Abstract
BACKGROUND:
Little is known about predictors of activity participation that are objectively measured in cardiac rehabilitation (CR) graduates. This prospective observational study aimed to determine the predictors of objectively measured activity participation among adults with chronic cardiac conditions who have completed Phase II CR.
METHODS:
Twenty-five adults with chronic cardiac conditions graduating from traditional CR program participated in the study. The outcome variable was an activity participation level measured by light-to-vigorous intensity physical activity (LVPA) minutes using ActiGraph GT9X Link accelerometer after CR discharge. Covariates were collected at the discharge from CR, and outcome variables were collected at 1 month, 3 months, and 9 months post CR discharge.
RESULTS:
Gender, standardized LVPA at CR discharge, body mass index, and motivation for physical activity and leisure were significantly associated with the activity participation levels at 1 month, 3 months, and/or 9 months post CR discharge. Gender, standardized LVPA at CR discharge, highest education completed, and depressive symptoms significantly predicted the activity participation levels at 1 month (R2 = 0.69, p < 0.001), 3 months (R2 = 0.65, p < 0.001), and/or 9 months (R2 = 0.80, p < 0.001) post CR discharge. Female CR participants who were more active, had more than high school education, and showed more depressive symptoms at CR discharge were more likely to be active post CR.
CONCLUSIONS:
CR participants may benefit from individualized approach to plan their days post CR and application-focused and education-level sensitive sessions to understand the importance of activity participation maintenance post CR.
Keywords
Introduction
Participation in cardiac rehabilitation (CR) is effective in improving short-term outcomes including physical activity (PA), exercise capacity, and symptoms of depression and anxiety, and long-term outcomes including cardiovascular risks, hospital admission, and health-related quality of life [1–4]. Mixed results have been reported in the maintenance of CR benefits after 10 days to 1 year from CR discharge [5–8]. It was suggested that more focus on light intensity PA and sedentary behaviors during CR may improve the maintenance post CR [5–7]. Higher PA levels, both light-intensity (1.5 to 3.0 metabolic equivalent [MET]) and moderate-to-vigorous intensity (> 3.0 MET) PA, have been consistently associated with positive health outcomes including prevention of cardiovascular diseases and reducing all-cause mortality in those with cardiac conditions [9–12]. Therefore, identifying CR participants at high risk of reducing PA levels after CR discharge and providing necessary interventions to these individuals during CR can be beneficial for long-term maintenance of PA levels.
To reduce cardiovascular risks and promote hea-lthy behaviors, traditional CR takes a comprehensive approach using the core components of promoting nutritional and PA counseling in addition to the exercise training [13]. PA counseling primarily focuses on providing recommendations to sustain levels of moderate-to-vigorous intensity PA only. In addition, predictors of PA for graduates of Phase II CR have been studied mostly for moderate-to-vigorous intensity PA using subjective outcome measures (e.g. PA questionnaires) and rarely studied for the full intensity levels of PA using an objective measure, such as activity tracking. In women after CR, comorbidity, benefits and barriers of PA, and social support for persistence were identified as predictors of exercise measured using a heart rate monitor [14]. In addition, in post CR adults, the length of CR participation, self-efficacy in barriers, task, and relapse, and social assimilation were identified as the predictors of self-reported exercise levels [15, 16]. However, at the present time, little data exist on predictors of comprehensive PA, or activity participation, that are objectively measured in graduates of Phase II CR. Therefore, the purpose of this study was to determine the predictors of objectively measured activity participation time among graduates of Phase II CR.
Methods
Design
We used a prospective observational design to identify the predictors of activity participation at three different times post CR. Covariates were collected at the discharge from CR, and outcome variables were collected at 1 month, 3 months, and 9 months post CR discharge.
Participants
Twenty-five adults with chronic cardiac conditions who had received CR at a local cardiovascular and pulmonary center in rural Southeastern State of US participated in the study. The inclusion criteria were [1] being scheduled for discharge from the CR program within 3 weeks, [2] no plan to attend a Phase III maintenance program at the cardiovascular and pulmonary center, [3] being willing to wear an activity monitor for 7–15 days, [4] living within 60 miles of the research center, and [5] having functional English fluency. Anyone who did not have chronic cardiac conditions, such as valve replacement without coronary artery disease, was excluded.
Outcome variable
The primary outcome variable was the participation in any types of light and moderate-to-vigorous intensity PAs that are non-sedentary. PA was measured using an ActiGraph GT9X Link accelerometer that is triaxial and equipped with a secondary acce-lerometer and gyroscope and magnetometer sensors to detect the position and rotation data. Participants wore the ActiGraph on their non-dominant wrist for 24 hours for 15 consecutive days, and only the date and time were displayed on the accelerometer screen. The non-dominant wrist as a measurement placement was shown to increase the wear compliance [17, 18]. The data were downloaded using Actilife v6 13.4 software [19]. Wear time validation was completed using the Choi (2011) algorithm [20] to exclude any days during which participants wore the ActiGraph less than 10 hours a day. Minutes for sedentary behavior, light-intensity PA, and moderate-to-vigorous intensity PA were downloaded using the Freedson Adult (1998) algorithm [21]. Finally, due to the varied wearing times, combined standardized minutes of sedentary behavior, light-intensity PA, and moderate-to-vigorous intensity PA for 1020 minutes (17 hours) were calculated. The standardized minutes of light-to-vigorous intensity PA (LVPA) at 1 month, 3 months, and 9 months from the CR discharge were used as an outcome variable. Data for 5 or more consecutive days were available for all participants.
Covariates – Demographic and health variables
The demographic variables were collected in-person at CR discharge: age, race, gender, highest education completed, marital status (Y/N), employment status (Y/N), and income level. The following health variables were collected through medical record review at CR discharge: body mass index (BMI) and numbers of cardiovascular diseases, cardiovascular risk factors, such as smoking history and diabetes [22], major diagnosis, and CR sessions completed. Although not included in the main analyses, information on the history of cardiovascular diseases and interventions received for cardiovascular disease during the past 2 years was collected through medical record review at CR discharge.
Covariates – Variables on body functions and activity and participation
The Center for Epidemiologic Studies Depression Scale (CES-D) measures depressive symptoms during the past 1 week using 20 self-reported questions. With 0–3 frequency answer options, the total score ranges 0–60. A higher score indicates more depressive symptoms, and a score of 16 or greater indicates depressive symptoms. CES-D has been validated in US adults with heart disease [23].
The Physical Activity and Leisure Motivation Scale (PALMS) measures motivation toward participation in leisure-time physical activity. Each item is self-rated using a 1–5 agreement rating scale, and a higher score indicates greater PA motivation. In our study, two combined total scores were calculated to represent the PA motivation related to self and PA motivation related to others. The PA motivation score related to self ranged 25–125 and included the subscales of mastery, physical condition, appearance, enjoyment, and psychological condition. The PA motivation related to others ranged 15–75 and included the subscales of affiliation, others’ expectations, and competition. PALMS has good reliability and criterion validity among community dwelling adults [24].
The Keitel Functional Test (KFT) measures body, spinal, and extremity movement with 24 items. In our study, one item on walking in a hallway and two items on walking 10 steps upstairs and downstairs were used, with a total score of 0–12. A higher score indicates slower performance and moredifficulty. KFT has acceptable psychometric properties in people with rheumatoid arthritis [25].
The Patient-Reported Outcomes Measurement Information System Fatigue measures self-reported fatigue intensity and frequency using a computer adaptive test. Individuals typically answer three to five questions generated by its algorithm and based on their experience during the past 7 days. An average standardized score of the general population is 50, and a higher score indicates worse fatigue. It has clinical validity in individuals with heart failure and excellent reliability in the general population [26, 27].
The 6-Minute Walk Test measures functional exercise capacity by measuring the total distance an individual can walk in 6 minutes [28]. A hallway that is 100 feet long and 7 feet wide was used in our study. It has excellent reliability and moderate-to-high convergent validity with relevant physiological factors in CR participants [29].
The Activity Measure for Post-Acute Care Outpatient Short Form measures self-reported activity limitations in basic mobility, daily activity, and applied cognitive activity using 18, 15, and 19 items, respectively. Individuals are asked to rate how much difficulty they have completing each item on 1–4 scale [30]. A T-scale score of 50 was the average of general rehabilitation population with a standard deviation of 10. A higher T-scale score indicates less activity limitation. It has good reliability and good construct validity in mixed, general rehabilitation population [31, 32].
The Participation Objective Participation Subjective (POPS) measures objective and subjective aspects of participation in complex activities using 26 items. We used the objective information of the participation frequency per week on seven social participation items: socializing with friends, relatives, and school mates/co-workers, going out to social events and occasions, going to places where they might meet new people, speaking with neighbors, and speaking with strangers. A high score indicates more frequent social participation. POPS has good reliability and convergent and discriminant validity in individuals with traumatic brain injuries [33].
Lastly, LVPA were measured by ActiGraph GT9X Link, and the standardized LVPA minutes at the CR discharge were used as a covariate.
Procedure
The research procedures were in accordance with the Helsinki Declaration of 1975, as revised in 1983. Participants were recruited from two local cardiovascular and pulmonary rehabilitation centers. After receiving an approval from the Institutional Review Board (UMCIRB 17-001329), research staff completed the eligibility screening over the phone and visited the center to obtain signed consent form. Participants were assessed at the research center at their CR discharge and 1 months and 3 months post CR discharge. Participants were assessed over the phone or in-person at 9 months post CR discharge. Participants were provided with the ActiGraph to wear 7–15 full days prior to each assessment date. Screeners and assessors were an occupational therapist or occupational therapy graduate students who were trained and showed competency in all assessments prior to the independent administration.
Statistical analyses
All variables were calculated as means with standard deviations for continuous variables and frequency with percentages for categorical variables. According to visual examinations using boxplots and normal QQ-plots, all outcome variables showed no strong deviation from normality. Bivariate associations with the outcome variables were assessed using one-way ANOVA or two-sample t-tests when the covariates were categorical or Pearson’s correlations when the covariates were continuous. Finally, general linear models were used to study the adjusted effects of the outcome variables. A forward model selection method was utilized, and covariates with P values no greater than 0.05 were kept. However, covariates that were selected in a previous time point were automatically kept for a later time point.
Results
The mean (SD) age was 64.6 (10.5) years, and the majority were White, male, and married. On average, participants had three cardiovascular diseases and completed 29 CR sessions. At CR discharge, participants spent 382.7 minutes out of 1020 minutes on average on LVPA during waking hours (Tables 1and 2).
Demographic and health information
Demographic and health information
Body function and activity and participation variable information
CES-D, center for epidemiologic studies depression scale; PALMS, physical activity and leisure motivation scale; KFT, keitel functional test; PROMIS, patient-reported outcomes measurement information system; 6MWT, 6-minute walk test; AM-PAC, activity measure for post-acute care outpatient short form; POPS, participation objective participation subjective; LVPA, light to vigorous physical activity; CR, cardiac rehabilitation.
Gender and the standardized LVPA minutes at CR discharge showed significant bivariate associations with the standardized LVPA minutes at 1 month (p = 0.03 and p < 0.001), 3 months (p = 0.02and p = 0.002), and 9 months (p = 0.009 and p = 0.001) post CR discharge. BMI showed significant negative bivariate association with the standardized LVPA minutes at 9 months post CR discharge (p = 0.046). Lastly, PALMS related to others showed significant negative bivariate associations with the standardized LVPA minutes at 1 month (p = 0.044), and 9 months (p = 0.022) post CR discharge (Table 3).
Results of bivariate associations between outcome variables and covariates
at. bF. *p < 0.05. LVPA, light to vigorous physical activity; CES-D, center for epidemiologic studies depression scale; PALMS, physical activity and leisure motivation scale; KFT, keitel functional test; PROMIS, patient-reported outcomes measurement information system; 6MWT, 6-minute walk test; AM-PAC, activity measure for post-acute care outpatient short form; POPS, participation objective participation subjective.
For the standardized LVPA minutes at 1 month post CR discharge, both gender (p = 0.012) and the standardized LVPA minutes at CR discharge (p < 0.001) showed significant effects in the general linear model. Males had on average 70 minutes lower LVPA at 1 month than females. One more minute of LVPA at CR discharge translated to on average 0.9 more minute of LVPA at 1 month.
For the standardized LVPA minutes at 3 months post CR discharge, gender (p = 0.009), thestandardized LVPA minutes at CR discharge (p < 0.001), and highest education (p = 0.035) showed significant effects in the general linear model. Males had on average 71 minutes lower LVPA at 3 months than females. One more minute of LVPA at CR discharge translated to on average 0.74 more minute of LVPA at 3 months. People with master’s or higher education and some college to bachelor education had respectively on average 31 and 92 more LVPA minutes at 3 months than those with up to high school education.
For the standardized LVPA minutes at 9 months post CR, gender (p = 0.16), the standardized LVPA minutes at CR discharge (p < 0.001), highest education (p = 0.02), and CES-D (p = 0.006) were kept in the general linear model. Males had on average 33 minutes lower LVPA at 9 months than females. One more LVPA minute at CR discharge translated to on average 0.96 more LVPA minute at 9 months. People with master’s or higher education and some college to bachelor education had respectively on average 17 and 79 more LVPA minutes at 9 months than those with up to high school education. For one unit ofCES-D score increase, the standardized LVPA minutes at 9 months was expected to increase by 5.3 minutes (Table 4).
Results of the general linear models between outcome variables and covariates
*p < 0.05. LVPA, light to vigorous physical activity; CR, cardiac rehabilitation; CES-D, center for epidemiologic studies depression scale.
This study investigated the predictors of objectively measured activity participation at 1 month,3 months, and 9 months post Phase II CR. We found that gender, activity participation time at CR discharge, highest education completed, and depressive symptoms at CR discharge significantly predicted objectively measured activity participation after CR discharge. According to our findings, female CR participants who were more active, had more than high school education, and showed more depressive symptoms at CR discharge are more likely to be active post CR compared to those who are not.
We found that gender was one of the significant predictors of activity participation through 3 months post CR, and women showed higher activity participation than men. There has been general consensus that women are less active in moderate-to-vigorous intensity PA than men in the US population [34]. However, older women were more active in light-intensity PA than older men [35], and there have been no consistent findings on the gender difference in sedentary time among older adults [36]. Therefore, our study offers a novel finding on the difference in activity participation between genders post CR. Our outcome measure focused on the comprehensive levels of PA ranging from light-intensity PA to vigorous-intensity PA, effectively measuring the opposite of the sedentary time. Considering there are positive health outcomes for all intensity levels of PA [9, 37] and female CR graduates are more active in comprehensive PA levels, individually brainstorming and planning ways to accumulate physically active tasks (e.g. pet care, home management, social outings) at the end of CR may be beneficial.
We found that the activity participation level atCR discharge significantly predicts activity participation for 9 months after the completion of CR. There have been mixed reports on the changes in activity participation level after CR completion. In one study, increase in sedentary time, decrease in moderate-to-vigorous intensity PA, and no change in light-intensity PA were found among CR graduates with angina or myocardial infarction [6], whereas individuals with acute coronary syndrome became more active following CR [8]. However, it is notsurprising that the activity participation at CR discharge can predict the post CR activity participation levels based on the relatively small increases or decreases in activity participation levels identified post CR in previous studies [6, 8]. Therefore, the gains of LVPA during CR may hold a significance in deciding CR graduates’ activity participation level following CR completion.
According to our findings, the education level predicts the activity participation level among CR graduates. Specifically, individuals with up to high school education showed almost 1.5 hour less activity participation than those with some or full college education. This result was not consistent with the previous findings that failed to identify the education level as one of the interpersonal correlates of PA during CR or post CR [38]. However, it is important to note that previous findings focused on exercise, whereas our finding focuses on the wider spectrum of activity participation, or LVPA. Because CR participants are more likely to have higher education level than non-CR participants [39] and CR graduates with more than high school education show higher activity participation level, the education level may be one of the important factors to consider during CR and discharge planning. This may include more assistance and time to help CR participants understand the information provided during CR sessions using examples and applications, rather than principles or textbook knowledge.
Lastly, depressive symptoms at CR discharge predict the activity participation level post CR among people with chronic cardiac conditions. Specifically, more depressive symptoms at CR discharge were associated with higher activity participation at 9 months post CR. This was a surprising finding because the presence of depressive symptoms is frequently associated with sedentary behaviors in people with cardiovascular diseases [40, 41]. However, it is important to note that our participants were CR graduates who successfully completed an average of 29 CR sessions when only over a half of the CR national cohort completed 25 or more sessions [42]. In addition, individuals with depression are 4 times more likely to attend CR than those without depression, and CR was effective in reducing depressive symptoms and mortality [43, 44]. Also, our participants maintained their activity participation level for 9 months after CR discharge. Lastly, only 4 out of 12 participants who were showing depressive symptoms based on the CES-D cutoff of 16 had a diagnosis of depression at CR discharge. Therefore, it is possible individuals with more depressive symptoms, not with a diagnosis of depression, at CR discharge may be better aware of the importance of being active and, hence, continue their activity participation post CR. This finding signifies the importance of education on maintenance of activity participation for both CR participants with and without depressive symptoms.
Despite the strengths of objective outcome measurements capturing comprehensive levels of PA for 9 months post CR, our study had some limitations. Although participants verbally confirmed that they did not go back to the CR center for the maintenance program during each assessment session, this was not confirmed by each CR center. However, this was very unlikely due to the limited number of CR centers in the area. In addition, we had a small sample size, and this may have caused low power to detect statistical significance and have decreased the generalizability of our findings to general CR graduates. Future studies can collect information on any structured PA programs CR graduates participated to eliminate any potential impact of such program participation. Also, future studies can recruit a statistically-powered number of participants to improve the generalizability of the findings.
Gender, activity participation time at CR discharge, highest education completed, and depressive symptoms at CR discharge can predict activity participation time after CR discharge among individuals with chronic cardiac conditions. CR participants may benefit from individualized approach to plan their post CR days and application-focused and education-level sensitive sessions to understand the importance of activity participation maintenance post CR.
Footnotes
Acknowledgment
This study was supported by East Carolina University Start-up Fund. The authors would like to thank clinicians in Vidant Cardiovascular and Pulmonary Rehabilitation in Greenville and Washington, NC for referring participants.
Conflict of interest
The authors have no conflict of interest to report.