Effects of Physical Activity, Diet, & Self-Efficacy on Physical Function in Older Adults

Background: Many older adults (OA) demonstrate decreased physical function (PF) which may lead to disability. Increased physical activity (PA), eating a healthful diet, and maintaining higher self-efficacy (SE) each improve PF in older adults, but few studies have examined if the combination of these three variables have an additive effect on PF. Additionally, few longitudinal studies have assessed the change in PF in relation to the changes in PA, diet, and SE. Purpose: The purpose of this study was to determine: (a) the relationship between PF and PA levels, fruit and vegetable intake (F&V), and exercise SE; (b) the percentage of variance in PF that is explained by PA levels, F&V intake, and exercise SE; and (c) if changes in PA levels, F&V intake, and exercise SE are related to changes in PF in community-dwelling OA. Methods: A secondary data analysis was conducted using data from the SENIOR II project. The participants (N=470) were community-dwelling OA (M=79.9, SD=5.8). PF was measured using the Timed Up and Go. F&V intake, PA and exercise SE were measured using the NCI Fruit and Vegetable Screener, the Yale Physical Activity Survey, and the Exercise Self-Efficacy Scale, respectively. Results: Pearson’s r correlation showed that PF was significantly related to PA (r=-.207, p<.001), F&V (r=-.125, p<.001), and SE (r=-.120, p<.01). Multiple regression analysis revealed that PA, F&V, and SE explained almost 15% of the variation of PF in OA. A repeated measures MANCOVA revealed that vigorous PA levels increased in individuals whose PF improved from baseline to 48-months. Conclusion: PA, F&V, and SE combined had an additive effect on PF in OA and explained a greater variance in PF than each individual variable. Declines in PA, F&V, and SE did not result in significant declines in PF in OA; however, results indicate that variable levels at a younger age may be related to PF in OA.


INTRODUCTION
Over the past 55 years, the average life expectancy in the US has increased from 69.8 years to 78.8 years. 1 As life expectancy increases, older adults (OA) are facing challenges not previously experienced by earlier generations, including increased risk and prevalence of chronic diseases, injuries, and disabilities; increased healthcare costs; and decreased quality of life. [2][3] Maintaining high levels of physical function (PF) can play an important role in delaying or preventing these challenges. [4][5][6] PF is an all-encompassing term that includes an individual's mobility as well as their capability to perform activities of daily living (ADLs) and instrumental activities of daily living (IADLs). 7 A decline in PF has been associated with increased mortality, rate of hospitalization and injury, and morbidity 4,8 , while improvements in PF have resulted in decreased hospital and nursing home admission and risk of falls 6 .
Physical activity (PA), fruit and vegetable consumption (F&V), and selfefficacy (SE) are related to PF in OA. [9][10][11] Cross-sectional studies have demonstrated that higher levels of PA and SE have been consistently correlated with higher levels of PF 12 , while F&V consumption has been inversely associated with disability 16 and positively correlated with lower extremity strength. [18][19] Unfortunately, levels of PA and F&V are consistently lower in OA compared to younger populations. Less than 25% of OA are meeting national guidelines for PA 20  respectively. 16 Lower SE is associated with decreased PF in OA. 15,20 SE demonstrates an inverse relationship with aging due to the perception that there is an automatic decline in PF with increasing age. 14,20 Interventions designed to improve PA, F&V, and SE in OA have resulted in significant improvements in PF. PA specific interventions have resulted in significant improvements in PF. 22 An intervention by Neville et al. designed to increase F&V found non-significant increases in PF 23 ; however, when a PA component was added to the F&V intervention there were significant improvements in PF. 24 Similarly, PA interventions that incorporate a SE component resulted in significantly greater improvements in PF in OA compared to interventions that do not include a SE component. [25][26] This may be due to the relationship found between changing PA and SE levels in OA. [27][28] No research has been identified on the combined effect of PA, F&V, and SE on PF in OA. Therefore, the purpose of this study was threefold to determine: (1) the relationship between physical function and PA levels, F&V intake, and exercise SE; (2) the percentage of variance in PF that is explained by PA levels, F&V, and exercise SE; and (3)

Physical Activity
Physical activity was measured using the Yale Physical Activity Survey (YPAS). The YPAS is an interviewer-administered survey developed to assess the PA levels of OA in a typical week within the past month. 31 Participants quantified the frequency and duration of participation in five activity dimensions: vigorous activity, leisurely walking, moving around on feet, standing, and sitting. Weights were assigned to each dimension. The frequency and duration scores were multiplied together and then multiplied again by each dimension's weighting factor to calculate an index for each dimension. The five individual indices were summed to calculate the summary index.

F&V Consumption
F&V consumption was measured using the National Cancer Institute Fruit and Vegetable (NCI F&V) Screener. The NCI F&V Screener is a 19-item questionnaire that assesses the number of servings of F&V consumed in an average day over the past month. 32 It includes consumption of whole F&V, as well as servings of F&Vs found in soups and juices.

Exercise Self-efficacy
The six-item exercise SE scale measured a subject's confidence in his/her ability to exercise despite adverse or challenging situations. 33 Self-efficacy scores range from 1 to 5, with higher scores indicating greater SE.

Physical Function
Physical function was measured using the Timed Up and Go (TUG) test. The TUG is a practical, reliable measure of functional mobility in older adults. 34 The TUG test measures, in seconds, the time taken by an individual to stand up from a standard chair, walk a distance of three meters, turn, and walk back to the chair and sit down again. The score is recorded in seconds taken to complete the task. TUG scores have been able to distinguish between OA who need assistance in ADLs, those who are independent, and those with somewhat impaired mobility. 35 Table 3).
Of the five YPAS Summary Index components, moving around (r=-.24, p<.001), followed by sitting (r=.24, p<.001) and vigorous activity (r=-.24, p<.001) had the strongest correlation with TUG scores, indicating that those who spent more time moving around on their feet while doing different daily tasks, participated in vigorous PA, and spent less time sitting had better PF based upon a lower TUG score.
Significant correlations were found between F&V intake (r=-.11, p<.01) and fruit intake alone (r=-.11, p<.01) with TUG scores, and no significant correlation was found between vegetable intake alone (r=-.06, p>.05) and TUG scores. These results indicate there is a significant, relationship between PA, F&V consumption, and SE and PF in OA.
To look at the independent effects of PA, F&V, and exercise SE on PF, a series of hierarchical regression models were constructed to examine the unique contribution of these variables on TUG scores (see Table 4). PA (r 2 =. 121 This supports the results of a previous study that found that OA who met the recommended servings of F&V had significantly lower odds of developing disability relating to IADLs, with fruit providing additional protection against developing lower extremity mobility and general disabilities. 17 No studies were found that analyzed the correlation between fruit and/or vegetable intake with PF in OA. However, correlational studies have been conducted using levels of micronutrients found in fruits vegetables. Vitamin C, vitamin E, beta carotene, and retinol were found to be significantly correlated with physical function in older adults. [40][41] The results of both studies indicate a weak correlation similar to that found in the present study.

Physical Function in Older Adults
Disability, according to Nagi's Disability Model, is defined as (a) limitation(s) in performing activities required for an individual's role in society--including self and home care, work, and community participation. 6 Disability includes limitations or impairments related to mobility and independence with activities of daily living (ADLs; basic self-care activities, including bathing and eating) and instrumental activities of daily living (IADLs; activities required for independent living within the community, including grocery shopping, and driving). 3,4 Individuals with disability demonstrate a lower life expectancy 4,7 , decreased quality of life 8 , increased risk of hospitalization 9-10 , increased health care costs 10 , and decreased independence. 10 Physical function (PF) is a precipitator of 6 and is highly correlated with disability 11 and successful aging. 12 Functional limitations are impairments at a performance level 6 that would include such things as impaired gait or decreased independence with transfers. A decline in PF has been associated with increased mortality 13 , rate of hospitalization and injury 14 , falls 15 , and a decline in cognitive function. 16 In a longitudinal study, Dapp, Minder, Anders, Golgert, & von Renteln-Kruse (2014) found that functional ability predicted mortality and that the need for nursing care was independent of age and gender. 17 Improvements in PF result in decreased hospital and nursing home admission and risk of falls. 18 There is a current trend towards the reduction of disability by 1.0% to 2.5% each year within the U.S. 3 ; however, this still leaves millions of OA with disabilities.
To further decrease disability prevalence among OA, the correlates of PF need to be identified. Once these correlates have been identified, evidence based interventions can be developed and implemented to improve PF, and ultimately disability rates, in OA. Three correlates that have been studied most frequently in the literature are physical activity (PA), diet, and self-efficacy (SE). This review of literature will present the status of these correlates within the OA population, the relationship of these correlates with PF, and the effects of interventions designed to improve upon these correlates.

Measuring Physical Function in Older Adults
To assess a change in PF in OA, it is important to utilize a reliable and valid measure that has established normative data and is sensitive to clinically significant changes for this population. The Timed Up and Go (TUG) test is one such measure. [19][20][21][22] Originally developed to assess disability in frail OA, this measure has been found to predict future ADL and IADL disability 19 , falls 22 , and mortality in community dwelling OA. [23][24] Equipment required for this test includes a stopwatch and a chair without armrests that allows the participant to sit with hips and knees in 90 degrees of flexion.
The participant is asked to stand up from the chair without the use of his or her upper extremities, ambulate 3 meters, navigate around a cone, ambulate back to the chair, and return to a seated position, once again without the use of upper extremities. The tester begins the stopwatch when he or she gives the command 'Go'. The score is the time, in seconds, for the participant to complete the task. Lower scores are associated with higher physical function, and higher scores are associated with lower physical function.

Physical Activity Levels in Older Adults
With increasing age, there is a trend towards decreased PA and increased sedentary time. [25][26][27][28][29]  The form and intensity of PA performed changes with age, as demonstrated by the prevalence of running, team sports, weightlifting, and aerobics in younger adults compared to OA, who in turn demonstrate a greater prevalence of walking, yardwork and gardening, golf, and bicycling, with walking, bicycling, and gardening accounting for over 75% of the PA performed by PA. 29 These trends reflect a transition from moderate to vigorous PA performed in younger adults to light to moderate intensity PA as adults age. 29 This transition to lower PA intensity with increasing age is often due to lack of time, fear of falling or injury, physiological impairments, and lack or resources, with the latter three specific to OA. [34][35] Physical Activity and Physical Function Consequences of decreased PA are decreased aerobic endurance, muscular strength, and balance, as demonstrated by the positive relationship between PA and these physiological measures of fitness in OA. [36][37][38] Sarcopenia, a reduction in muscle mass and strength due to aging, affects 7% to 50% of the population ≥65 years of age and increases the risk of disability by 79%. 39 PA has been shown to reduce or prevent the loss of strength associated with sarcopenia. 36,40 Strength, balance, and aerobic endurance are required to safely and independently perform activities of daily living (ADLs) 36 , which is a component of physical function (PF); thus, it is reasonable to explore the direct relationship between PA and PF.

PA positively correlates with an individual's level of PF and negatively
correlates with an individual's level of disability. 33,[41][42][43][44] This relationship is independent of body mass index (BMI) or weight 33,42,45 , with research showing that physically active overweight or obese individuals have higher levels of PF and decreased disability compared to their normal weight, sedentary peers. However, one study found that PA preserves PF in obese males but not females. 46 This relationship remains significant when accounting for covariates, such as age, gender, education, and smoking 42 , indicating that PA may be more predictive of PF than those covariates.
When defining PA as steps taken per day (including steps during leisurely activity instead of structured exercise), higher step counts were positively correlated with PF, indicating the benefit of total daily activity on PF. 47 Researchers PF outcome measures. 41,48 One study found that subjects who participated in light (N=17, mean age 70.3±5.7) and vigorous PA (N=17, mean age 69.8±4.4) improved overall balance, only vigorous PA significantly improved dynamic balance as demonstrated by gait and sit to stand movements. 56 Other cross-sectional studies indicate that light PA is as beneficial as MVPA to preserve PF [48][49]54 , with two studies indicating no statistically significant difference in PF between groups who performed different intensities of objectively and subjectively reported PA. 49,57 The relationship between the time spent being physically active and PF has also been investigated. Studies suggest that greater total time spent performing PA correlates with higher levels of PF. [48][49]57 Adults who meet the PA recommendations have higher functional scores than their peers who do not meet the recommended PA levels. 41,[58][59] Dropping below the recommended PA threshold results in a clinically significant decline in PF. 59 In contrast, two studies found a strong positive relationship between regular PA and PF even in individuals who did not meet national PA guidelines 44,60 , with an increase of 10 minutes/day of even low intensity PA resulting in a significant improvement in PF. 60 Gebel et al. (2014) found that, independent of total MVPA, a 1% increase in total time spent being physically active resulted in 0.3% decreased risk of a decline in PF. 58 Nonetheless, it remains unclear, if PA initiated later in life is as protective of PF in OA as PA levels in adults <65 years of age. Following a 9-month PA intervention, OA between 65-74 years of age were able to maintain significant improvements in PF, whereas OA 75 years of age and older did not maintain significant improvements in PF. 61 Stenholm et al. (2015) found that individuals with a higher level of PA in early adulthood and late midlife had a smaller decline in PF compared to individuals with lower levels of PA in early adulthood. 62 These results support the findings by Manini  The lack of consensus among the studies cited could be secondary to poor adherence and different exercise protocols, both of which were not adequately described in all studies included in the review.
PA interventions have been successful in improving PF in OA 38,64,67-68 , with differences found between individuals aged 64 to 74 years and 75 and older in changes in strength. 61 Ip et al. (2013) demonstrated that OA between the ages of 70 and 89 who participated in a PA program decreased the odds of experiencing a decline in PF by 60%. 64 However, 56% of individuals participating in both the PA and successful aging groups did not experience any change in PF from baseline to 12 months, demonstrating the variability of response to a PA intervention and the possibility of additional correlates of PF not addressed by the intervention.
PA interventions have been implemented in unhealthy OA, including those classified as having dementia 70 or frail. 71 In OA with dementia, PA interventions have been found to increase static and dynamic balance 70 , but only static balance in frail OA. 71 In frail OA, exercise was found to improve gait speed, but not Timed Up and Go scores [71][72] , and to have mixed results on ADL performance. 72 One study found that not all participants improved in PF following an exercise intervention, with some participants actually demonstrating a decrease in PF. 73 Following two PA interventions, it was found that PA decreased the risk of moving to a lower state of PF by 60%. 64 PF has been strongly correlated with physiological variables, such as strength, balance, and endurance, as well as PF in a variety of OA subpopulations.
Consequently, there have been many exercise interventions implemented to preserve or improve PF in this population. Among those studies, however, lies a lack of consensus regarding the benefits of a PA intervention. This is due to differences in participant characteristics, the number of participants, the duration of the intervention, and the design of the intervention. Further studies need to be conducted to analyze the effects of PA on the PF and physiological variables required for OA to carry out ADLs.

Diet of Older Adults
The overall American diet quality does not meet the recommendations set forth in the Dietary Guidelines for Americans. 74

Interventions to Improve Diet and Physical Function
Although there is significant evidence for a relationship between F&V consumption and PF in OA, there are few F&V intervention studies with a PF outcome published. [100][101] There are multiple studies with a general or protein-specific intervention to improve PF, often in combination with increased PA or weight loss, that have reported a significant improvement in PF. [96][97][98] However, one study found no significant change in PF among participants receiving a dietary intervention group, 99 although the results of this study may have been confounded by high levels of PF at baseline.
There have been some interventions implemented to increase F&V consumption in OA; however, only two were found that had PF as the primary outcome of the study. Those two studies were multimodal interventions--combining PA and dietary components. [100][101] Only one study found significant improvements in PF following an increase in PA and F&V consumption 100 , while the other study found no significant improvement. 101 Only one intervention that solely implemented a F&V or F&V-related intervention used a subject population of OA. 102

Relationship between Self-Efficacy and Physical Function
The level of an individual's SE has been found to be significantly related to PF 111-112 and mortality. 106,109 This is supported by findings that SE mediates the effect of age on walking performance in OA 113 and Timed Up and Go (TUG) 114  The effects of SE on PF are mediated by PA. Higher SE is a determinant in adopting and maintaining a physically active lifestyle [118][119][120][121][122] and participating in higher intensities of PA. 118 This is due to the effects of an individual's SE on his/her ability to exercise for various amounts of time, overcome barriers to exercise, and recover from failures or setbacks. 119-120 This influence has been demonstrated across age groups. 123 Declining PA levels associated with increased age are also correlated with low SE. 124 Additionally, changes in PA levels have resulted in significant changes in SE 116 , with increased PA levels improving the SE of OA. 122,[125][126] This relationship has also been demonstrated in frail and diseased OA. 122,125 This positive change in SE of OA due to PA may mediate the improvements in PF observed with increased PA. 111 Interventions to Increase Self-Efficacy and Physical Function There are few published interventions designed to specifically improve SE in OA with the aim of improving PF. There are multiple PA interventions for OA with PF as a primary outcome and SE as a secondary outcome. 127 One study found that PA improved SE; however, the improvement in SE was not related to the improvement demonstrated in PF. 127 Interventions that include a SE specific component for OA have resulted in increased PF 125,[128][129][130] or quality of life. 131 SE has increased significantly following an exercise intervention for OA, which in turn was significantly related to improved PF. 125,130 Individuals who received a SE+PA intervention improved PF significantly more than individuals who received a PA-only intervention. 129 SE has been found to be a significant predictor of PF post-intervention compared to exercise alone; however, the overall intervention effect was due to the exercise component. 128