Examining the Relationship Between Physical Activity and Self-Efficacy for Exercise Among Overweight and Obese Marshallese Adults

Abstract

INTRODUCTION

This study examined associations between self-efficacy for exercise (SEE) and physical activity frequency among overweight and obese Marshallese adults in the US.

METHODS

Physical activity frequency was assessed using two items adapted from the Physical Activity Questionnaire. SEE was assessed using the nine-item scale. Logistic regression and ordinal logistic regressions were employed for the analysis of total physical activity and moderate and vigorous physical activity.

RESULTS

SEE was significantly associated with increased odds of engaging in overall sufficient physical activity and higher odds of engaging in both moderate and vigorous physical activity. Unemployed participants reported lower odds of moderate physical activity. Younger male participants who reported being in excellent or good health had higher odds of participating in vigorous physical activity.

DISCUSSION

Physical activity is a modifiable lifestyle behavior associated with many of the chronic disease disparities faced by the Marshallese community and the broader NHPI population.

BACKGROUND

Native Hawaiian and Pacific Islander (NHPI) community members experience significant health disparities compared with the general population, with higher rates of obesity, cardiovascular disease, and type 2 diabetes [1–8]. Marshallese are an NHPI population from the Micronesian nation the Republic of the Marshall Island. Studies have documented high rates of obesity and type 2 diabetes specifically among Marshallese adults living in the United States [8]. In a study of 401 Marshallese adults in Arkansas, 28.0% were overweight, 61.7% were obese, 32.6% had pre-diabetes, and 38.4% had type 2 diabetes [7].

A lack of sufficient physical activity can contribute to negative health outcomes including all-cause mortality, obesity, coronary heart disease, hypertension, and type 2 diabetes [9]. However, only 26% of men and 19% of women in the US reported physical activity levels that met the Physical Activity Guidelines for Americans recommendations [9]. Similarly, the majority of Native Hawaiian and Pacific Islanders (NHPI) do not meet physical activity recommendations. In a sample of 100 NHPI participants living in California, only 18.0% met the physical activity guidelines for moderate physical activity [10]. A meta-analysis of physical activity studies among NHPI concluded that less than half of NHPI adults met physical activity guidelines for moderate physical activity, and only one-third met the guidelines for vigorous physical activity [11]. In a study that assessed eight diabetes-related self-care behaviors among a sample of Marshallese adults with type 2 diabetes (n=111), engaging in recommended levels of physical activity was the second-least practiced self-care behavior [12]. Only 33.3% of participants reported engaging in any kind of physical activity (moderate or vigorous) four or more times per week [12]. Another study of Marshallese adults (n=376) found 28.7% reported engaging in physical activity zero times per week, and only 26.9% reported engaging in physical activity four or more times per week [13].

Self-efficacy refers to an individual’s confidence or belief that they can carry out a task and achieve a desired outcome [14]. Studies have shown self-efficacy for exercise (SEE) to be an important predictor of intention to engage in physical activity, while others found SEE to act as a mediator for change in physical activity behaviors [15–19]. One study examined changes in both SEE and physical activity among NHPI residing in Hawaii [20]; however, this study did not assess the relationship between the two.

Despite the importance of SEE observed in other populations, no prior studies to our knowledge have examined the relationship between SEE and physical activity among NHPI living in the continental US. This study examines the relationship between SEE and physical activity among a large sample of overweight and obese Marshallese Pacific Islander adults living in the continental US.

METHODS

This study used cross-sectional baseline data from a faith-based Diabetes Prevention Program (DPP) randomized trial (clinicaltrials.gov ID: NCT03270436). The DPP trial is one of many research activities conducted by the academic-community partnership in northwest Arkansas [21–23] that uses a community-based participatory research (CBPR) approach [24–26].

A CBPR approach was utilized to engage Marshallese community members in the design and execution of a DPP interventions. The curricula focused on reducing weight, eating healthy, and increasing physical activity. The study was conducted in 30 Marshallese-serving churches in Arkansas and Oklahoma. Churches were chosen by the community as the most appropriate setting for the intervention. In order to reduce issues of contamination, churches served as the unit of randomization. All educational sessions were conducted in the churches in a group setting. Full details of the study design, recruitment process, and engagement process have been published elsewhere [27]. All study protocols and procedures were approved by the University of Arkansas for Medical Sciences Institutional Review Board (IRB#207034).

Data Collection

The study sample comprised 378 Marshallese adults aged 18 years and older with a body mass index (BMI) ≥25. Study recruitment began in October 2017 and continued until recruitment goals were met in May 2019. Potential participants were recruited from community locations by bilingual Marshallese study staff. Study staff described the study in detail and answered questions about the study prior to obtaining informed consent.

Measures

Physical activity frequency

Physical activity frequency was assessed using two items adapted from the brief self-report Physical Activity Questionnaire [20, 28]. The adaptation incorporated culturally relevant examples of physical activity (e.g., gardening, playing softball, Zumba) based on the feedback from local Marshallese stakeholders. Participants were asked to self-report how often they engaged in moderate physical activity and vigorous physical activity using the following two questions: 1) “For the past month, about how often have you taken part in moderate physical activity (mowing your lawn, gardening, taking long walks)?” and 2) “For the past month, about how often have you taken part in vigorous physical activity (such as jogging, playing basketball or softball, Zumba)?” Both questions used a 4-point response scale: 1) Rarely or Never; 2) Once a week; 3) 2–4 times a week; and 4) More than 4 times a week.

A total physical activity variable was created based on methods adapted from a study assessing the psychometric properties of the scales [28]. Each 4-point scale for moderate and vigorous physical activity was weighted: 0=Rarely or Never; 1=Once a week; 2=2–4 times a week; and 4=More than 4 times a week. The weights for each participants’ response to the moderate and vigorous physical activity variables were then summed and dichotomized as follows: ≥4 = sufficient physical activity and <4 = insufficient physical activity.

SEE was assessed using the nine-item scale [14, 29]. The response scale was culturally adapted based on Marshallese stakeholder requests to decrease the number of response options from the original 11 (0, Not Confident – 10, Very Confident) to three and change the response options to the following: 1) No/Not Sure at All; 2) Maybe/Not Sure; and 3) Yes, Completely Sure. This approach is similar to previous adaptations used to measure SEE among other NHPI populations for whom English is not their first language and/or for which there are no equivalent concept (i.e., confidence in degrees) in their language or worldview [20]. Internal reliability-consistency was 0.84, well within the range of acceptable standards [30]. All but three items had item-total correlation coefficients greater than the recommended 0.50 [31], ranging from 0.36 to 0.65. Each item was prefaced with the following phrase: “I can exercise at least 60 minutes in total each week even if…” 1) The weather was bothering me; 2) I was bored by the exercise program or activity; 3) I felt discomfort when exercising; 4) I had to exercise alone; 5) I did not enjoy it; 6) I was too busy with other activities; 7) I felt tired; 8) I felt stressed; and 9) I felt depressed. A total SEE scale score was calculated across the nine items (possible range=9–27) and the sum was divided by the number of items, resulting in an overall mean SEE score for each participant (possible range=1.0–3.0), with higher scores indicating higher levels of self-efficacy.

Covariates

Age

Age was calculated based on self-reported date of birth and survey completion date. All values were rounded down to the nearest whole number (i.e., a person with a calculated age of 37.8 would be considered 37 years old). Sex was reported as: 1) Male or 2) Female. Marital status was captured with a single item. Response options included: 1) Single; 2) Married; 3) Divorced/Separated; 4) Widowed; and 5) A member of an unmarried couple. For purposes of analysis, the number of categories was reduced from five to two: 1) Married and 2) Not married (previously 1, 3, 4, and 5) due to small numbers of responses in some categories. Employment status asked participants to select the option that best described their current employment status. Response categories included: 1) Employed for wages (35+ hours a week); 2) Employed for wages part time (34 hours per week or less); 3) Self-employed; 4) Out of work for 1 year or more; 5) Out of work for less than 1 year; 6) Taking care of your family and home; 7) A student; 8) Retired; and 9) Unable to work. For analysis purposes, the number of categories was reduced from nine to two: 1) Employed (previously 1, 2, and 3); 2) Unemployed (previously 4, 5, 6, 7, 8, and 9) due to small numbers of responses in some categories.

Education level

Education level was assessed with a single item regarding highest grade of school completed. Response options included: 1) Never attended school or only attended kindergarten; 2) Grades 1 through 8 (Elementary); 3) Grades 9 through 11 (Some high school); 4) Grade 12 or GED (High school graduate); 5) College 1 year to 3 years (Some college or technical school); and 6) College 4 years or more (College graduate). For analysis purposes, the number of categories was reduced from six to three: 1) Less than high school diploma (previously 1, 2, and 3); 2) High school graduate; and 3) Education beyond high school (previously 5 and 6) due to small numbers of responses in some categories.

Number of comorbidities

Number of comorbidities was a sum of affirmative responses to four questions asking participants if they had ever been diagnosed by a healthcare professional with the following chronic conditions: cancer; diabetes; hypertension; and heart disease.

BMI

BMI was measured using participants’ height and weight ((weight in pounds/[height in inches]²)*703). Weight, with clothing but without shoes, was measured to the nearest 0.5 pound using a calibrated digital scale. Height, without shoes, was measured to the nearest 0.25 inch using a stadiometer.

General health status

General health status was assessed based on the participants’ self-reported perception of their current health. A four-point scale was used: 1) Excellent; 2) Good; 3) Fair; and 4) Poor. Due to the extremely low number of responses of “Poor”, this scale was reduced to three categories, combining “Fair” and “Poor” into a single category re-termed Fair/Poor.

Analysis

Data were analyzed with SAS/STAT 14.2 [32]. Descriptive statistics including means with standard deviations, frequencies, and proportions were calculated for demographic, self-efficacy and physical activity variables, conditional on the measurement scale. The association between first dependent variable, total physical activity, and SEE was assessed using multivariable logistic regression. The association between the second and third dependent variables (moderate physical activity and vigorous physical activity) were assessed with ordinal logistic regression [33]. In each instance, the assumption of proportional odds was tested. Covariates included for adjustment were sex, age, marital status, employment status, education level, total number of comorbidities, BMI, and general health status. The results were considered significant at the alpha level of 0.05.

RESULTS

As shown in Table 1, the average age of participants was 42.1 years (SD=11.6), over half were female (56.6%), and half had a high school education or greater (51.3%). The majority were married (79.6%), more than half were employed (54.5%), and most reported being in good or excellent health (79.8%).

Table 1.

Descriptive statistics for socio-demographic and health-related explanatory variables

Measures	n(%) or Mean±SD
Age (years)	42.1±11.6
Sex
Male	164 (43.4)
Female	214 (56.6)
Marital Status a
Single	45 (11.9)
Married	301 (79.6)
Divorced/Separated	3 (0.8)
Widowed	17 (4.5)
Member of unmarried couple	12 (3.2)
Education Level
Never attended school	1 (0.3)
Elementary school	54 (14.3)
Some high school	129 (34.1)
High school graduate	132 (34.9)
Some college or technical school	57 (15.1)
College graduate	5 (1.3)
Employment Status b
Employed – Full Time	166 (43.9)
Employed – Part Time	23 (6.1)
Self-employed	17 (4.5)
Out of work for 1 year or more	59 (15.6)
Out of work for less than 1 year	50 (13.2)
Taking care of family and home	42 (11.1)
Student	6 (1.6)
Retired	7 (1.9)
Unable to work	8 (2.1)
Comorbidities c	1.5±0.7
Body Mass Index	33.7±5.4
General Health Status
Excellent	49 (13.0)
Good	251 (66.8)
Fair	73 (19.4)
Poor	3 (0.8)

Notes: Percentages may not total 100 due to rounding. SD=standard deviation.

^aSingle includes those who indicated they were currently single, divorced, widowed, and member of an unmarried couple. Married includes only those who indicated they were currently married.

^bEmployed includes those who indicated they were currently working full time, part time, and who were self-employed. Unemployed includes those who were out of work for less than 1 year, out of work for 1 year or more, disabled, students, retired, or taking care of their family member or home.

^cComorbities include self-reported previous diagnoses of cancer, diabetes, heart disease, or hypertension.

Table 2 provides a summary of responses to the physical activity and SEE items. Of the 378 participants enrolled in the study, 375 responded to all nine SEE items. The mean SEE score was 2.5 (SD=0.48) out of a possible 3.0 (observed range=1.0–3.0). Total physical activity was evenly split, with approximately half of participants engaging in insufficient weekly physical activity (49.9%) and half engaging in sufficient weekly physical activity (50.1%).

Table 2.

Descriptive statistics for self-efficacy for exercise and self-reported engagement in moderate and vigorous physical activity

Measures	N(%) or Mean±SD
Self-Efficacy for Exercise (SEE)	2.5±0.48
SEE Itemsa “I can exercise at least 60 min. each week even if…”
I had to exercise alone	333 (88.6)
I was bored by the exercise program or activity	288 (76.2)
The weather was bothering me	255 (67.5)
I did not enjoy it	250 (66.1)
I felt stressed	239 (63.2)
I felt discomfort when exercising	223 (59.0)
I felt tired	221 (58.6)
I was too busy with other activities	220 (58.2)
I felt depressed	204 (54.0)
Total Physical Activity b
Insufficient	189 (49.9)
Sufficient	190 (50.1)
Moderate Physical Activity c
More than 4 times a week	143 (37.8)
2–4 times a week	101 (26.7)
About once a week	50 (13.2)
Rarely or never	84 (22.2)
Vigorous Physical Activity d
More than 4 times a week	41 (10.9)
2–4 times a week	98 (25.9)
About once a week	37 (9.8)
Rarely or never	202 (53.4)

Notes: Percentages may not total 100 due to rounding.

^aNumber and percent responding “Yes, completely sure” to each item.

^bTotal Physical Activity score was created by combining and weighting each participants’ responses to the moderate and vigorous physical activity items.

^cModerate Physical Activity was assessed with a single self-report item.

^dVigorous Physical Activity was assessed with a single self-report item.

More than one-third (37.8%) of participants reported engaging in moderate physical activity 4 times a week or more, but only 10.9% of participants reported engaging in vigorous physical activity 4 times a week or more. Approximately one-third (35.4%) of participants reported engaging in moderate physical activity one time a week or less, and almost two-thirds (63.2%) reported engaging in vigorous physical activity one time per week or less. Furthermore, one-fifth (20.1%) of participants reported rarely or never engaging in either moderate or vigorous physical activity.

Total Physical Activity

As shown in Table 3, after controlling for the eight covariates, for every 1 unit increase in SEE, the odds of engaging in sufficient physical activity increased by over one and a half times (odds ratio [OR]=1.70, 95% confidence intervals [CI]: 1.07–2.69, p=0.024). Additionally, the odds of engaging in sufficient physical activity for unemployed participants was less than two-thirds that of participants who were employed (OR=0.60, CI: 0.38–0.94, p=0.026). None of the remaining covariates were significantly associated with frequency of engaging in sufficient weekly physical activity.

Table 3.

Effect of Self-Efficacy for Exercise on Self-Reported Levels of Total, Moderate, and Vigorous Physical Activity: Results of Logistic and Ordinal Logistic Regression Analyses

	Total PA		Moderate PA		Vigorous PA
	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
Self-Efficacy for Exercise	1.70 (1.07–2.69)	0.024	2.23 (1.49–3.33)	<0.0001	2.13 (1.36–3.36)	0.001
Age	0.98 (0.96–1.00)	0.089	1.00 (0.98–1.02)	0.880	0.98 (0.96–0.99)	0.021
Sex
Male	0.74 (0.47–1.18)	0.208	1.09 (0.72–1.64)	0.695	2.67 (1.71–4.15)	<0.001
Female†	—	—	—	—	—	—
Marital Status a
Single†	—	—	—	—	—	—
Married	1.49 (0.86–2.59)	0.208	1.21 (0.76–1.95)	0.418	1.19 (0.70–2.04)	0.517
Education
Less than HS Diploma†	—	—	—	—	—	—
HS Diploma	0.97 (0.60–1.58)	0.908	1.02 (0.67–1.57)	0.921	1.11 (0.70–1.76)	0.659
Some College/College Degree	0.94 (0.50–1.74)	0.831	1.07 (0.62–1.84)	0.801	0.92 (0.51–1.68)	0.797
Employment Status b
Employed†	—	—	—	—	—	—
Unemployed	0.60 (0.38–0.94)	0.026	0.59(0.40–0.89)	0.010	1.20 (0.78–1.86)	0.412
General Health Status
Excellent	1.95 (0.89–4.29)	0.097	1.22 (0.61–2.41)	0.576	3.14 (1.45–6.77)	0.003
Good	1.62 (0.92–2.85)	0.098	1.53 (0.94–2.49)	0.090	2.06 (1.13–3.75)	0.018
Fair/Poor†	—	—	—	—	—	—
Body Mass Index	0.98 (0.94–1.02)	0.255	0.98 (0.95–1.02)	0.311	1.01 (0.97–1.05)	0.635
Comorbidities c	0.89 (0.63–1.24)	0.489	0.86 (0.64–1.15)	0.302	0.83 (0.60–1.15)	0.255

Notes: Statistically significant p-values are bolded. PA=Physical Activity; HS=High School; OR=Odds Ratio; CI=Confidence Intervals.

^†=reference category.

^aSingle includes those who indicated they were currently single, divorced, widowed, and member of an unmarried couple. Married includes only those who indicated they were currently married.

^bEmployed includes those who indicated they were currently working full time, part time, and who were self-employed. Unemployed includes those who were out of work for less than 1 year, out of work for 1 year or more, disabled, students, retired, or taking care of their family member or home.

^cComorbities include self-reported previous diagnoses of cancer, diabetes, heart disease, or hypertension.

Moderate Physical Activity

Controlling for the eight covariates, for every 1 unit increase in SEE, the odds of engaging in more frequent moderate physical activity increased by over two times (OR=2.23, CI: 1.49–3.33, p<0.0001). Additionally, the odds of engaging in more frequent moderate physical activity for unemployed participants was approximately half that of participants who were employed (OR=0.59, CI: 0.40–0.89, p=0.01). None of the other covariates were significantly associated with frequency of engaging in moderate activity.

Vigorous Physical Activity

Controlling for the eight covariates, for every 1 unit increase in SEE, the odds of engaging in more frequent vigorous physical activity increased by over two times (OR=2.13, CI: 1.36–3.36, p=0.001). Additional variables were found to have an influence on vigorous physical activity frequency. Males had nearly three times greater odds of engaging in more frequent vigorous physical activity compared to females (OR=2.67, CI: 1.71–4.15, p<0.0001). For every one year increase in age there was a decrease in the odds of engaging in more frequent vigorous physical activity (OR=0.98, CI: 0.96–0.99, p=0.021). Those who reported excellent health had over three times the odds of more frequent vigorous physical activity than those who reported fair/poor health (OR=3.14, CI: 1.45–6.77, p=0.003). Finally, those who reported being in good health had twice the odds of more frequent vigorous physical activity compared to those who reported fair/poor health (OR=2.06, CI: 1.13–3.75, p=0.018).

DISCUSSION

Half of the sample (50.1%) reported engaging in sufficient levels of weekly physical activity, based on a total activity score computed from combining the number of times per week that moderate and vigorous physical activity was reported. While this result cannot be directly compared to nationally representative studies or the Physical Activity Guidelines for Americans due to the structure of the physical activity items (i.e., they do not ask respondents specifically about minutes of physical activity), it aligns fairly closely with past findings among NHPI. Specifically, a meta-analysis of nine studies examining engagement in physical activity among NHPI adults found that 48.7% of NHPI adults met guidelines for the combination of moderate and vigorous physical activity [11].

Despite half of the sample engaging in sufficient weekly physical activity, 20.1% of the sample reported rarely or never engaging in any kind of physical activity. This finding is consistent with inactivity levels observed among the general US population (23.8%), but notably lower than the prevalence of inactivity among the general Arkansas population (31.0%) [34]. There were also low levels of vigorous physical activity reported, with 63.2% engaging in vigorous physical activity once a week or less, and only 10.9% engaging in vigorous physical activity more than 4 times a week. This finding is consistent with past studies, which have found unhealthy levels of physical activity among Marshallese adults with type 2 diabetes in Arkansas [12, 13] and NHPI adults in California and Hawaii [10, 35].

The primary objective of this study was to examine the association of physical activity and SEE among a sample of overweight and obese Marshallese adults. SEE was significantly associated with more regularly engaging in sufficient weekly physical activity, and more regularly engaging in both moderate and vigorous physical activity. This finding aligns with prior literature, which shows that SEE is associated with physical activity levels after controlling for age, sex, and other sociodemographic characteristics [15–19]. While these variables have been explored separately among NHPI adults in Hawaii [20], this is the first study to our knowledge to document the associations between SEE and physical activity among NHPI in the US.

Participants who were unemployed at the time of data collection reported engaging in lower levels of sufficient weekly physical activity and lower levels of moderate physical activity. This finding is consistent with past research that has found significant associations between employment status and physical activity in other populations [36, 37]. While these studies also documented that the activity level of the job is important to this relationship, it is unclear whether the association observed in this study may be due to the type of jobs (active vs. sedentary) held by participants in our sample, as this data was not available. This study also found associations between vigorous physical activity frequency and being male, being younger, and being in better self-reported general health. These associations have been observed in past studies conducted with large, nationally-representative samples in the US [38–41].

The findings of this study should be interpreted with limitations in mind. First, the study sample consists of overweight and obese Marshallese adults recruited for a faith-based DPP randomized trial. This limits the generalizability of the study results to other racial/ethnic groups and other geographic locations. Second, the study is limited to analyzing cross-sectional data and no causal inferences can be made. The directionality of the relationship between SEE and engaging in physical activity is difficult to disentangle, given the reciprocal nature [15–19, 42–45]. Third, the outcomes of interest (e.g., physical activity and SEE) are self-reported, which can result in biased data. Fourth, in addition to being self-reported, the moderate and vigorous physical activity measures used to estimate total physical activity did not include a time component, thus limiting comparisons to the Physical Activity Guidelines for Americans and nationally representative statistics. Finally, the study did not assess participants’ sedentary time.

NEW CONTRIBUTION TO THE LITERATURE

To our knowledge, this is the first study to document the association between SEE and physical activity among Marshallese Pacific Islanders or NHPI in the continental US. Physical activity is a modifiable lifestyle behavior that is associated with many of the chronic disease disparities faced by the Marshallese community, as well as the broader NHPI population [1–3, 8, 12, 47]. The findings of this study will be useful for practitioners and researchers working to address these disparities with health promotion and disease prevention interventions. Specifically, these results highlight the need for physical activity interventions that can successfully engage women, older adults, those who are unemployed, and those who are in worse overall health.