Inpatient rehabilitation patients are more active during rehabilitation times and days than non-rehabilitation times and days: An observational study

Abstract

Introduction:

The effectiveness of inpatient rehabilitation services has been well supported. However, patients receiving inpatient rehabilitation may be physically inactive outside of therapy. The primary objectives were to determine differences in activity levels between rehabilitation time and non-rehabilitation time and rehabilitation day and non-rehabilitation day during inpatient rehabilitation stay. The secondary objective was to determine differences among rehabilitation services.

Methods:

In all, 50 participants wore an ActiGraph GT9X Link on their non-dominant wrist during their inpatient rehabilitation stay.

Results:

We found significant differences in total activity counts between rehabilitation time and non-rehabilitation time (z = 6.15, p < 0.001) and between rehabilitation day and non-rehabilitation day (z = 5.82, p < 0.001). We found significant differences in total activity counts between occupational therapy and physical therapy (z = 5.71, p < 0.001), occupational therapy and speech-language pathology (z = 3.41, p < 0.001), and physical therapy and speech-language pathology (z = 3.07, p = 0.002). The degrees of all differences were large.

Conclusion:

To increase the activity levels during non-rehabilitation times and days, patients receiving inpatient rehabilitation may benefit from strategies which increase activity levels such as an individualized program that can be safely and independently performed outside of rehabilitation sessions.

Introduction

Individuals admitted to an inpatient rehabilitation facility in the United States receive an intensive, multidisciplinary rehabilitation including occupational therapy (OT), physical therapy (PT), and/or speech-language pathology (SLP). Most of these individuals are in the process of recovery from one or more conditions such as neurological, orthopedic, and cardiopulmonary conditions. The effectiveness of inpatient rehabilitation services in improving functional outcomes has been well documented (Galloway et al., 2013). However, patients are sedentary during their inpatient rehabilitation stay. Stroke patients receiving inpatient rehabilitation spent 87% of their weekday waking hours sedentary, measured by an activity monitor and became more sedentary during weekend (Barrett et al., 2018). Similar patterns were observed among older adults receiving inpatient rehabilitation whose activity levels were significantly lower than their counterparts in the community (Smith et al., 2008). These low activity levels during inpatient rehabilitation are concerning due to their potential negative impact on health and rehabilitation outcomes. Among patients in acute inpatient settings, lower activity levels were associated with longer hospital stay following surgery (Abeles et al., 2017). In addition, the more sedentary the acute stroke patients were, the lower physical functioning they showed during daily activities (Mattlage et al., 2015). Lastly, a low activity level has been associated with negative health outcomes, such as cardiovascular mortality, recurrent falls, functional limitation, and cognitive decline (Cacciatore et al., 2019; Cunningham et al., 2020).

The presence of a healthcare professional may help individuals increase activity levels during the day. This is especially important for those individuals who are not safe to mobilize without supervision. When therapists or nurses were present, acute stroke patients were most active and frequently engaged in mobile activities (Bernhardt et al., 2007). While the support from healthcare professionals is available throughout the day during inpatient stay, at post-discharge, patients’ access to support and services that help maintain health and function is greatly diminished. This decreased availability for support and services can lead to decreased activity levels following discharge to the community (Ezeugwu and Manns, 2017). Therefore, experience of increased activity level during rehabilitation process can be a learning opportunity for patients to be motivated to perform activities outside of rehabilitation sessions.

The primary objectives of this study were to determine differences in activity levels between (1) the time of receiving rehabilitation services (rehabilitation time) and the time of not receiving rehabilitation services (non-rehabilitation time) and (2) the day of receiving rehabilitation services (rehabilitation day) and the day of not receiving rehabilitation services (non-rehabilitation day) during inpatient rehabilitation stay. The secondary objective was to determine differences in activity levels among OT, PT, and SLP. We hypothesized that activity levels during the time of receiving rehabilitation services and the day of receiving rehabilitation services are significantly higher than the time of not receiving rehabilitation services and the day of not receiving rehabilitation services, respectively.

Method

Study design and participants

This study used a prospective observational design to investigate activity levels of patients during their inpatient rehabilitation stay. This design was selected for our study because it allowed us to collect activity level data prospectively (longitudinal data collection) by observation without controlling rehabilitation services (a lack of intervention control). Participants were patients who were receiving OT, PT, and/or SLP in the 55-bed inpatient rehabilitation facility attached to the level-1 trauma medical center in Greenville, NC in the United States. Inclusion criteria were (1) being 18 years or older and (2) recovering from one or more of the following conditions, but not limited to: stroke, brain injury, specified neurologic conditions, major multiple trauma, congenital deformity, burns, amputation, systemic vasculitis with joint involvements, fracture of the hip, knee or hip replacement, active polyarthritis, and/or severe or advanced osteoarthritis (Forrest et al., 2019). Exclusion criteria were (1) admitting diagnosis and/or past medical history of spinal cord injuries; (2) having isolation precautions; (3) the presence of moderate-to-severe cognitive impairment according to the Six-Item Screener score of 3 or below (three or more errors) (Callahan et al., 2002); (4) inability to wear ActiGraph on their wrist 24-h per day for 9–11 days; and (5) being expected to stay in the facility less than three full rehabilitation days. Those diagnosed with spinal cord injury or having isolation precautions were excluded due to the potential excessive limitation of activities. The length of the data collection was up to 2 weeks, depending on their length of inpatient rehabilitation stay. The data collection began September 2020 and was completed in September 2021 during which restrictions due to coronavirus disease 2019 (COVID-19) were in place. Three occupational therapists who worked in the inpatient rehabilitation facility and also were study team members referred participants to the study. This study was approved by the University and Medical Center Institutional Review Board of East Carolina University and ECU Health in Greenville, NC in the United States.

Instrumentation

Demographic and medical information

Demographic information including age, gender, and marital status was collected during the initial intake visit. Medical information, such as acute care and inpatient rehabilitation lengths of stay, admitting diagnoses, admission and discharge scores of Self-Care and Mobility Section GG, and therapy dates and times, was collected through electronic health record review after participant’s discharge from the facility. Self-care and Mobility Section GG consisting of 24 items is the outcome measure on self-care and mobility that are required for all post-acute care settings in the United States (American Occupational Therapy Association, 2023).

ActiGraph GT9X Link

ActiGraph GT9X Link accelerometer was used to collect objective activity levels during OT, PT, and SLP and non-rehabilitation times. ActiGraph GT9X Link uses triaxial accelerometer capturing vertical, antero-posterior, and medio-lateral motions. Using the motion data, ActiGraph can monitor various activity levels, such as total activity counts (TAC), step counts, and metabolic equivalent of tasks (ActiGraph, 2023; Wolff-Hughes et al., 2015; Kelly et al., 2013). ActiGraph objectively measures activity levels by quantifying complex motions in real time, compared to self-reported activity levels that rely on individuals’ memory recall and were shown to be under or over-estimated (Prince et al., 2008). ActiGraph showed good reliability and validity with favorable overall accuracy in older adults (Heesch et al., 2018). In addition, ActiGraph may be more appropriate to wear on the wrist under free-living conditions due to more frequent hand and wrist movement than waist movement in free-living conditions (Tudor-Locke et al., 2015). Participants wore the ActiGraph on their non-dominant or less-affected wrist, rather than waist, to increase the wearing compliance. Participants were asked to wear the ActiGraph for 24 h per day until a study member returned to collect it.

Data, specifically vector magnitude TACs that is a combination of triaxial data (ActiGraph, 2020b), were downloaded and analyzed using the ActiLife Software (Pensacola, FL, USA) (ActiGraph, 2020a). TACs are considered as the alternative for the PA volume (Wolff-Hughes et al., 2015), and the counts algorithm for the TACs has been recently published (Neishabouri et al., 2022). For wear time validation, Choi algorithm (Choi et al., 2011) was used with the “use vector magnitude” option selected. The “exclude non-wear time analysis” option was also selected to ensure that correct wear time duration was calculated for rehabilitation day and non-rehabilitation day durations. A minimum of 3 days of wear time was required for rehabilitation days. A minimum of 1 day (>6 h) of wear time was required for non-rehabilitation days to be included in the data analysis for the difference between rehabilitation day and non-rehabilitation day (Primary Objective 2). Only the TACs collected during 6 am–6 pm, except for the last data collection date, were considered for data analyses. For the last data collection date, only TAC collected during 6 am–4 pm were included in data analyses because a study member collected ActiGraph after 4 pm. For the TACs of rehabilitation times, only the TACs during rehabilitation time durations that were identified by electronic health record review were downloaded. According to the electronic health record review, no more than one rehabilitation discipline was provided at the same time. Lastly, TACs per minute for each rehabilitation discipline, rehabilitation and non-rehabilitation time, and rehabilitation and non-rehabilitation day were calculated. This calculation for TAC per minute was necessary because the total durations of rehabilitation and non-rehabilitation times and days varied among participants.

Procedure

After potential participants were referred to the study, a study member visited the patient’s room for eligibility screening. Once participants are determined eligible for the study, they voluntarily signed on the written informed consent form that was written in no greater than sixth grade level, in short sentences, and in non-technical and non-professional terms to describe the study. Following obtaining the signed consent form, the study member interviewed a participant to collect demographic information. Then, the study member put the ActiGraph on the non-dominant or less-affected wrist. When possible, a study member collected the ActiGraph a day before the participant’s discharge date. Otherwise, the ActiGraph was collected from the nursing staff in the facility. After participant’s discharge from the inpatient rehabilitation facility, a study member accessed the electronic health record to collect medical information. The interviewers and health record reviewers were occupational therapists or occupational therapy graduate students who were trained in all study-related activities prior to the independent administration.

Data analysis

Due to the absence of previous data, power analysis was not able to be completed. Power analysis is used to estimate sample size of the study using a level of significance, desired power, and an expected effect size based on a similar previous study (Portney, 2020). Instead, a goal of recruiting 50 participants was established based on the potentially available patient pool in the facility, the planned study period (1 year), and the availability of participant incentives. Demographic and medical information was reported as means with standard deviation for continuous variables, such as age, and frequencies with percentages for categorical variables, such as gender. For assumption checking, appropriate datasets for each analysis were visually inspected using histograms and box plots for normal distribution. Assumptions of statistical tests are conditions to be met to ensure the accuracy of test statistics and p value (Field, 2013). Due to a lack of normal distribution of the datasets that is one of the assumptions, the nonparametric Wilcoxon Signed Rank Test (Field, 2013:228) was used to determine the differences in TACs between rehabilitation and non-rehabilitation times, rehabilitation and non-rehabilitation days, and among OT, PT, and SLP. z and p values from the Wilcoxon Signed Rank Test were reported. For the differences among OT, PT, and SLP, an alpha level of 0.017 by applying the Bonferroni correction was used to be conservative (Field, 2013: 67–69). For all other statistical analyses, an alpha level of 0.05 was used. When a p value is smaller than an alpha level, a null hypothesis is rejected. IBM Statistical Package for Social Sciences for Mac, Version 27.0 (Armonk, NY) was used for all analyses. In addition to the statistical significance, the effect size r was calculated for each difference. The degrees of difference according to r were interpreted as: negligible for 0–0.09, small for 0.10–0.23, medium for 0.24–0.36, large for 0.37–0.70, and very large for 0.71 or greater (Tickle-Degnen, 2001:318–319).

Results

Out of 74 individuals referred, 60 individuals participated in the study with data from 50 participants available for analyses (Figure 1). The mean (SD) ages were 64.2 (12.0) years, and over a half of participants were female (54%). Approximately equal numbers of participants were admitted to the inpatient rehabilitation facility with orthopedic (34%), cardiopulmonary (32%), or neurological and other diagnoses (34%). On average, participants required partial assistance to supervision/touching assistance (verbal cues, touching, and/or contact guard assistance) at admission and setup assistance (setting-up or cleaning-up assistance prior to or following the activity) at discharge in self-care and mobility according to the Section GG items (Table 1).

Figure 1.

A flow diagram for participant recruitment.

Table 1.

Participants’ demographic and medical information.

Demographic and medical information	n = 50
Age (year), M (SD)	64.2 (12.0)
Sex, n (%)
Female	27 (54.0)
Male	23 (46.0)
Married, n (%)	23 (46.0)
Acute care length of stay (day), M (SD)	14.8 (24.5)
Inpatient rehabilitation length of stay (day), M (SD)	13.2 (3.6)
History of major diagnoses, n (%)
CVA	11 (22.0)
Depression	9 (18.0)
COPD	7 (14.0)
Cancer	7 (14.0)
PVD	6 (12.0)
Anxiety	5 (10.0)
Hypothyroidism	3 (6.0)
Inpatient rehabilitation admitting diagnosis, n (%)
Orthopedic	17 (34.0)
Cardiopulmonary	16 (32.0)
Neurological & Other	17 (34.0)
Section GG: Self-care items, M (SD)
Admission score	3.6 (0.6)
Discharge score (n = 49)	5.2 (0.7)
Section GG: Mobility items, M (SD)
Admission score	3.3 (0.6)
Discharge score (n = 49)	5.0 (0.7)

CVA: cerebrovascular accident; COPD: chronic obstructive pulmonary disease; PVD: peripheral vascular disease.

The mean (SD) wear time of ActiGraph per day during the rehabilitation days was 697.9 (58.1) minutes and the mean (SD) wear time during non-rehabilitation days was 682.1 (87.0) minutes. We found significant differences in TACs between rehabilitation time and non-rehabilitation time (z = 6.15, p < 0.001). We also found significant differences in TACs between rehabilitation day and non-rehabilitation day (z = 5.82, p < 0.001). The degrees of both differences were large (r = 0.62 and 0.61, respectively) (Table 2). We found significant differences in TACs between OT and PT (z = 5.71, p < 0.001), OT and SLP (z = 3.41, p < 0.001), and PT and SLP (z = 3.07, p = 0.002). The degrees of these differences were also large (r = 0.57, 0.62, and 0.56, respectively) (Table 3). Only 15 participants received SLP services; therefore, the data from those 15 participants were analyzed to compare OT and SLP, and PT and SLP. According to the electronic health record review, there was no overlap among rehabilitation service times, indicating no group therapy among OT, PT, or SLP.

Table 2.

Results of differences in activity levels between rehabilitation time and non-rehabilitation time and rehabilitation day and non-rehabilitation day.

Rehabilitation and non-rehabilitation times and days	Total activity count per minute, Mdn (Q1, Q3)	r, z, p value
Rehabilitation and non-rehabilitation times (n = 50)
Rehabilitation time	1693.45 (1152.14, 2157.45)	0.62, 6.15, <0.001
Non-rehabilitation time	703.11 (509.26, 1303.79)
Rehabilitation and non-rehabilitation days (n = 46)
Rehabilitation day	964.54 (669.18, 1523.84)	0.61, 5.82, <0.001
Non-rehabilitation day	712.59 (473.00, 1123.93)

Total activity counts were measured by ActiGraph.

Mdn: median; Q: quartile.

Table 3.

Results of differences in activity levels among occupational therapy, physical therapy, and speech-language pathology (n = 50).

Rehabilitation discplines	Total activity count per minute, Mdn (Q1, Q3)
Occupational therapy	2024.37 (1490.21, 2586.85)
Physical therapy	1381.09 (968.64, 1744.71)
Speech-language pathology a	761.30 (323.03, 1610.01)
Occupational therapy–physical therapy	0.57, 5.71, <0.001 b
Occupational therapy–speech-language pathology a	0.62, 3.41, <0.001 b
Physical therapy–speech-language pathology a	0.56, 3.07, 0.002 b

Total activity counts were measured by ActiGraph.

Mdn: median; Q: quartile.

^an = 15.

^br, z, p value.

Discussion

In our study, we assessed objective activity levels of inpatient rehabilitation patients to investigate the differences in activity levels between rehabilitation and non-rehabilitation times and days and among different rehabilitation services.

Activity levels during inpatient rehabilitation

We found that patients at the inpatient rehabilitation facility were significantly more physically active during rehabilitation times and days than non-rehabilitation times and days, as hypothesized. These results are consistent with previous findings. Significantly reduced uptime in older adults in inpatient rehabilitation on weekend days compared to weekdays was found, whereas there was no difference in community-dwelling older adults (Smith et al., 2008). The similar results were also found among patients receiving inpatient rehabilitation where patients showed decreased activity levels, fewer steps, and less time in upright activities during weekends (Bear-Lehman et al., 2001; Peiris et al., 2012). Our findings are more detailed and accurate because we collected objective activity levels that are measured by a tri-axial accelerometer worn at wrist. The previous findings were based on the observations or a uniaxial accelerometer worn on hip which may not accurately detect complex movements during rehabilitation.

Our findings on the differences in activity levels may have been due to pervasive sedentary behaviors during non-rehabilitation times and days at inpatient rehabilitation facilities. Treatments during non-rehabilitation times typically consist of patient care and leisure activities (Whiteneck et al., 2011) that are not as physically involved or intensive as the time spent during rehabilitation services. In addition, time not spent in hospital care often includes personal leisure activities, such as resting, visiting with family and/or friends, watching television, and reading most of which are considered sedentary. There can be even fewer opportunities to participate in activities during pandemic outbreaks.

Another potential reason can be the high usage of individual sessions over group sessions during rehabilitation times compared to non-rehabilitation times and other treatment sessions. According to a study done at nine United States and one Canadian inpatient rehabilitation facilities, only 13.7% of therapy sessions, specifically 10% for OT, 8% of PT, and 16% of SLP services, were provided in a group (Hammond et al., 2015). Our participants were guided and assisted to perform therapeutic activities one-to-one by the therapists without any group sessions. In addition, our participants required partial assistance to supervision/touching assistance in self-care and mobility at admission to the inpatient rehabilitation facility. This means that most of our participants, if not all, were on fall risk precautions in this facility. Therefore, during the individual sessions, it is likely that participants received more attention from therapists and were more encouraged to actively move without safety concerns because therapists are required to closely monitor and guide the sessions. Furthermore, our participants may have had higher energy and less fatigue during non-rehabilitation times of the rehabilitation days compared to non-rehabilitation days. Among older adults, lower numbers of steps and counts from ActiGraph were associated with higher levels of fatigue (Egerton et al., 2016). This may be similar among patients receiving inpatient rehabilitation because increased activity levels during rehabilitation may have lowered their levels of fatigue and, in turn, increased their activity levels outside of the rehabilitation sessions.

Activity levels during different types of rehabilitation services

We also found significantly higher activity levels measured at the wrist during OT followed by PT and SLP. These differences may have been due to the different focuses and goals of rehabilitation services. OT treatment sessions at inpatient rehabilitation facilities typically aim to improve performance in daily activities and utilize meaningful activities, or occupations, to the individuals to meet this goal (American Occupational Therapy Association, 2020). These OT interventions may be motivating for patient engagement in therapy sessions because the elements of occupations may elicit feelings of self-determination and autonomy responsibility (Kennedy and Davis, 2017). In the previous study, most of PT sessions (25–38%) were spent on gait activities regardless of mobility limitations (Beaulieu et al., 2015). They also found that most of SLP sessions were spent on swallowing activities (20%), followed by problem-solving/reasoning activities (18%) and verbal expression activities (10%) in multiple inpatient rehabilitation facilities (Beaulieu et al., 2015). While these PT and SLP focuses are important to clients, it may be useful to jointly explore ways to increase activity levels during rehabilitation and non-rehabilitation time. Therefore, rehabilitation sessions incorporating a wide variety of activities may offer more therapy options for individuals with differing abilities.

Implications for practice

Based on our findings, rehabilitation services may hold a critical role to reduce the overall sedentary and bedrest time among individuals receiving inpatient rehabilitation. While rest is necessary for healthy recovery in all adults, bedrest as a general or primary treatment does not improve or can even worsen patient outcomes (Allen et al., 1999). Therefore, increasing activity levels through early, intensive, and possibly longer rehabilitation sessions may be needed for better outcomes for individuals in inpatient rehabilitation facilities. In addition, enhancing generalizability of therapy interventions through adaptations, such as performing motivated activities in fall risk-minimized settings (sitting or standing with a chair behind), and/or individualized programs that can be implemented during non-rehabilitation times may help to increase activity levels over the course of inpatient rehabilitation stay. A group of researchers investigated the effectiveness of patient-directed activity program in increasing activity levels among individuals post stroke in an inpatient rehabilitation facility (Swank et al., 2020). This patient-directed program consisted of two 30-min functional individualized activities that were developed by therapists and physiatrists, who are medical doctors specialized in physical medicine and rehabilitation, and independently performed by participants. The experimental group significantly increased the number of steps at discharge and 3-month follow-up from admission compared to the control group. Therefore, patients may benefit from an individualized program that is a safe extension of guided rehabilitation sessions and can be performed independently outside of rehabilitation sessions. Patient education on the importance of independently performing the program and tracking compliance are also essential to maximize the program impact. When assessing the effectiveness of an individualized program and patient education to increase activity levels, it would be also important to take needed resources, such as staffing and cost–benefit, into consideration.

Limitations of the study and opportunities for future research

This study has several limitations. Although our participants’ admitting diagnoses were evenly split among orthopedic, cardiopulmonary, and neurological and other diagnoses, we did not target certain diagnoses, such as stroke or heart failure. Therefore, our results are applicable to the general inpatient rehabilitation population. Increasing numbers of recruitment sites and participants stratified by certain diagnoses groups may resolve this limitation in a future study. Another limitation is the placement of ActiGraph. We chose to use the wrist ActiGraph placement to improve potential wearing compliance without much monitoring support. It is unclear how the wrist placement may have affected our results compared to the waist placement in this population due to the lack of previous literature. For example, our participants may have used an ambulation device, such as a rolling walker or a cane, during rehabilitation sessions, and it is unknown how the use of ambulation devices may affect the accelerometer outcomes. Therefore, placing ActiGraph on both wrist and waist and adopting the higher number between these two placements may result in more accurate outcomes. A support from nursing staff for monitoring the wearing compliance would be also critical.

Conclusion

Patients receiving inpatient rehabilitation showed higher activity levels during rehabilitation times and days than non-rehabilitation times and days and during OT compared to PT or SLP. To increase the activity levels during non-rehabilitation times and days, patients receiving inpatient rehabilitation may benefit from strategies which increase activity levels such as an individualized program that can be safely and independently performed outside of rehabilitation sessions. Future studies can consider recruiting participants in multiple inpatient rehabilitation facilities and using multiple accelerometers for better generalizability and more accurate outcomes.

Key findings

• Inpatient rehabilitation patients are less active outside of rehabilitation sessions.

• Patients are most physically active (as measured at the wrist) during occupational therapy followed by physical therapy and speech-language pathology.

What the study has added

• This study confirms the low activity levels of individuals receiving inpatient rehabilitation and shows differences in activity levels measured at the wrist among different rehabilitation services.