors that explain outcomes for high-functioning patients might be different from those explaining outcomes for low-functioning patients. We split high- and low-functioning patients based on toileting score for the self-care scale ( 5 for high functioning) and based on toilet transfer score for the mobility scale ( 5 for high functioning). These cutoff points were based on reports from clinicians who described these factors as important for discharge planning. No differences were found by level of function; therefore, we presented only results for models across all patients. Clinicians suggested that patients in different settings might differ significantly by comorbid conditions that would affect their ability to engage in and benefit from rehabilitation; specifically, anemia, diabetes, and mood disorders. We computed models that included these comorbid conditions, but found no difference in results. Therefore, models with these comorbid conditions were not reported. We computed models separately by mobility and self-care because we found evidence of multicollinearity when both self-care and mobility at admission were included in a single model. Consequently, we did not include mobility at admission in self-care models or self-care at admission in mobility models. We also tested models that included either total therapy minutes by discipline and therapy intensity (total minutes a day) as covariates. These covariates were not statistically significant; therefore, we did not report those models.
4 COMPARING POSTACUTE CARE SETTINGS, Mallinson 715 Table 1: Demographic Characteristics, Health Conditions, and Functional Status Variable SNF (n 83) IRF (n 76) HHA (n 71) Total (N 230) Patient demographics Sex Women 66 (79.5) 53 (69.7) 44 (62.0) 163 (70.9) Men 17 (20.5) 23 (30.3) 27 (38.0) 67 (29.1) Race White 82 (98.8) 69 (90.8) 68 (95.8) 219 (95.2) Black 0 (0.0) 6 (7.9) 3 (4.2) 9 (3.9) Other 1 (1.2) 1 (1.3) 0 (0.0) 2 (0.8) Marital status Single 6 (7.2) 5 (6.6) 1 (1.4) 12 (5.2) Married 39 (47.0) 40 (52.6) 54 (76.1)* 133 (57.8) Widowed 33 (39.8)* 21 (27.6) 13 (18.3)* 67 (29.1) Separated/divorced 5 (6.0) 8 (10.5) 3 (4.2) 16 (7.0) Missing 0 (0.0) 2 (2.6) 0 (0.0) 2 (0.8) Living location before episode Home 82 (98.8) 73 (96.1) 71 (100.0) 226 (98.3) Board & care (assisted living) 0 (0.0) 3 (3.9) 0 (0.0) 3 (1.3) Missing 1 (1.2) 0 (0.0) 0 (0.0) 1 (0.4) Living situation Lives alone 41 (49.4)* 22 (28.9) 14 (19.7)* 77 (33.5) Lives with others 42 (50.6)* 53 (69.7) 57 (80.3)* 152 (66.3) Missing 0 (0.0) 1 (1.3) 0 (0.0) 1 (0.4) Age (y) * Health conditions Time from surgery (d) * Joint replacement Hip replacement 35 (42.2) 26 (34.2) 23 (32.4) 84 (36.8) Knee replacement 48 (57.8) 50 (65.8) 48 (67.6) 146 (63.2) Revision 4 (4.8) 4 (5.3) 2 (2.8) 10 (4.4) No. of comorbid conditions * * * Obesity 2 (2.41)* 14 (18.42)* 3 (4.23) 19 (8.26) Impaired vision, MDS (8.43) 21 (27.63)* 1 (1.41)* 29 (12.61) Cognition at admission, MDS 2.0 Problems with short-term memory 14 (16.9) 4 (5.3) 10 (14.1) 28 (12.2) Problems with long-term memory (0.0) (0.0) 1 (1.4) 1 (0.4) Problems with daily decision making 32 (38.6)* 13 (17.1)* 21 (29.6) 66 (28.7) Bladder continence, MDS 2.0 Incontinent 16 (19.3) 13 (17.1) 27 (38.0)* 56 (24.3) Bowel continence, MDS 2.0 Incontinent 3 (3.6) 3 (3.9) 7 (9.9)* 13 (5.7) Functional status Self-care at admission, IRF-PAI * Mobility at admission, IRF-PAI * NOTE. Values expressed as no. (%) or mean SD. Chi-square tests or ANOVAs are used for categorical and continuous variables, respectively. Differences among settings for ANOVAS were tested by using the Tukey HSD test. *Significant values (P.05). Self-care and mobility data from the IRF-PAI: self-care (6 items: eat, groom, bath, toilet, upper body dressing, lower body dressing), mobility (7 items: bed transfer, toilet transfer, stairs, tub transfer, shower transfer, walking, wheeling). Self-care measures range from 6 to 42; mobility measures, from 7 to 49, with higher scores indicating more independent function. Cognitive function and bladder/ bowel continence data are from the MDS 2.0. RESULTS Table 1 lists patient characteristics by PAC provider. Of 230 patients, 63.2% had a knee replacement and 36.8% had a hip replacement; revisions made up 4.4% of the sample. Overall, most patients were white, women, and lived at home before admission to the acute-care hospital. Key differences by setting at admission included the following: HHA patients were more likely to be married, live with others, have a longer time from surgery, and more frequently be incontinent than either IRF or SNF patients. SNF patients were older and more likely to be widowed, live alone, and have more problems with daily decision making. IRF patients had a greater number of comorbid conditions and greater proportions were obese or had vision problems than HHA or SNF patients. HHA patients were less dependent in mobility and self-care function at admission than IRF or SNF patients. IRF and SNF patients did not differ significantly in functional status at admission. Table 2 lists facility characteristics. Half the HHAs were hospital based and half were freestanding. All IRFs were freestanding. One SNF was hospital based and 4 were freestanding. With only a few exceptions, there were at least 4 SNFs and 5 HHAs within 10 miles of each PAC site, whereas
5 716 COMPARING POSTACUTE CARE SETTINGS, Mallinson Facility Type Medicare/ Census Region Licensed Beds Annual Visits in 2008 (for HHAs) Table 2: Descriptors of Data Collection Facilities Profit Status Approximate No. of Medicare Discharges 2007 FS vs HB No. of SNFs Within 10-Mile Radius No. of IRFs Within 10-Mile Radius No. of HHAs Within 10-Mile Radius HHA VII/Midwest NA 120,324 Voluntary not-for-profit, private 2445 FS HHA I/Northeast NA 29,398 Voluntary not-for-profit, other 4299 FS HHA VII/Midwest NA 67,769 Proprietary, other 9976 FS HHA II/Northeast NA 130,000 Voluntary not-for-profit, other 6900 HB HHA V/Midwest NA 60,000 Voluntary not-for-profit, other 4001 HB HHA III/Northeast NA 189,503 Voluntary not-for-profit, other 33,797 HB IRF V/Midwest 155 NA Voluntary not-for-profit, other 2480 FS IRF V/Midwest 120 NA Voluntary not-for-profit, church 2443 FS IRF VII/Midwest 72 NA Proprietary, other 817 FS IRF II/Northeast 322 (3 sites) NA Proprietary, corporation 6406 FS SNF VIII/West 50 NA Proprietary, other 450 FS SNF V/Midwest 103 (2 sites) NA Voluntary not-for-profit, church 558 FS SNF V/Midwest 43 NA Voluntary not-for-profit, other 200 FS SNF VII/Midwest 17 NA Proprietary, other 340 HB SNF II/Northeast 22 NA Voluntary not-for-profit, other 464 FS Abbreviations: HB, hospital based; FS, free standing; NA, not applicable. there often was only 1 other IRF; for 1 HHA and 1 SNF, there were no IRFs. Except for these supply constraints, patients had access to all 3 types of PAC. Table 3 lists patients LOSs, therapy minutes, and discharge locations by type of provider. HHA patients had significantly longer lengths of treatment and received significantly fewer PT and OT total minutes compared with IRF and SNF patients. Few (n 12) HHA patients received OT services. IRF and SNF patients had similar LOSs and total number of days on which therapy was received, but IRF patients received significantly more OT, PT, psychology, and total therapy minutes. IRF patients received a greater number of minutes a day (therapy intensity). Almost all IRF and SNF patients were discharged to home with additional PAC services, including HHA and outpatient services. Only 20% of patients from SNFs received additional outpatient therapy, whereas 37% of HHA and 42% of IRF patients were discharged to these services. Table 4 lists unadjusted mean mobility and self-care measures at discharge by setting and unadjusted functional change from admission to discharge. There were significant differences among settings for self-care function at admission (F 2, ; P.0001) and discharge (F 2, ; P.0001). HHA patients were significantly less dependent than SNF and IRF patients at admission, but significant differences among settings were present at discharge. Mobility differed across settings at admission (F 2, ) (P.0001; see table 1) and discharge (F 2, ; P.0001) (see table 4), with Tukey HSD indicating that HHA patients were significantly less dependent in self-care than SNF and IRF patients at both times. For the hierarchical regression self-care model (table 5), setting alone explained 16% of the variance, with patients in IRF and HHA settings significantly different from SNF patients (the reference category) in self-care at discharge. Adding patient demographic characteristics accounted for little additional variance. Procedure, comorbid conditions, and admission selfcare status explained an additional 29% of the variance and the HHA coefficient was no longer significant, suggesting that significant findings for HHA patients were due to these patients being less medically and functionally complex than SNF and IRF patients. LOS added little to the explained variance in self-care, suggesting that the longer LOS in HHAs was not a significant factor in functional outcomes. IRF patients had greater dependency in discharge self-care even after controlling for demographic and illness factors. For mobility (table 6), setting alone explained 19% of the variance, with only HHA patients significantly different in mobility at discharge. Adding patient demographic characteristics added little to the model. There was a marked decrease in size of the HHA coefficient with the addition of procedure, comorbid conditions, and admission mobility status, reflecting that HHA patients were significantly less complex than SNF or IRF patients. Urinary incontinence was the only statistically significant comorbidity after controlling for mobility at admission. The IRF coefficient approached significance with the addition of comorbid conditions and mobility at admission, indicating that the lower discharge mobility scores for IRF patients were not due to lower (measurable) functioning at admission (see table 1). DISCUSSION Based on our study sample, 3 main findings emerged. (1) Different PAC providers admitted different kinds of patients with lower-extremity joint replacement; (2) for healthier, less dependent patients with social support, discharge to the home with home health care appeared to be the optimal setting; and (3) for other patients, the more therapy-intensive IRF setting did not appear to offer an advantage over the SNF setting in terms of functional outcomes. Different types of PAC providers admitted different kinds of patients with lower-extremity joint replacement. In general, HHAs admitted patients who needed less assistance with mobility, were younger, had more social support, and had fewer comorbid conditions. IRF patients had more comorbid conditions and higher prevalences of obesity and impaired vision. SNF patients were more likely to live alone and have problems with decision making. Other investigators 6,19,29-31 also reported that compared with patients with TJR discharged to home or home care, institutional PAC users tended to be older. Within institutional TJR PAC users, SNF patients tended to be older than IRF patients. 16,18 The literature also suggested that overall institutional TJR PAC users were more likely to be women, 16,19,24,31-34 and SNF patients were more likely to be women than IRF patients. 21 Compared with patients with TJR
6 COMPARING POSTACUTE CARE SETTINGS, Mallinson 717 Table 3: LOS, Therapy Minutes, Therapy Intensity, and Discharge Location SNF (n 83) IRF (n 76) HHA (n 71) Total (n 230) LOS (d)* 10 (7 13) n 10 (7 13) n 19 (14 24) n 11 (8 16) n Total therapy min OT* PT* SLP Respiratory therapy Psychology* Other* Total* No. of therapy days OT* PT* SLP Respiratory therapy Psychology* Other* Discipline intensity (therapy min/ no. of days) OT* PT* SLP Respiratory therapy Psychology* Other* Overall therapy intensity (therapy min/los)* Discharge location* Home w/o therapy 18 (21.7) 0 (0.0) 43 (60.6) 61 (26.4) Home with HHA 45 (54.2) 39 (51.3) 2 (2.8) 86 (37.2) Home with OP 17 (20.5) 32 (42.1) 26 (36.6) 75 (32.5) Other 3 (3.6) 4 (5.3) 0 (0.0) 7 (3.0) Missing 0 (0.0) 1 (1.3) 0 (0.0) 1 (0.4) NOTE. Values expressed as median (interquartile range), mean SD, or n (%). Abbreviations: OP, outpatient; SLP, speech language pathology; w/o, without. *Significantly different (P.05). Other discharge locations for IRF include 1 SNF and 3 board care and assisted living. Other discharge locations for SNF include 1 IRF, 1 board care and assisted living, and 1 other location. discharged to home care, institutional PAC users were more likely not to have a caregiver available in the home. 23,29,31-35 The literature for the impact of comorbid conditions on discharge setting was equivocal. Some investigators 31,32,33 found no influence of comorbid conditions on discharge location. Others found that compared with patients with TJR discharged to home, IRF 29,33 and SNF patients 36 were sicker. It was unclear whether IRF or SNF patients with TJR were sicker because some studies found IRF patients to be sicker, 21 whereas others found SNF patients to be sicker. 17,18 Differences in how severity and comorbidities are defined in each study likely contributed to this variation in findings. We found that SNF and IRF patients with lower-extremity joint replacement were fairly similar in motor functional status at admission and had similar LOSs. Differences in functional status at PAC admission have been harder to determine from the literature because few studies had consistent measures of function across PAC provider types. A few studies 23,24 found Table 4: Unadjusted Functional Status at Discharge and Functional Change From Admission to Discharge Variable SNF (n 83) IRF (n 76) HHA (n 71) Total (n 230) Functional status at discharge Self-care * * * Mobility * Functional change Self-care * * Mobility NOTE. Values expressed as mean SD. ANOVA was used for significant difference testing and differences among settings was tested post hoc by using Tukey HSD test. *Significant (P.05).
7 718 COMPARING POSTACUTE CARE SETTINGS, Mallinson Table 5: Results of Stepwise Regression Analysis on Self-Care Function at Discharge Self-Care Function at Discharge Step l Step 2 Step 3 Step 4 Step 4 95% CI Setting IRF 2.14* 2.16* 1.84* 1.90* 2.84 to 0.95 HHA to 2.14 Demographics Age to 0.03 Sex to 0.16 Severity Procedure 1.75* 1.70* 0.88 to 2.52 Revision to 0.87 No. of comorbid conditions to 0.02 Urinary incontinence 1.80* 1.78* 2.73 to 0.82 Fecal incontinence to 0.21 Self care at admission 0.35* 0.34* 0.21 to 0.47 Therapy LOS to 0.01 R Adjusted R NOTE. Base categories are setting: skilled nursing facilities; sex: woman; diagnosis: hip replacement; revision: no revision; urinary incontinence: continent at admission; fecal incontinence: continent at admission. Abbreviation: CI, confidence interval. *P.05; P.01; P.001. no differences in functional status between patients with TJR discharged home versus PAC, whereas others 29,32 found that patients who had more limited functional status on discharge from acute-care hospitals were more likely to go to a PAC setting. Studies comparing IRF and SNF TJR patients found that IRF patients had lower functional status on admission to PAC. 13,21 Lavernia et al 6 found that HHA patients had higher functional status at admission than SNF or IRF patients. For healthier patients, particularly those with assistance at home, our study found that receiving home health care offered a significant advantage in recovery of functional status. Although home health patients generally were less dependent at admission, after controlling for this factor and other clinical issues, home health patients still achieved greater discharge functional status. It may be that patients homes offer the optimal rehabilitation environment because patients are likely to be more active in their own home, for example, walking to and using the bathroom and getting in and out of chairs or bed, compared with an institutional setting. This finding was consistent with a growing literature that suggested that discharging Variable Table 6: Results of Stepwise Regression Analysis on Mobility Function at Discharge Mobility Function at Discharge Step l Step 2 Step 3 Step 4 Step 4 95% CI Setting IRF to 0.12 HHA 5.09* 4.62* to 4.08 Demographics Age to 0.05 Sex to 1.95 Severity Procedure to 1.72 Revision to 0.97 No. of comorbid to 0.21 conditions Urinary incontinence to 0.67 Fecal incontinence to 3.27 Mobility at admission 0.57* 0.58* 0.39 to 0.77 Services LOS to 0.10 R Adjusted R NOTE. Base categories are setting: skilled nursing facilities; sex: woman; diagnosis: hip replacement; revision: no revision; urinary incontinence: continent at admission; fecal incontinence: continent at admission. Abbreviation: CI, confidence interval. *P.05; P.01; P.001.
8 COMPARING POSTACUTE CARE SETTINGS, Mallinson 719 patients with lower-extremity joint replacement directly to home is feasible and beneficial. Schneider et al 35 examined results of fast-track discharge (from acute care) to home of 100 consecutive patients off a waiting list without exclusion criteria. Seventy-three percent of patients with TKR and 84% of patients with THR achieved the target of discharge 3 to 5 days postsurgery. Fast-track patients did not have a greater need for home care, but had 1 to 2 more postdischarge OT contacts. Mahomed et al 22 conducted a randomized controlled trial of discharge to home or inpatient rehabilitation for 234 patients (mean age, 68y) receiving unilateral total hip or total knee arthroplasty (not related to a fracture). They found no difference between the 2 groups in rates of complications, pain, or physical function measures at 3- or 12-month follow-up. In contrast to other studies, 14,20,21 we found that IRFs did not offer a significant benefit compared with SNFs in terms of recovery of functional status after controlling for patient characteristics. However, IRF patients in this study had significantly more comorbid conditions than either SNF or HHA patients. Part of the decision to admit sicker patients to an IRF may have been a need for 24-hour physician and nursing care that is not available in most SNFs. Unlike DeJong et al, 11 we did not find an advantage from more therapy minutes, suggesting that patients in this study did not benefit from the higher therapy intensity that IRFs must provide. One implication may be that there is a need for higher intensity medical services with less intensive therapy services. This is an option not generally available in PAC settings, but could be addressed with longer acute-care stays. The facilities participating in this study were likely to be of higher than average quality relative to the average facility of each type because of their willingness and ability to participate in the study. DeJong 11 found a variation between and within facilities. In contrast, we found IRF and SNF patients had similar LOSs, which were similar to those for IRF patients in DeJong s 11 study. The number of IRF OT hours per patient in our study was similar to those in DeJong s 11 study, whereas IRF PT hours were higher. SNF OT and PT hours per patient in our study were both on the low end of those reported in DeJong s 11 study. Future research should focus on facilityspecific characteristics as predictors of therapy variations within types of PAC facilities and as predictors of functional outcomes. MedPAC has noted the fragmentation of the health care system and described how current payment incentives, including prospective payment, encourage providers to reduce financial risk by discharging patients to other settings. 37 MedPAC suggested payment bundling for acute and postacute providers to better coordinate care across episodes of care while maintaining quality and outcomes. The American Medical Rehabilitation Providers Association has proposed creation of continuing care hospitals to address fragmentation in PAC and tailor services to patients with medical and rehabilitation needs. 38 The Patient Protection and Affordable Care Act includes pilot studies for bundling and continuing care hospital concepts to examine how these approaches may be operationalized and affect patient outcomes. The focus of our study was on functional status after postacute rehabilitation services because function is a key quality indicator for rehabilitation settings. However, our results suggest that medical outcomes, such as rehospitalization rates, may be another important indicator in evaluating postacute facilities given the clinical complexity of patients. Nurses in the study noted that the medical complexity of IRF patients has increased in recent years; they frequently are admitting patients who require more intensive medical services. They also noted that patients are more frequently treated at bedside because they are too sick to attend the therapy gym or activities of daily living area, and meeting daily therapeutic activity goals has become increasingly challenging. Study Limitations Limitations of this study included the small sample size. Although sufficient to power the analyses, these data were from only 2 census areas and were not representative of the entire United States. In particular, like other studies comparing rehabilitation outcomes, we recruited high-quality providers. 11 Although we collected data for the presence of comorbid conditions, we did not collect data for the severity of comorbid conditions, for example, body mass index, hemoglobin levels for anemic patients, extent of depressive symptoms, and degree of diabetes control. In addition, the methods of collecting data, although similar across settings, may have influenced which comorbid conditions were reported. A limit of 10 comorbid conditions may have created a ceiling. We attempted to statistically control for differences in patient characteristics across settings. To the extent that we may not have measured all relevant indicators, we may have over- or underestimated the effect of setting on patient outcomes. CONCLUSIONS For the patients in our U.S.-based study, direct discharge to home with home care was the optimal strategy for patients after TJR surgery who were healthy and had social support. For sicker patients, availability of 24-hour medical and nursing care may be needed, but intensive therapy services did not seem to lead to additional improvement of functional recovery in these patients. Acknowledgments: We thank the facilities, research nurses, administrators, and coordinators from data collection sites who gave so generously of their time; Nancy Richman, OTR/L, FAOTA, Dr Pamela Smith, DNS, RN, and Linda Murakami, RN who provided training on the PAC instruments; and Dr Bill Shuart, PhD, Pat Wilson, RN, Beth Souder, PT, Dr Kathleen Ruroede, PhD, MEd, RN, Dr Santiago Toledo, MD, and Peggy Kirk, RN for comments and feedback on earlier versions of this manuscript. References 1. Tian W, DeJong G, Brown M, Hsieh CH, Zamfirov ZP, Horn SD. Looking upstream: factors shaping the demand for postacute joint replacement rehabilitation. Arch Phys Med Rehabil 2009;90: Kurtz S, Mowat F, Ong K, Chan N, Lau E, Halpern M. Prevalence of primary and revision total hip and knee arthroplasty in the United States from 1990 through J Bone Joint Surg Am 2005;87: Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to J Bone Joint Surg 2007;89: Wilson NA, Schneller ES, Montgomery K, Bozic KJ. Hip and knee implants: current trends and policy considerations. Health Aff (Millwood) 2008;27: MedPAC. Report to the Congress: issues in a modernized Medicare Program. Washington (DC): MedPAC; 2005 June 15, Lavernia CJ, D Apuzzo MR, Hernandez VH, Lee DJ, Rossi MD. Postdischarge costs in arthroplasty surgery. J Arthroplasty 2006; 21(6 Suppl 2): Bozic KJ, Rubash HE, Sculco TP, Berry DJ. An analysis of Medicare payment policy for total joint arthroplasty. J Arthroplasty 2008;23(6 Suppl 1): Buntin MB, Colla CH, Escarce JJ. Effects of payment changes on trends in post-acute care. Health Serv Res 2009;44:
9 720 COMPARING POSTACUTE CARE SETTINGS, Mallinson 9. Munin MC, Seligman K, Dew MA, et al. Effect of rehabilitation site on functional recovery after hip fracture. Arch Phys Med Rehabil 2005;86: Wodchis WP. Physical rehabilitation following Medicare prospective payment for skilled nursing facilities. Health Serv Res 2004; 39: DeJong G, Hsieh CH, Gassaway J, et al. Characterizing rehabilitation services for patients with knee and hip replacement in skilled nursing facilities and inpatient rehabilitation facilities. Arch Phys Med Rehabil 2009;90: FitzGerald JD, Boscardin WJ, Ettner SL. Changes in regional variation of Medicare home health care utilization and service mix for patients undergoing major orthopedic procedures in response to changes in reimbursement policy. Health Serv Res 2009;44: Buntin MB, Partha D, Escarce JJ, Hoverman C, Paddock S, Sood N. Comparison of Medicare spending and outcomes of beneficiaries with lower extremity joint replacements. Arlington: RAND Health; June 2005 Report No DeJong G, Tian W, Smout RJ, et al. Use of rehabilitation and other health care services by patients with joint replacement after discharge from skilled nursing and inpatient rehabilitation facilities. Arch Phys Med Rehabil 2009;90: Liu K, Wissoker D, Rimes C. Determinants and costs of Medicare post-acute care provided by SNFs and HHAs. Inquiry 1998;35: Gage B. Impact of the BBA on post-acute utilization. Health Care Financ Rev 1999;20: Gage B, Morley M, Spain P, Ingber M. Examining post acute care relationships in an integrated hospital system. Washington (DC); Centers for Medicare and Medicaid Services, Department of Health and Human Services; p Buntin MB, Garten AD, Paddock S, Saliba D, Totten M, Escarce JJ. How much is postacute care use affected by its availability? Health Serv Res 2005;40: Bozic KJ, Wagie A, Naessens JM, Berry DJ, Rubash HE. Predictors of discharge to an inpatient extended care facility after total hip or knee arthroplasty. J Arthroplasty 2006;21(6 Suppl 2): Walsh MB, Herbold J. Outcome after rehabilitation for total joint replacement at IRF and SNF: a case-controlled comparison. Am J Phys Med Rehabil 2006;85: DeJong G, Horn SD, Smout RJ, Tian W, Putman K, Gassaway J. Joint replacement rehabilitation outcomes on discharge from skilled nursing facilities and inpatient rehabilitation facilities. Arch Phys Med Rehabil 2009;90: Mahomed NN, Davis AM, Hawker G, et al. Inpatient compared with home-based rehabilitation following primary unilateral total hip or knee replacement: a randomized controlled trial. J Bone Joint Surg Am 2008;90: Tribe KL, Lapsley HM, Cross MJ, Courtenay BG, Brooks PM, March LM. Selection of patients for inpatient rehabilitation or direct home discharge following total joint replacement surgery: a comparison of health status and out-of-pocket expenditure of patients undergoing hip and knee arthroplasty for osteoarthritis. Chronic Illness 2005;1: Mahomed NN, Koo Seen Lin MJ, Levesque J, Lan S, Bogoch ER. Determinants and outcomes of inpatient versus home based rehabilitation following elective hip and knee replacement. J Rheumatol 2000;27: Lai JVCA, Linacre, JM. Adjusting for rater severity in an unlinked functional independence measure national database: an application of the many facets Rasch model. Physical medicine & rehabilitation: state of the art reviews. Philadelphia: Hanley & Belfus Inc; p Marais I. Response dependence and the measurement of change. J Appl Meas 2009;10: Wright BD. Rack and stack: time 1 vs. time 2. Rasch Measure Trans 2003;17: Bond T, Fox C. Applying the Rasch model: fundamental measurement in the human science. Mahwah: Lea; Munin MC, Kwoh CK, Glynn N, Crossett L, Rubash HE. Predicting discharge outcome after elective hip and knee arthroplasty. Am J Phys Med Rehabil 1995;74: Forrest G, Fuchs M, Gutierrez A, Girardy J. Factors affecting length of stay and need for rehabilitation after hip and knee arthroplasty. J Arthroplasty 1998;13: Oldmeadow LB, McBurney H, Robertson VJ. Predicting risk of extended inpatient rehabilitation after hip or knee arthroplasty. J Arthroplasty 2003;18: de Pablo P, Losina E, Phillips CB, et al. Determinants of discharge destination following elective total hip replacement. Arthritis Rheum 2004;51: Epps CD. Length stay, discharge disposition, and hospital charge predictors. AORN J 2004;79:975-6, , Chimenti CE, Ingersoll G. Comparison of home health care physical therapy outcomes following total knee replacement with and without subacute rehabilitation. J Geriatr Phys Ther 2007;30: Schneider M, Kawahara I, Ballantyne G, et al. Predictive factors influencing fast track rehabilitation following primary total hip and knee arthroplasty. Arch Orthop Trauma Surg 2009;129: Bini SA, Fithian DC, Paxton LW, Khatod MX, Inacio MC, Namba RS. Does discharge disposition after primary total joint arthroplasty affect readmission rates? J Arthroplasty 2010;25: MedPAC: Medicare Payment Advisory Committee. Report to the Congress: Medicare Payment Policy. Washington (DC): Med- PAC; March 1, American Medical Rehabilitation Providers Association An option for the future of medical rehabilitation and other post acute care hospital providers: the continuing care hospital. Concept paper. Washington (DC): AMRPA; May 4, Suppliers a. Winsteps, Available at: b. StataCorp LP, 4905 Lakeway Dr, College Station, TX