Progressive Multicomponent Intervention for Older Adults in Home Health Settings Following Hospitalization: Randomized Clinical Trial

Abstract

Objectives

Reduced physical function following hospitalization places older adults at risk of adverse health events. Many older adults receive home health physical therapy to reverse their deconditioning; however, optimal approaches to improve physical function are currently not known. This study aimed to evaluate the effectiveness of a home health care approach comprised of high-intensity exercise, enhanced care transition, and protein supplementation.

Methods

Eligible participants included adults aged 65 years or older referred to home health care following hospitalization. Two hundred older adults who are medically complex were enrolled and were randomized 1:1 to (1) a high-intensity progressive, multi-component (PMC) intervention or (2) enhanced usual care (UC) comparison group. All participants received 12 visits over 60 days. The primary study outcome was change in the Short Physical Performance Battery (SPPB) from baseline to 60 days. Secondary outcomes included gait speed (usual, fast), modified Physical Performance Test, grip strength, Fatigue Severity Scale, Falls Efficacy Scale-International, physical activity (step count), and adverse events (falls, emergency department visits, hospitalizations). All outcomes were collected at baseline, then 30, 60, 90, and 180 days post baseline.

Results

There was no difference in 60-day SPPB change between groups with both groups experiencing significant improvements (PMC = 1.53 [95% CI: 1.00–2.05]; enhanced UC = 1.39 [95% CI = 0.89–1.88]). Differences were also not observed in secondary measures or adverse events at any time point.

Conclusion

An intervention consisting of high-intensity exercise, enhanced care transition, and protein supplementation was not associated with greater functional improvement at 60 days compared to enhanced UC in older adults receiving home health physical therapy.

Impact

The findings of this study demonstrate that a high-intensity progressive, multi-component intervention results in similar physical functional changes as an enhanced UC intervention in older adults who are medically complex and receiving home health care following hospital-associated deconditioning.

Exercise, Geriatrics, Health, High Intensity Exercise, Home Care Services, Intervention

INTRODUCTION

Hospital-associated deconditioning, a sharp downward trajectory in physical function following acute hospitalization, is a growing concern for older adults.^1–3 Older adults discharged home after hospitalization are 60 times more likely to develop limitations in activities of daily living (ADLs) compared to older adults who are not hospitalized.¹ Similarly, 68% of older adults who require institutional post-acute care (skilled nursing facility [SNF] or inpatient rehabilitation) after hospitalization are discharged home below their pre-hospitalization level of function.⁴ Declines in physical function have been strongly linked to the increased rates of rehospitalization, disability, and mortality.^5,6 More than 3 million older adults receive access to home health physical therapy following hospitalization,^7,8 to attenuate the impact of deconditioning and improve physical function. However, many factors may influence the degree of benefit received from participation in home health physical therapy including the physiologic capacity for patients to improve their physical function and the type of intervention delivered.

Prior research has demonstrated that high-intensity resistance training is a safe and effective intervention for promoting physical function in older adults who are frail.^9–11 Specifically, a Cochrane review detailed that high-intensity resistance training in older adults improves muscle strength, gait speed, and pain.¹² There is a need to test high-intensity principles within home health physical therapy given unique aspects of the home care setting (eg, homebound patients with recent illness, environmental constraints, increased treatment focus on ADLs and safety, and frequency of treatment). Additionally, there are several factors beyond high-intensity resistance training that may influence the effectiveness of home health rehabilitation and its ability to influence health outcomes. For instance, greater integration of home health therapists in the holistic care transition from hospital to home has potential to improve health outcomes in older adults above and beyond direct delivery of physical therapy.¹³ Enhancing care transition requires effective information transfer from the hospital to the home health rehabilitation team (eg, medical concerns, plan of care) to reduce the risk of rehospitalization¹⁴ while optimizing functional capacity.^15,16 Nutrition also affects the functional capacity of older adults participating in high-intensity exercise. Older adults with chronic medical conditions have elevated dietary protein requirements and additional protein intake may facilitate improvements in muscle mass from high-intensity exercise.¹⁷ Despite the presence of these likely influential factors, the efficacy of an intervention combining high-intensity exercise, enhanced care transition, and protein supplementation has not been evaluated.

Thus, the purpose of this study was to assess whether a high-intensity progressive, multi-component (PMC) intervention, initiated upon admission to home health following discharge from a hospital or SNF over a 60-day episode of care, improves physical function more than enhanced usual care (UC) physical therapy. We hypothesized that the PMC group would experience greater improvements in physical function 60 days following discharge from a hospital or SNF compared to the enhanced UC group.

METHODS

Study Design

This was a single blinded (assessors), randomized controlled trial in older adults referred to home health physical therapists following discharge home from a hospital or SNF (following a hospitalization). One study group received a progressive multicomponent program including high-intensity home health physical therapy, enhanced care transition, and nutritional protein supplements, while the second group received enhanced UC home health physical therapy (Figure 1).

Figure 1

Abbreviated Study Design. PMC = progressive multicomponent care; UC = usual care.

The trial was registered at clinicaltrials.gov (NCT02905370). The detailed study protocol has been published previously. Please refer to the protocol paper for a full description of the intervention groups and outcome measures.¹⁸

Participants

Eligible participants included adults 65 years of age and older referred to home health physical therapy with acute medical deconditioning from illness requiring hospitalization, defined by therapy being ordered on discharge from hospital or from SNF following a hospitalization. Participants also had to have at least three comorbid conditions. Exclusion criteria included the inability to ambulate 10 feet without human assistance, with or without an assistive device, at time of discharge with a gait speed between 0.3 to 1.0 m/s and the presence of factors precluding full participation in high-intensity resistance training.¹⁹ These included acute lower extremity fracture with weight-bearing restriction, acute joint replacement surgery, lower extremity amputation, terminal illness, cancer, pulmonary embolus (<6 weeks), acute stroke, progressive neurogenerative diseases, use of illegal substances, dementia, lack of cognitive capacity to follow directions, and lack of capacity to provide informed consent. If indicated in their course of care, capacity was determined via a formal assessment by hospital medical staff. If a formal capacity evaluation was not documented, capacity was assessed by research staff through targeted questioning related to comprehension of the intervention and study involvement and subsequently reassessed by an in-home assessor during baseline testing. A cut-off of <20 on the St. Louis University Mental Status exam was initially used to determine exclusion due to dementia; however, this exclusion criteria was removed in August of 2018 due to the study team’s recognition of the prevalence of impaired cognition, and subsequent high exclusion rates in the home health population. Instead, study team members subjectively evaluated cognition with an emphasis on ensuring capacity to safely participate in high intensity exercise. Pre-consent screening was conducted by trained research assistants via chart review with consultations from physicians, rehabilitation practitioners, and case managers to ensure appropriateness. The number and reasons potential participants were excluded (before and after consent) are provided in Figure 2 and Supplementary Tables 1 and 2.

Figure 2

CONSORT Flow Diagram.

Recruitment and Randomization

Eligible participants (n = 200) were recruited during an acute care hospitalization at UCHealth (University of Colorado Health) or by referral from a partnering hospital or SNF (58 potential recruitment sites). Following informed consent and within 72 h of discharge home, study staff conducted the assessment within the participant’s home. Upon completion of the assessment, the study coordinator opened a pre-packed envelope with the randomization assignment and conveyed the information to the home health care agency coordinator who assigned the treating therapist. Prior to patient approach for consent and again prior to randomization, the study coordinator confirmed with the home health care agency coordinator that therapists for both randomization assignments were available to treat the patient. Randomization was performed using randomly permuted block sizes of 4 and 6, stratified by discharge location (SNF or hospital). Randomization schemes were produced using the PLAN procedure in SAS v9.4 (SAS Institute, Inc; Cary, NC, United States) with randomly generated seeds. Once the concealment process was complete, only the trial’s biostatistician accessed the schemes.

Interventions

Physical therapists from two home health agencies in Metro Denver provided interventions which were initiated within 5 days after discharge home as part of routine home health care. After an evaluation, all participants received 12 home health physical therapist visits over the 60-day intervention period, comprising eight visits during the first 30 days (two visits per week) followed by four visits over the last 30 days (one visit per week). Physical therapist visits for all participants consisted of three primary components (strengthening exercises, ADL training, and gait training); however, the intensity of these components differed between intervention groups. Physical therapists were trained by members of the study team to deliver either PMC or enhanced UC interventions (12 therapists per intervention) and were blinded to the alternative. Two study staff physical therapists were trained to provide both interventions for gap coverage to ensure continuity of care. Participants received agency standards of home health care including nursing, occupational therapy, and social work visits as needed.

Progressive Multicomponent Intervention

Participants in the PMC intervention group performed high-intensity strengthening exercises targeting their lower extremities. Strengthening exercises included a single leg press in supine and unilateral standing hip extension with resistance applied via a portable progressive resistance exercise machine employing latex bands that provided 6 to 100 pounds of resistance (Shuttle MiniPress, Contemporary Design Company, Glacier, WA, United States). The amount of resistance was evaluated to be at an intensity of an eight-repetition maximum and participants performed the exercises at a volume of three sets of eight repetitions for each leg. Intensity was reevaluated weekly, and resistance was increased if the participant was able to complete eight repetitions without failure due to muscle fatigue. This dosage and intensity were chosen as they align with the American College of Sports Medicine guidelines for improving muscle strength.¹⁹ Further, significant functional improvements have occurred in previous intervention studies employing this dosage and intensity with similar populations.¹¹ The ADL training component involved functional exercises necessary for independence in the home setting including bed mobility, sit to stand, and floor to stand. Progression of ADL training difficulty occurred by task or environmental adjustment when a participant successfully completed eight repetitions with functional tasks progressed per an algorithm developed by the study team.¹⁸ The final exercise component was a gait training program that aimed to improve gait speed and quality. This program included stepping and walking activities that progressed in complexity, speed, and accuracy once completed with an 80% success rate to ensure there was adequate challenge. Each patient in the PMC intervention group also received a home exercise program (HEP), to be completed up to twice daily, that included the functional ADL exercises of chair rises and step-ups as well as recommending a 20-min walking program.¹⁸

Along with the 12 physical therapist visits and home exercise program, all patients in the PMC intervention group received assistance with their care transition home via a Care Transitions Coordinator, a study-staffed registered nurse or physical therapist. The Care Transitions Coordinator (1) communicated with hospital and home health staff to ensure accurate transfer of patient information using a structured transition communication tool, (2) established a patient-centered, interdisciplinary plan of care, and (3) coordinated with study team physicians and treating therapists to clarify discrepancies or questions identified in the transition communication log or new medical concerns identified by the therapist. If a participant responded to Care Transitions Coordinator, a 1-h visit discussing care transition would occur. Should follow-up be needed, a subsequent visit between the participant and Care Transitions Coordinator would occur. This resulted in 1 to 3 h spent between a participant and the Care Transitions Coordinator.

Lastly, participants within the PMC intervention group received a high-protein (15 g) daily nutritional supplement (Boost High Protein or Boost Glucose Control). The 15 gram supplement was provided to support muscle strength gains as research indicates that older adults with chronic medical conditions may require elevated dietary protein.¹⁷ Therapists encouraged participants to consume protein shakes at the end of each therapy session and on non-therapy days. Participants were requested to log their daily protein supplement intake and exercises performed within their HEP to track adherence.

Enhanced Usual Care

Participants in the enhanced UC group received equal frequency, duration, and domain of intervention for their physical therapist visits. Strengthening therapeutic exercises included active range of motion or unweighted exercise, for one set of 10 repetitions each. Participants performed four exercises at the start of the program, adding up to one additional exercise each week up to a maximum of seven therapeutic exercises. The ADL training component included practicing transfers with an emphasis on safety, such as using the upper extremities to assist in rising and sitting. Gait training also emphasized safety and was tailored to each patient and could include ambulating inside/outside of the home with/without an assistive device. Participants within the enhanced UC group also received an HEP containing two exercises from their physical therapist visits with the goal of completing 10 repetitions daily. Participants in the enhanced UC group did not have any contact with the Care Transitions Coordinator nor did they receive any nutritional advice or supplementation.

This intervention group was termed “enhanced” UC due to the higher number of home health physical therapist visits received in comparison to real world home health care as well as the standardized delivery and progression of physical therapist exercises. With the adoption of the Patient Driven Groupings Model, a new home health payment model, the study team noted greater differences in visit numbers when comparing the enhanced UC intervention to real world care and consequently added in a third, non-randomized, group while the study was ongoing in August 2018—true UC. The true UC group consisted of participants who received home health care without any input from the study team and was therefore reflective of real-world home health care. Analyses examining how true UC group outcomes compared to the PMC and enhanced UC groups will be examined in a subsequent manuscript.

Outcomes

Outcomes assessments were performed by clinically trained (physical therapist or physical therapist assistant) research assistants within participants’ homes. Reliability testing was performed every 6 months between active outcome assessors (up to four active assessors at any time). The primary study outcome was the Short Physical Performance Battery (SPPB), a well-studied global measure of lower extremity physical function which consists of walking speed, chair stands, and balance. The SPPB is a composite measure of lower extremity function and a strong predictor of disability, institutionalization, and morbidity in older adults.^20,21 The continuous scoring system minimizes ceiling effects, and scores range from 0 to 12 with higher scores indicative of better performance. The SPPB was collected at baseline (within 72 h of hospital/SNF discharge), then 30, 60, 90, and 180 days following baseline. The primary end point, 60 days post discharge, was chosen for clinical relevance as it parallels the end of a Medicare-defined episode of care.

Baseline sociodemographic and descriptive measures included age, sex, race, ethnicity, education level, marital status, height, weight, discharge location, functional comorbidity index (FCI), and the Patient Activation Measure (PAM-10). The FCI assesses the presence/absence of comorbidities (out of a possible total of 18)²² while the PAM-10 probes an individual’s self-reported knowledge, skill, and confidence for managing their own health.^23,24 All sociodemographic and descriptive measures were collected via patient self-report at baseline by blinded outcomes assessors or via electronic medical record.

Intervention adherence was assessed (sessions attended/total number of sessions prescribed) with daily treatment logs, collected by the treating therapist, using previously established methods for monitoring therapist adherence.²⁵

The following secondary outcomes were collected at baseline, 30, 60, 90, and 180 days post hospital discharge: fast 4-meter walking speed (4 MW fast),²⁶ upper and lower extremity function via the modified Physical Performance Test (mPPT),²⁷ grip strength,²⁸ impact of fatigue on activities of daily living (ADL) via the Fatigue Severity Scale (FSS),²⁹ fear of falling via the Falls Efficacy Scale-International (FES-I),³⁰ and adverse events. Physical activity data were collected at baseline, 60, and 180 days post discharge via a thigh-mounted activPAL (PAL Technologies, Glasgow, Scotland) worn for 24 h per day over a minimum of 7 days.^31–33 Adverse events assessed included occurrences of falls, emergency department visits, hospitalizations, and deaths throughout the 180-day study period. Adverse events were determined via physical therapist queries at each visit and 120- and 150-day phone calls, supplemented by study-staff chart reviews. Patients were also given a fall diary to record any falls to facilitate documentation. Falls were defined as an unintentional change in position resulting in coming to rest on the ground or other lower level.

Power and Sample Size

We estimated statistical power from a pilot study of 18 participants (eight randomized to PMC, 10 to the enhanced UC. The 60-day change in SPPB from baseline was 2.6 (3.0) (mean [SD]) and 0.5 (1.4) in the PMC and enhanced UC groups, respectively. Assuming a 2.1 pooled SD for both groups, a sample size of 150 patients (75 per group) provided 82% power to detect a between group difference of 2 points in mean SPPB score (using a 2-tailed, alpha = 0.05 level, 2-group t test), which is double the minimal clinically important difference.³⁴ Based on an estimated dropout rate of 25%, a total of 200 participants (100 per group) were recruited.

Statistical Analyses

Participant demographics and baseline physical function were examined using counts with percentages or means with standard deviations, depending on variable type. The primary analysis compared the differences between treatment groups in amount of change in SPPB scores 60 days following baseline assessment. This was accomplished via a linear regression model with a response variable of SPPB 60-day change and explanatory variables of treatment group (PMC vs. enhanced UC), the stratification variable of discharge facility (SNF vs. hospital), and baseline SPPB. In this model, the coefficient for treatment group was the statistical measure for the difference between groups while controlling for facility and baseline SPPB scores. This analysis was repeated for other periods of change (30, 90, 180 days) and secondary variables (4 MW fast, mPPT, grip strength, FSS, FES-I, and step counts). The median cumulative number and occurrence (yes/no) of each adverse event (falls, emergency department visits, hospitalizations) were calculated and compared at each timepoint via Wilcoxon rank-sum and chi-square tests, respectively. A Wilcoxon rank-sum test was also used to compare median cumulative deaths at 180 days. Analyses were performed using SAS v9.4 and assumed a two-tailed hypothesis test and significance level of 0.05.

Role of Funding Source

The funding sources had no role in data collection or analysis.

RESULTS

Between the dates of October 1, 2016, and December 4, 2021, we enrolled and randomized 200 older adults to the PMC intervention (n = 100) and the enhanced UC comparison group (n = 100) (Figure 2). The study groups’ sociodemographic measures are presented in Table 1. Participants were primarily female (58%) with an average age of 78.1 and an average of five function-limiting chronic conditions (FCI; PMC = 5.02 [SD 2.2], enhanced UC = 4.87 [SD 2.3]). Participants were primarily recruited following discharge directly from the hospital (PMC = 73%, enhanced UC = 74%). Outcomes at baseline are reported in Table 2.

Table 1

Demographics^a

Characteristic	Overall (N = 200)	PMC (N = 100)	EUC (N = 100)
	Mean (SD) or N (%)	Mean (SD) or N (%)	Mean (SD) or N (%)
Discharged from
Hospital	147 (73.5%)	73 (73%)	74 (74%)
SNF	48 (24%)	26 (26%)	22 (22%)
Acute rehab	5 (2.5%)	1 (1%)	4 (4%)
Age (y)	78.1 (7.7) 78 (65–98)	78.7 (7.9) 79 (65–95)	77.6 (7.6) 77 (65–98)
Sex
Male	84 (42%)	42 (42%)	42 (42%)
Female	116 (58%)	58 (58%)	58 (58%)
Race
White	154 (77%)	73 (73%)	81 (81%)
Black	25 (12.5%)	16 (16%)	9 (9%)
Native American	6 (3%)	2 (2%)	4 (4%)
Asian	4 (2%)	1 (1%)	3 (3%)
Multiracial/Other	11 (5.5%)	8 (8%)	3 (3%)
Hispanic
Yes	20 (10%)	11 (11%)	9 (9%)
No	180 (90%)	89 (89%)	91 (91%)
Education (N = 196)
≤ 12 y, no diploma	16 (8%)	8 (8%)	8 (8%)
High school diploma	42 (21%)	20 (21%)	22 22%)
Some college	59 (30%)	32 (33%)	27 (27%)
Associate/bachelor’s degree	39 (20%)	14 (14%)	25 (25%)
Master’s/doctoral degree	40 (20%)	23 (24%)	17 (17%)
Marital status
Single	24 (12%)	9 (9%)	15 (15%)
Married/cohabitating	76 (38%)	34 (34%)	42 (42%)
Divorced/separated	31 (15.5%)	19 (19%)	12 (12%)
Widowed	69 (34.5%)	38 (38%)	31 (31%)
Height (cm)	166.0 (10.7)	165.8 (10.2)	166.2 (11.2)
Weight (kg)	77.8 (22.6)	76.4 (19.8)	79.1 (25.2)
Functional Comorbidity Indexb	4.95 (2.2)	5.02 (2.2)	4.87 (2.3)

^a

EUC = enhanced usual care; PMC = progressive multicomponent care; SNF = skilled nursing facility.

^b

Functional comorbidity index (FCI) scores range from 0–18, with higher scores representing a greater number of comorbidities.

Table 2

Outcomes at Baseline^a

Outcome	PMC		EUC
	N	Mean (SD)	N	Mean (SD)
Short Physical Performance Batteryb	98	5.05 (2.2)	98	5.39 (2.2)
Gait speed (m/s)	99	0.45 (0.17)	98	0.47 (0.16)
Gait speed fast (m/s)	98	0.64 (0.23)	94	0.71 (0.23)
Modified Physical Performance Testc	95	12.4 (6.1)	94	13.0 (5.8)
Grip strength (kg)	92	19.7 (7.5)	91	21.0 (7.5)
Fatigue Severity Scale (sum)d	100	42.2 (13.9)	96	43.1 (12.4)
Falls Efficacy Scale-Internationale	97	36.9 (11.9)	93	36.3 (12.6)
Patient Activation Measure-10 Scoref	99	62.2 (14.8)	94	63.0 (14.5)
Patient Activation Measure-10 Level	99	2.80 (0.91)	94	2.84 (0.91)
St Louis University Mental Statusg	99	24.0 (4.0)	98	24.5 (3.7)

^a

EUC = enhanced usual care; PMC = progressive multicomponent care.

^b

Short Physical Performance Battery scores range from 0 to 12. Higher scores are indicative of better performance.

^c

Modified Physical Performance Test task scores range from 0 (unable to complete) to 4 (performed quickly and easily) with a maximum test score of 28.

^d

Fatigue Severity Scale scores range from 9 (no impact) to 64 (profound impact of fatigue on activities of daily living).

^e

Falls Efficacy Scale-International scores range from 16 (not at all concerned) to 64 (very concerned about falling).

^f

Patient Activation Measure scores range from 0 to 100, with higher scores representing higher self-management and understanding of health conditions. Levels are the corresponding sub-categories from 1 (low activation/motivation) to 4 (high activation/motivation).

^g

St Louis University Mental Status examination scores range from 0 to 30. Cut-off scores include 27–30 considered normal in a person with a high school education), 21–26 (suggest a mild neurocognitive disorder), and 0–20 (indicate dementia).

Participants in the PMC group attended 10.1 out of 12 physical therapist visits (SD = 3.6). PMC therapy notes indicated that therapeutic strengthening exercise, ADL training, and gait training occurred during 95.4%, 90.1%, and 83.8% of completed sessions, respectively. Within the comparison, enhanced UC group participants attended 10.2 of out 12 visits (SD = 3.3). Therapeutic strengthening exercise, ADL training, and gait training occurred during 97.3%, 67.4%, and 94.8% of enhanced UC visits, respectively. Fidelity data on the HEP and protein supplementation was not analyzed secondary to infrequent adherence log completion.

Both the PMC and enhanced UC groups experienced 60-day SPPB improvement. However, there was no significant difference in the primary endpoint for SPPB change between groups while controlling for discharge location and baseline SPPB (PMC = 1.53 [95% CI =1.00–2.05); enhanced UC = 1.39 [95% CI = 0.89–1.88]; P = 0.68) (Table 3). Similarly, there were no differences in the secondary measures and time points (Table 3; Suppl. Tables 3–5), excluding 90-day mPPT change which was greater in the enhanced UC group (PMC = 2.45 [95% CI = 1.04–3.85]; enhanced UC = 4.84 [95% CI = 3.46–6.23]; P = 0.01).

Table 3

60-Day Model Estimates^a

Model Response Variable	PMC		EUC		Estimated Difference in Change (95% CI)	P
	N	LS Mean Change (95% CI)	N	LS Mean Change (95% CI)
Short Physical Performance Batteryb	77	1.53 (1.00 to 2.05)	83	1.39 (0.89 to 1.88)	−0.14 (−0.82 to 0.54)	.68
Gait speed (m/s)	78	0.13 (0.09 to 0.17)	83	0.14 (0.11 to 0.18)	0.01 (−0.04 to 0.06)	.63
Gait speed fast (m/s)	76	0.20 (0.13 to 0.26)	79	0.17 (0.11 to 0.23)	−0.03 (−0.11 to 0.05)	.47
Modified Physical Performance Testc	72	3.39 (2.12 to 4.66)	75	3.82 (2.58 to 5.05)	0.42 (−1.23 to 2.08)	.61
Grip strength (kg)	73	0.95 (0.06 to 1.84)	76	1.23 (0.37 to 2.10)	0.28 (−0.89 to 1.45)	.64
Fatigue Severity Scale (sum)d	80	−4.11 (−7.06 to −1.15)	83	−3.97 (−6.86 to −1.07)	0.14 (−3.76 to 4.04)	.94
Falls Efficacy Scale-Internationale	74	−2.60 (−5.13 to −0.07)	79	−4.47 (−6.94 to −1.99)	−1.87 (−5.23 to 1.50)	.27
Step count	36	758.74 (−861.62 to 2379.09)	45	443.69 (−465.98 to 1353.36)	314.19 (−46.78 to 675.16)	.39

^a

EUC = enhanced usual care; LS = least squares; PMC = progressive multicomponent care.

^b

Short Physical Performance Battery scores range from 0 to 12. Higher scores are indicative of better performance.

^c

Modified Physical Performance Test task scores range from 0 (unable to complete) to 4 (performed quickly and easily) with a maximum test score of 28.

^d

Fatigue Severity Scale scores range from 9 (no impact) to 64 (profound impact of fatigue on activities of daily living).

^e

Falls Efficacy Scale-International scores range from 16 (not at all concerned) to 64 (very concerned about falling).

No significant differences were observed between study groups in any adverse events (falls, emergency department visits, hospitalizations, deaths) throughout the study period (Table 4).

Table 4

Adverse Events^a

Through Timepoint	Event	EUC (N = 100)		PMC (N = 100)		Kruskal Wallis Chi Square P (the Test for Any [N])	Wilcoxon P (the Test for Number Median [Range])
		Any (N)	Number Median (Range)	Any (N)	Number Median (Range)
0–30 d	Falls	12	0 (0–3)	12	0 (0–7)	>.99	.96
	ED visits	12	0 (0–3)	15	0 (0–3)	.67	.55
	Hospitalizations	20	0 (0–2)	14	0 (0–2)	.35	.31
	Death	2	-	0	-	.52	-
0–60 d	Falls	20	0 (0–4)	21	0 (0–13)	>.99	.70
	ED visits	18	0 (0–4)	20	0 (0–3)	.86	.71
	Hospitalizations	27	0 (0–3)	25	0 (0–3)	.84	.83
	Death	3	-	0	-	.24	-
0–90 d	Falls	28	0 (0–7)	26	0 (0–13)	.88	.95
	ED visits	23	0 (0–4)	27	0 (0–3)	.62	.53
	Hospitalizations	30	0 (0–3)	30	0 (0–4)	>.99	.94
	Death	5	-	1	-	.21	-
0–180 d	Falls	41	0 (0–7)	34	0 (0–13)	.38	.47
	ED visits	36	0 (0–4)	30	0 (0–4)	.46	.39
	Hospitalizations	43	0 (0–3)	35	0 (0–4)	.32	.44
	Deaths	8	-	4	-	.37	-

^a

All patients are included in all denominators as the total counts are cumulative. ED = emergency department; EUC = enhanced usual care; PMC = progressive multicomponent care.

DISCUSSION

Older adults are at substantial risk for functional decline and poor health outcomes following an acute hospitalization,^1–3 highlighting the need to implement safe and effective physical therapist approaches in post-acute settings, particularly in home health. We designed the PMC intervention to improve physical function and health outcomes in older adults recently discharged from the hospital using high-intensity principles during strength, gait, and functional ADL exercises. Moreover, the PMC intervention included an enhanced care transition protocol and protein supplementation to ensure there was capacity for functional improvements.

Previous research has demonstrated that each component of the PMC intervention (high-intensity exercise,^9,11,12 enhanced care transition,^14–16 and protein supplementation)¹⁷ improves outcomes in older adults who are frail. The findings of this study add to this body of research by illustrating that an intervention combining all three components is safe for the general population of older adults who are medically complex and receiving home health care after discharge from a hospital or SNF. While the feasibility and safety of this intervention was supported, findings did not support our hypothesis that the PMC intervention would lead to greater physical function improvements compared to enhanced UC delivered over a target of 12 visits during a 60-day period. Participants in both intervention groups experienced similar increases in SPPB scores, and these improvements exceeded the established minimally clinically important difference (MCID) for SPPB of 1.0 point.³⁴ Moreover, the MCID-exceeding increase persisted at 180 days.

Though the lack of differences between intervention groups was contrary to initial expectations, there are several reasons which might explain this result. For example, the comparison group received standardized care, which included 12 physical therapist visits over a 2-month period. The number and rate of physical therapist visits received by patients in both intervention groups is considerably higher than what occurs in real world care for home health patients as data has shown that between 2015 to 2020 the average home health patient received three physical therapist visits every 30 days.³⁵ This elevated number and frequency of visits for the enhanced UC group, which equaled that of PMC group, may have contributed to the clinically significant changes that did not differ from the PMC group.

Second, both treatment protocols were delivered by clinicians who received standardized training from the research staff with high fidelity. In contrast, dosing and progression is often variable in rehabilitation settings. Thus, it is possible that protocols focused on standardized dosing and progression result in similar physical function gains by reducing variability, regardless of approach. In addition, training on both interventions emphasized motivation and positive reinforcement throughout the exercise sessions; the resulting therapeutic alliance may have contributed to improved physical function in the enhanced UC group. Two aspects that differed between intervention groups (HEP intensity and protein supplementation) were tracked but not analyzed due to inconsistent participant logs, despite consistent reminders by physical therapists and study staff. Consequently, we cannot rule out that the effect of the PMC intervention was attenuated secondary to inconsistent adherence of the high-intensity HEP and protein supplementation.

Lastly, the incorporation of progressive physical challenges outside of therapy sessions may be necessary to optimize physical function gains seen with each intervention. Individual patient factors such as reduced cognitive capacity, limited caregiver support, increased pain, or lower motivation could have potentially negatively impacted exercise adherence and general physical activity outside of physical therapist sessions. However, further research is needed to determine whether these factors influence physical activity. In the current study sample, there was increased potential for social isolation due to the COVID-19 pandemic that may have contributed to reduced overall physical activity and potential for improvements in physical function. Specifically, 40 participants in the PMC group and 39 participants in the enhanced UC group were recruited during pandemic-effected months (after March 2020). However, it is unlikely the lack of differences in study outcomes between the PMC and enhanced UC groups is due to this potential reduced physical activity considering participants in both groups could have been impacted by the above factors.

Strengths, Limitations, and Future Research

There are several strengths and limitations of this study. Notably, this was a pragmatic study design that was scientifically rigorous yet performed in a real-world setting. Further, while there was no difference between groups in outcomes, this study has detailed two interventions (PMC and enhanced UC) that produced clinically significant changes. This allows clinicians multiple options when deciding on an intervention that is suitable and effective for their patient. The primary limitation of this study is the dearth of information on participants’ HEP, protein supplementation adherence, and care transition received secondary to incomplete participant, therapist, and coordinator logs. Given both exercise dose and protein intake greatly influence high-intensity outcomes, we cannot definitively determine whether the PMC intervention would produce greater outcomes in ideal conditions (ie, high adherence to all components). Within this study, formal evidence-based behavior change strategies were not employed to enhance adherence to intervention components performed outside of the therapy session (ie, HEP, protein supplementation). Future studies should consider use of behavior change strategies, such as motivational interviewing, to encourage high adherence for all intervention components. Greater adherence to exercise outside intervention sessions may be especially important with a limited number of intervention sessions (12 visits over 60 days). Another potential limitation was that protein supplementation was guided by general recommendations and was not individually tailored to each participant. Future studies may consider the use of a study nutritionist to provide protein supplementation that aligns with each participant’s needs, based on body mass and current diet, to ensure that participants have capacity for physical function improvements. An additional limitation is the heterogeneity of the study sample. While heterogeneity can be a strength as the study results are applicable to larger population, it may also weaken study effects as certain characteristics within the sample (eg, baseline function, presence of certain comorbidities, motivation) can influence a participant’s capacity to improve via the PMC intervention. In line with the heterogeneity of baseline scores and demographics, there was also large heterogeneity in intervention response as evidenced by outcome confidence intervals (Table 3). While heterogeneity was present in our sample and participants’ response to the interventions, it should be noted that participants in this study are not reflective of all older adults receiving home health care. Our selection criteria aimed to ensure that deconditioned and multimorbid participants could safely complete the intervention. Therefore, some common diagnoses like lower extremity fractures or acute stroke were excluded as were those with coexisting barriers (eg, insurance incompatibility or profound cognitive deficits). This resulted in a high number of patients being excluded. Future studies examining home health may seek to expand inclusion criteria and employ pre-planned subgroup analyses to understand if certain participants respond more or less favorably to specific interventions.

A high-intensity progressive, multicomponent intervention is safe and feasible but not more effective at improving physical function in older adults who are medically complex and recently discharged from a hospital or SNF than enhanced UC. Future work is needed to 1) identify whether shared aspects of these interventions promoted clinically significant changes in physical function, 2) evaluate the optimal frequency of therapy to facilitate functional improvements, 3) explore who is most likely to benefit from high intensity or low intensity exercise, and 4) determine whether the PMC intervention can be optimized in the home health setting to approximate the health outcomes improvements seen in other clinical populations secondary to high-intensity exercise.

CRediT – CONTRIBUTOR ROLES

Alexander J. Garbin (Formal analysis [Supporting], Visualization [Lead], Writing—original draft [Lead], Writing—review & editing [Lead]), Jason Falvey (Methodology [Supporting], Writing—review & editing [Supporting]), Ethan Cumbler (Investigation [Supporting]), Danielle Derlein (Data curation [Lead], Project administration [Lead], Validation [Lead]), Deborah Currier (Investigation [Supporting]), Amy Nordon-Craft (Methodology [Supporting]), Robert Will (Investigation [Supporting]), Maegan Olivos (Investigation [Supporting], Writing—review & editing [Supporting]), Jeri E. Forster (Data curation [Supporting], Formal analysis [Lead], Writing—review & editing [Supporting]), Kathleen Kline Mangione (Conceptualization [Supporting], Investigation [Supporting], Methodology [Supporting], Writing—review & editing [Supporting]), Jennifer E. Stevens-Lapsley (Conceptualization [Lead], Funding acquisition [Lead], Investigation [Lead], Methodology [Lead], Supervision [Lead], Writing—review & editing [Supporting]).

ACKNOWLEDGMENTS

The authors extend a sincere thank you to Berkley Home Health Care in Denver, Colorado, and the Colorado Visiting Nurse Association for their invaluable support of this study through their agencies. The authors also thank the following individuals for their contributions to this work: Allison Gustavson, PT, DPT, PhD; Julie Anderson, PT, DPT; Tanya Budnikova, MBA; Mike Gummerson, PT; Megan Mitchell, PT; Jared Poole, PT, DPT; Gary Ruvins, JD, MBA; Derek Smith, PT, DPT; Julie Stutzbach, PT, DPT, PhD; and David Williams, PT, DPT.

ETHICS APPROVAL

This randomized controlled trial received ethics approval from the University of Colorado Multiple Institutional Review Board.

FUNDING

This trial is funded by the National Institute on Nursing Research and the Eunice Kennedy Shriver National Institute on Child Health and Human Development (grant number R01NR016209; https://clinicaltrials.gov/ct2/show/NCT02905370). Boost Nutritional Drinks were provided via a research grant from Nestlé HealthCare Nutrition, Inc.

CLINICAL TRIAL REGISTRATION

The trial was registered at clinicaltrials.gov (NCT02905370).

DISCLOSURES

The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.

Kathleen Mangione had a subcontract with the NIH through the University of Colorado.

Although Robert Will was employed at Denver’s safety net hospital during the clinical trial, he did not refer patients to the study.

DATA AVAILABILITY

Data will be made available upon request.