Abstract
Analysis of 12-mo post–inpatient rehabilitation discharge interviews found a relationship between exercise and hospital readmission among patients with SCI; follow-up research is needed to determine whether regular exercise reduces hospital readmissions among this population.
Spinal cord injury (SCI) occurs after trauma to the spinal cord as a result of a car accident, fall, gunshot, or some other impact to the spine resulting in bruising, stretching, piercing, tearing, or severing of the spinal cord (National Institute of Neurological Disorders and Stroke, 2019). Damage to the spinal cord can result in complete loss of or partial impairment in motor and sensory functions at and below the level of injury. SCI often results in mobility impairments, requiring reliance on assistive devices such as walkers or manual or power wheelchairs.
After a traumatic SCI, the highest risk of hospital readmission occurs in the first year postinjury, with readmission rates as high as 45% (DeJong & Groah, 2015; DeJong et al., 2013; National Spinal Cord Injury Statistical Center [NSCISC], 2019; Skelton et al., 2015). Engagement in regular physical activity and exercise has been shown to reduce hospital readmissions in other populations (Courtney et al., 2009 ; Esteban et al., 2014); however, evidence is limited regarding the relationship of engagement in routine exercise by patients with SCI and hospital readmission rates among people with SCI during the first year postinjury.
Exercise is defined as a subcomponent of physical activity, an intentional activity to promote health through physical fitness (e.g., strength training, running). Physical activity is a broader term that includes any activity that expends energy and engages the musculoskeletal system, such as taking a walk, playing a sport, performing manual labor, or climbing stairs (Dasso, 2019). After SCI, participation in physical activity and exercise are reported to be challenging as a result of mobility impairments, accessibility of the environment, level of assistance needed to participate, financial limitations, knowledge of how to overcome obstacles to participate in physical activity and exercise, and fear of injury (Kehn & Kroll, 2009).
Hospital admissions, use of health care services, and expenditures on health care services increase after a person sustains an SCI (NSCISC, 2019). After initial SCI, the median hospital length of stay (LOS) is 11 days; in an inpatient rehabilitation setting, the initial LOS is 34 days (NSCISC, 2019). The most common causes of hospital readmissions after SCI are genitourinary infections, skin breakdown (pressure injuries), respiratory deficits, digestive disorders, circulatory system disease, and musculoskeletal injuries (DeJong et al., 2013; NSCISC, 2020; Skelton et al., 2015). Previous evidence has indicated that exercise after SCI can mitigate many of these risks, resulting in improvement in circulatory system function (Phadke et al., 2019), prevention of pressure injuries (Krause & Broderick, 2004), improvement in pulmonary function (Li et al., 2018), and enhancement of musculoskeletal function (Phadke et al., 2019).
In addition to these benefits, exercise can provide many other health benefits, including improved health-related quality of life; longer lifespan; independence in daily life activities; reduction in pain, blood pressure, weight, and hemoglobin A1c level; and improved cognition, sleep quality, endurance, and strength post-SCI (Agho & John, 2017). Other studies have found that sedentary behaviors post-SCI were associated with higher rates of depression (Mulroy et al., 2016); increased hospitalizations, days in hospital, and number of physician visits (Krause et al., 2013); and a decline in pulmonary function, higher body mass index, lower inspiratory muscle strength, and reduced fitness ability (Postma et al., 2013).
Despite the benefits of exercise identified in the evidence, no national physical activity or exercise guideline has been established for people with SCI because of the inconsistent exercise types studied, small sample sizes, limited longitudinal data, and limited research investigating any associated risks (National Institutes of Health [NIH], Office of Disease Prevention, 2021). In addition to the lack of guidelines, people with SCI experience barriers that restrict their access to and participation in sustained exercise post-SCI. Physical deficits experienced after SCI can make exercise challenging without proper education and guidance from a qualified professional. Environmental and community barriers that may be experienced are transportation limitations, building inaccessibility (both inside and outside), inaccessible gym equipment, staff’s lack of experience with and knowledge about SCI, and trainers’ and health care professionals’ lack of knowledge about adaptive exercise programming (Gorgey, 2014). This situation presents both a need and an opportunity for occupational therapy practitioners to address the many barriers to accessing, participating in, and independently engaging in sustained client-centered meaningful physical activity in the community or home after SCI.
The purpose of this study was to add to the current body of knowledge about the health-related benefits of exercise after SCI by evaluating the significance of a correlation between exercise and hospital readmission during the first year after discharge from inpatient rehabilitation.
Method
Research Design
We conducted a secondary analysis of a data set that was previously collected as part of a prospective longitudinal cohort study by the SCIRehab Project (Whiteneck, 2018). The SCIRehab study includes data collected from participants ages 12 yr and older who sustained an acute SCI and were admitted to one of six participating inpatient rehabilitation facilities in the United States between 2007 and 2009. The SCIRehab study collected more than 6,000 variables from 1,376 individuals and is publicly available online (https://www.icpsr.umich.edu/web/ADDEP/studies/36724). Each participant sustained either a complete or an incomplete SCI and required intensive inpatient rehabilitation that consisted of occupational therapy, speech therapy, physical therapy, therapeutic recreation, social work, psychology, and nursing services.
For this study, we analyzed data from 520 participants in the SCIRehab Project study who completed an inpatient rehabilitation facility stay, completed the 12-mo postdischarge interview, and provided responses on self-reported hospital readmissions and self-reported monthly (number of months since discharge), weekly (average days per week over the past 12 mo), and daily (average minutes per day over the past 12 mo) exercise frequency since discharge from inpatient rehabilitation. The SCIRehab study defined exercise as participation in general exercise outside of scheduled therapy sessions, which could include functional electrical stimulation or other forms of exercise (Whiteneck, 2018). Participants were included in the study if they reported no exercise since discharge or monthly exercise since discharge (i.e., minimum of 1 time per month for 12 mo).
Data Analysis
Descriptive statistics consisted of frequencies and percentages for 12-mo hospital readmission frequency, length of hospital readmission stay, age at injury category, sex, and neurological injury level. Then, to assess the relationship between monthly exercise and a hospital readmission occurrence, we used a χ2 statistical test to measure the relationship of monthly exercise since discharge to hospital readmission since discharge (Dunlap et al., 2000). The χ2 analysis was conducted for the entire sample and then for each neurological injury level (complete paraplegia, incomplete paraplegia, complete tetraplegia, and incomplete tetraplegia). After determining statistical significance with a χ2 test, a follow-up φ correlation coefficient was calculated to determine the strength of the relationship. Pearson correlations were computed to determine linear relationships between the continuous variables: weekly exercise in days, daily exercise in minutes, number of hospital readmissions, and number of hospital readmission days. The Pearson correlation is an appropriate statistical method to determine whether there is a relationship between two continuous variables (Kellar & Kelvin, 2013).
Results
Demographics
Data from 520 SCIRehab Project participants were used in this study. The extracted data sample was 80% male (n = 416) and 20% female (n = 104) and ranged in age from 12 yr to >80 yr. The percentage of the sample in each age category at the time of SCI was as follows: 20–29 yr, 27.5%; ages 30–39 yr, 15%; 40–49 yr, 13.1%; 50–59 yr, 14.2%; 60–69 yr, 10.4%; 70–79 yr, 2.9%; and ≥80 yr, <1%.
Each participant was assigned an SCI classification at discharge from inpatient rehabilitation, which was determined by a physician using the American Spinal Cord Injury Association Impairment Scale (Rupp et al., 2021). The six SCI classifications and corresponding frequencies are as follows: incomplete paraplegia, n = 70; complete paraplegia, n = 137; paraplegia with minimal deficits, n = 4; incomplete tetraplegia, n = 173; complete tetraplegia, n = 103; tetraplegia with minimal deficits, n = 5 (28 participants had unknown or missing classifications; Table 1). All neurological injury classification and age at injury categories had more male than female participants. Across each of the different neurological condition categories, most injuries had occurred among participants in the 20–29 yr range, apart from incomplete tetraplegia, in which the highest rate of injury had occurred among participants in the 60–69 yr range. Of the 520 participants, 229 indicated they had not exercised at all from discharge to 12 mo postdischarge, whereas 291 indicated they had exercised monthly since discharge from inpatient rehabilitation (see Table 1). Of the 520 participants, 155 had at least one or more hospital readmissions. Mean hospital readmission rates and days spent in the hospital were higher for the no-exercise group (M = 0.86, SD = 1.44) than for the exercise group (M = 0.58, SD = 1.07; see Table 1). Total hospital readmission rates and days spent in the hospital were lower across all neurological categories in the exercise group except the complete paraplegia category.
Demographics of the Sample by Neurological Injury Category
Note. The paraplegia with minimal impairment and tetraplegia with minimal impairment categories were not included because of low sample size (n = 4 and n = 5, respectively). Twenty-eight participants had unknown or missing SCI classifications and are included in the All SCI category. SCI = spinal cord injury.
Exercise and Hospital Readmissions
We analyzed two binomial variables using Pearson χ2; respondents answered yes or no regarding whether they had exercised monthly since discharge and whether they had experienced a hospital readmission since inpatient rehabilitation discharge. A statistically significant negative correlation (φ = −.091, p = .038) was identified between monthly exercise and hospital readmission over the first 12 mo after traumatic SCI, χ2(1, N = 520) = 4.30, p = .038, with an 8.4% greater hospital readmission rate experienced by the no-exercise group (Table 2). Although the φ coefficient was statistically significant, −.091 indicates a small effect size.
Exercise and Hospital Readmission Rates (N = 520)
Note. Difference in hospital readmission rate between exercise and no-exercise group = 8.4%.
Exercise and Hospital Readmissions by Neurologic Injury Category
We conducted a Pearson χ2 for each of the four SCI neurologic injury categories, using the same dichotomous binomial variables, monthly exercise and experiencing a hospital readmission over the 12 mo postdischarge. The results for each of the four SCI categories are as follows: incomplete paraplegia, χ2(1, N = 70) = 3.48, p = .062, φ = −.223; complete paraplegia, χ2(1, N = 137) = 0.17 p = .684, φ = −.035; incomplete tetraplegia, χ2(1, N = 173) = 2.82, p = .093, φ = −.128; and complete tetraplegia, χ2(1, N = 103) = 0.16, p = .694, φ = .039. We found no statistically significant relationships between monthly exercise and experiencing a hospital readmission for any of the four individual SCI classification categories (Table 3).
Relationship of Exercise and Hospitalization by SCI Neurological Injury Category
Note. Twenty-eight participants had unknown or missing SCI classifications and are included in the All SCI category. SCI = spinal cord injury.
p < .05.
Exercise Frequencies and Hospital Readmission Rates
We used a Pearson correlation to analyze the relationships of continuous variables, including self-reported average days per week and average minutes per day spent exercising and number of hospital readmissions since discharge from inpatient rehabilitation. The results revealed a statistically significant negative relationship between exercise duration (average minutes per day) and number of hospital readmissions, r(518) = −.093, p = .034. The results also revealed no statistically significant relationship between exercise frequency (average days per week) and number of hospital readmissions, r(518) = −.083, p = .059.
Exercise Frequencies and Hospital Readmission Days
We used a Pearson correlation to conduct a bivariate analysis to determine the relationship of duration or frequency of exercise to number of hospital readmission days experienced by participants. The results indicated no statistically significant relationship between exercise duration (average minutes per day) and hospital readmission days, r(518) = −.084, p = .056. However, the results did indicate a statistically significant negative relationship between exercise frequency (average days per week) and hospital readmission days, r(518) = −.101, p = .022.
Discussion
The average hospital readmission rate (29.8%) found for the entire sample was lower than the 45% rate found in a recent study by Skelton et al. (2015) and lower than the 36%–37% readmission rate found in other studies (DeJong & Groah, 2015; DeJong et al., 2013; NSCISC, 2019). Regarding exercise frequency, almost half (44%) of the sample did not exercise over the 12 mo postdischarge. This rate cannot be compared with those found in other studies because of the lack of longitudinal data collected on exercise behaviors post-SCI. According to Kaźmierczak and Lisiński (2018), the first 3 yr after SCI is the time when physical activity participation routines are developed.
We found a statically significant negative relationship between monthly physical exercise and hospital readmissions for the entire sample, but this relationship was insignificant when each neurologic injury category was analyzed individually. Insignificant outcomes could be explained by the sensitivity of χ2 analysis to larger sample sizes compared with smaller sample sizes (Hung et al., 2017). Despite the significance of the exercise and hospital readmission relationship for the entire sample population, the φ correlation coefficient indicated a weak relationship between exercise and hospital readmission. A weak φ correlation coefficient could be a result of several factors. Hospital readmission rates may be only partially dependent on regular exercise, and other confounding variables may explain a more accurate picture. Confounding variables, such as independence level after injury, severity of injury, or comorbidities, could explain why participants could not exercise while also explaining a higher hospital readmission rate in this group (Johnson & Christensen, 2020).
Epidemiological and clinical studies of exercise post-SCI have several inconsistencies. The first is the inconsistent and vague definitions of physical activity. Each participant has a unique perception of what exercise is, and it likely varies among participants, which therefore makes it challenging to develop inferences from the data and results. The inclusive definition of exercise in the SCIRehab study provided the ability to collect data that included many different forms of exercise, such as electrical stimulation, passive range of motion, daily activities that cause exertion, or other lower level exercise forms that might not otherwise have been included. This provides the opportunity for people with higher and more severe SCI to document participation in exercise. Although this definition is inclusive, it makes it challenging to isolate benefits because each form of exercise (strength training, cardio, stretching, passive movement, electrical stimulation) can produce different health outcomes (Lai et al., 2017).
The second inconsistency is the lack of current longitudinal studies of participation in exercise after SCI. The NSCISC is the largest longitudinal database on SCI in the world; data are collected from each participant every 5 yr, but (outside of formal therapy) no exercise data are collected (NSCISC, 2019). The SCIRehab study, now more than 10 yr old, collected data from 2007 to 2010 and was the largest and most recent longitudinal public data set that collected data on physical activity after SCI (Whiteneck, 2018).
Although studies have shown a range of both mental and physical health–related benefits of exercise after an SCI, the type of exercise, duration, intensity, and outcome measures used lack consistency. Recently, the NIH identified the need for physical activity guidelines for people with disabilities, including people with SCI, who make up a large part of this population. The NIH reviewed the current body of evidence of the benefits and risks of physical activity as part of an interdisciplinary (including occupational therapy) collaborative workshop in December 2020 and determined that there is not yet enough knowledge or evidence to establish physical activity guidelines for this population. The NIH is recommending future randomized controlled trials, observational studies, and large multisite collaborations to further knowledge of the risks and potential benefits of exercise in this population with the goal of establishing future exercise guidelines (NIH, 2021).
These limitations in the current body of evidence support the NIH’s 2021 request for larger multicenter collaborative longitudinal cohort studies and randomized controlled trials of exercise behaviors among wheelchair users to help guide the development of physical activity guidelines for this population.
Limitations
This secondary analysis of data from the SCIRehab Project has several limitations. The original data collection included self-report measures that were subject to social desirability bias (i.e., reporting more desirable outcomes than what is accurate). Recalling and estimating frequency of exercise is also subject to recall errors. Another limitation of the study is the term exercise. There was no description of what type of exercise each participant engaged in as part of the original data collection. Type and amount of health-related benefit from exercise is dependent on the type of exercise, which would have been helpful information in this current study.
Last, there may have been a difference between the cohort who reported exercising monthly and the cohort who reported no exercise since discharge. The cohort who engaged in exercise regularly may have been healthier or more independent at discharge from inpatient rehabilitation than the cohort who did not exercise. The cohort who did not exercise may have experienced more medical complications and comorbidities and lower levels of independence that prevented them from exercising but that also made them more vulnerable to hospital readmissions because of their baseline limitations. These factors were not taken into consideration in this study.
Future Research Recommendations
This was a nonexperimental, secondary analysis that investigated correlational relationships between exercise frequency and hospital readmission among data on SCI participants from a national database. These results can help guide future research in determining whether a cause-and-effect relationship exists between exercise and rehospitalization and exploring the health-related outcomes of different exercise frequencies. A regression analysis of multiple variables (independence level, comorbidities, type of exercise, etc.) would assist in identifying which variables most accurately predict higher hospital readmissions in this population.
Implications for Occupational Therapy Practice
Exercise post-SCI is an area in which occupational therapists can flourish and be a valuable resource for people with SCI. Opportunities range from providing inpatient, outpatient, home, and community interventions to addressing the many different exercise barriers people experience after SCI. Example interventions include exploring different exercise opportunities in a client’s community and identifying a meaningful form of exercise, grading and adapting exercises to the client’s specific performance skills, and promoting long-term adherence. Assisting clients with navigating environmental barriers in the gym or exercise location or with equipment and educating them on available adapted exercise equipment (e.g., modified hand grips to hold a barbell) can promote self-efficacy and exercise independence. Collaborating with community health centers to establish exercise opportunities for people with SCI and other conditions can provide for greater inclusion for and support of diverse community member needs. Further opportunities include advocating and working with fitness establishments to improve accessibility of gym environments and gym equipment.
This study has the following implications for occupational therapy practice: Almost one-third (29.8%) of the people included in this study experienced at least one hospital readmission in the first year after their SCI. Exercise occurs as part of the rehabilitation process after an SCI; however, after discharge from inpatient rehabilitation, 229 of the 520 participants in this study (44%) reported no exercise. There is a need to promote self-efficacy and independence in a meaningful and sustainable exercise routine after discharge from rehabilitation services in the SCI population. Occupational therapists are knowledgeable about disability and equipped to grade, adapt, and promote health and independence in meaningful activities that can assist with filling this gap. Establishing community exercise opportunities for this population is an area of opportunity for occupational therapists and future researchers to explore.
Conclusion
Exercise is a meaningful occupation that provides many physical and mental health–related benefits, including for people with SCI. However, after a person experiences a traumatic SCI, their participation in exercise becomes minimal or often nonexistent, risking occupational deprivation and health-related consequences, and it could contribute to the high hospital readmissions in the first year after injury. Limited exercise participation can be explained in part by the many barriers experienced after SCI, including environmental, performance skills, knowledge, and self-confidence. Occupational therapists have the training to and are capable of addressing each of these barriers, whether in an inpatient facility, clinic, home, or community, and they can promote and increase participation in exercise after SCI.