Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Jun 1.
Published in final edited form as: Health Psychol. 2017 Mar 9;36(6):568–576. doi: 10.1037/hea0000479

Morning Self-Efficacy Predicts Physical Activity Throughout the Day in Knee Osteoarthritis

Ruixue Zhaoyang 1, Lynn M Martire 1,2, Martin J Sliwinski 1,2
PMCID: PMC5466814  NIHMSID: NIHMS857642  PMID: 28277696

Abstract

Objective

The purpose of this study was to examine the within-day and cross-day prospective effects of knee osteoarthritis (OA) patients’ self-efficacy to engage in physical activity despite the pain on their subsequent physical activity assessed objectively in their natural environment.

Methods

Over 22 days, 135 older adults with knee OA reported their morning self-efficacy for being physically active throughout the day using a handheld computer and wore an accelerometer to measure moderate activity and steps.

Results

Morning self-efficacy had a significant positive effect on steps and moderate-intensity activity throughout that day, above and beyond the effects of demographic background and other psychosocial factors as well as spouses’ support and social control. The lagged effect of morning self-efficacy on the next day’s physical activity and the reciprocal lagged effect of physical activity on the next day’s self-efficacy were not significant. Positive between-person effects of self-efficacy on physical activity were found.

Conclusions

Future research should aim to better understand the mechanisms underlying fluctuations in patients’ daily self-efficacy, and target patients’ daily self-efficacy as a modifiable psychological mechanism for promoting physical activity.

Keywords: osteoarthritis, self-efficacy, physical-activity, older adults, daily diary

Osteoarthritis (OA) is a common chronic health condition and a leading cause of pain among adults, especially older adults. According to the Centers for Disease Control and Prevention (CDC, 2005), OA affected 33.6% (12.4 million) of American adults 65 years and older. Physical activity is well-documented as one of the most effective non-drug treatments to reduce or manage OA symptoms (Gay, Chabaud, Guilley, & Coudeyre, 2016). Previous research has demonstrated that physical activity is associated with reduced pain (Focht, 2006), improved physical function (Pisters et al., 2010), and higher quality of life for people with OA (Bieler et al., 2016). Therefore, the US national physical activity guideline recommends that every adult, including those with arthritis, should accumulate at least 150 minutes per week of moderate-intensity physical activity (The Physical Activity Guidelines Advisory Committee, 2008).

Despite the potential health benefits of physical activity, the majority of people with OA are not physically active (Wallis, Webster, Levinger, & Taylor, 2013). According to a recent review, only 13–48% of people with knee OA meet current guidelines and recommendations for moderate-intensity physical activity and daily steps; and people with OA are 25% less physically active than age-matched peers without arthritis (Wallis et al., 2013). Previous research has demonstrated that compared with healthy individuals, people with knee OA are more likely to avoid physical activity in order to prevent activity-related pain and stiffness in their knees, which in turn, is associated with deterioration of muscle strength and greater limitations in activities in the long-term (Pisters, Veenhof, van Dijk, & Dekker, 2014). Therefore, it is important to understand the proximal factors that may facilitate physical activity in people with OA.

Self-efficacy and physical activity in people with OA

The present study focused on one of the most well-established psychological factors that impact physical activity in the general population, self-efficacy for physical activity, and investigated whether fluctuations in morning self-efficacy predict changes in physical activity throughout the day for people with knee OA. Previous research has demonstrated that the physical activity adherence in people with OA is influenced by multiple factors including demographic, physical, and psychological factors (see Stubbs, Hurley, & Smith, 2015; Veenhof, Huisman, Barten, Takken, & Pisters, 2012 for reviews). For example, age, female gender, higher BMI, non-white ethnicity, and greater severity of OA symptoms, especially greater pain (Murphy, Niemiec, Lyden, & Kratz, 2016), are associated with less physical activity in adults with OA, while improved physical function and support from the spouse are positively associated with physical activity (Martire et al., 2013). As noted by previous researchers, however, research that identifies psychological factors that impact physical activity among people with OA remains limited (Stubbs et al., 2015; Peeters, Brown, & Burton, 2015).

Among a number of psychological factors, self-efficacy–one’s confidence in the ability to execute specific actions required to achieve specific outcomes (Bandura, 1997)–has been found to be the most consistent and robust predictor of physical activity in healthy adults (David et al., 2014; Dunton, Atienza, Castro, & King, 2009; McAuley et al., 2011), people with chronic health conditions such as obesity, diabetes or cancer (Basen-Engquist et al., 2013; Teixeira et al., 2015), as well as people with mental health conditions (Vancampfort et al., 2015). According to Social-Cognitive Theory (Bandura, 1997), people’s beliefs in their capabilities to produce desired effects by their own actions are the most important determinants of their subsequent behaviors and how much they persevere in their efforts in the face of obstacles and challenges. Typically, a high sense of self-efficacy helps people to approach more challenging tasks, maintain strong commitment to them, and sustain or even increase efforts in the face of obstacles, barriers, and aversive situations (Bandura, 1997). Such efforts to overcome challenges might be especially important for people with OA due to the fact that physical activity is often associated with pain and pain-related fears and avoidance (Pisters et al., 2014). Therefore, it is reasonable to expect that for people with OA, how confident they feel about their ability to engage in physical activity despite pain should have a significant influence on their future physical activity.

To date, only a few studies have investigated the association between self-efficacy and physical activity in people with OA, showing a positive association between the two constructs (e.g., Peeters et al., 2015). However, these studies relied on subjective self-report measures to assess physical activity (e.g., Peeters et al., 2015), which have been shown to be less reliable than objective measurement (Downs, Van Hoomissen, Lafrenz, & Julka, 2014). In addition, most previous studies were cross-sectional, and as a result, could not examine the temporal order of self-efficacy and physical activity. To our knowledge, only one study used longitudinal data to examine the effect of self-efficacy on physical activity. In this study, Peeters and colleagues (2015) found that among middle-age adults with arthritis, participants who reported improved self-efficacy over two years had 2.20 higher odds of increased physical activity over the same time period than those who did not. This study, however, did not establish the temporal order between self-efficacy and physical activity because these two variables were assessed concurrently. Finally, previous studies also have focused exclusively on between-person associations, which only address the question of whether people with OA who have higher self-efficacy are more physically active than those with lower self-efficacy. However, self-efficacy as a dynamic self-regulatory process has been shown to fluctuate from day to day among healthy adults (Dunton et al., 2009), cancer survivors (Basen-Engquist et al., 2013) and postmenopausal women (David et al., 2014). Therefore, it is reasonable to expect that the self-efficacy for physical activity will vary from day to day for people with OA, and it is important to examine how such a within-person variation in self-efficacy impacts subsequent physical activity: when people with OA experience higher levels of self-efficacy relative to their own mean level at one point in time, whether they are more likely to engage in more physical activity relative to their own mean level at a subsequent point in time. Examining such a temporal-ordered within-person association between self-efficacy and physical activity is a crucial first step toward establishing the causal relationship between self-efficacy and physical activity. It is also of particular interest of future interventions that are designed to causally induce changes in physical activity among people with OA: whether boosting people’s self-efficacy for physical activity could produce immediate and direct impact on their physical activity.

The current study

The current study used data collected with handheld computers and accelerometers over a 22-day period among a sample of people with knee OA. By disaggregating the within-person and between-person effects of self-efficacy for physical activity reported in the morning on the subsequent physical activity that day (i.e., number of steps, and minutes of moderate-intensity activity), the current study predicted that 1) on days when people with knee OA felt greater efficacy than usual to be physically active, they would engage in more physical activity throughout the day (within-person effect); and 2) consistent with previous findings, people with knee OA who were more confident in their ability to be physically active would engage in more physical activity than their less confident counterparts in general (between-person effect). It is notable that the design of the current study–self-efficacy was assessed every morning before patients engaged in any physical activity–allowed us to clearly examine the temporal associations between self-efficacy and physical activity. Furthermore, in order to provide a robust test of our hypothesis, we also examined the predictive effects of patients’ self-efficacy for physical activity independent of their background characteristics (e.g., age, gender, education, general health status) and other psychological factors that have been showed to be related to physical activity at the daily level, such as patients’ daily pain (Murphy, Smith, Clauw, & Alexander, 2008; Murphy et al., 2016), positive and negative affect (Liao, Shonkoff, & Dunton, 2015), and spouses’ support and control of patients’ physical activity (Martire et al, 2013). In addition, we examined whether bidirectional associations between self-efficacy and physical activity can extend across days.

Method

Participants and Procedures

Data for the current study were from a larger couple study in which one partner (i.e., the patient) was diagnosed with knee OA (for a detailed description, see Martire et al. 2013). This study included three in-person interviews conducted over a period of 18 months and a 22-day daily assessment immediately after the first interview. During the daily assessment, patients and spouses used handheld computers to complete a short survey three times per day (i.e., morning, afternoon, and end-of-day). Each patient and spouse also wore an accelerometer continuously for 22 days in order to provide objective data on their daily physical activity. The University of Pittsburgh Institutional Review Board (IRB) approved this study (IRB # 07030113: Daily individual and marital processes in osteoarthritis), and an IRB-approved form was used to obtain written informed consent from all participants. The current study used patients’ data from the first interview, morning and end-of-day assessments, and accelerometers.

Primary sources of recruitment were research registries for rheumatology clinics and for older adults interested in research, flyers distributed at the University of Pittsburgh, and word of mouth. To be eligible for this study, patients had to be diagnosed with knee OA, experience usual knee pain of moderate or greater intensity, be at least 50 years of age, and be married or in a long-term relationship and living with their spouse or partner. Exclusion criteria for couples included: (1) patient had a comorbid diagnosis of fibromyalgia or rheumatoid arthritis; (2) patient planned to have hip or knee surgery in the next six months; (3) spouse had arthritis pain of moderate or greater intensity, or required assistance with personal care activities; (4) either patient or spouse used a wheelchair to get around; and (5) either patient or spouse was not cognitively functional as indicated by the accuracy of their answers to questions regarding the current date, day of the week, their age, and birth date.

A total of 606 couples were screened for eligibility. Of these, 221 couples declined to participate; the most frequent reasons were lack of interest (N=87) and illness in the family (N=55). An additional 233 couples were ineligible for the study; the most frequent reasons were lack of OA of the knee (N=55) and OA pain that was mild (N=47). A total of 152 couples were enrolled in the study and a total of 145 couples completed the diary assessment section. Due to missing accelerometer data, the sample size for the current analyses was 135 patients. Table 1 provides descriptive information for the sample.

Table 1.

Demographic Characteristics of the Sample

Variable M or % S.D.
Gender (male) 43.70%
Age (years) 65.71   9.83
White 86.67%
Employed (yes) 42.96%
Education (years) 15.94   2.03
Years of Marriage 34.33   16.80
Household Income $64,798.00   $11,000.00
Body Mass Index 31.25   6.00
Duration of Knee OA (years) 16.33   12.37
Open in a new tab

Note. N=135. OA= Osteoarthritis.

Measures

Self-efficacy for physical activity

Patients reported their level of confidence in being physically active during the day ahead using two items in the morning assessment: how confident are you that you can be physically active today despite pain (McAuley, 1992), and how confident are you that you can be physically active today in ways that will help your arthritis symptoms (developed for this study). Before answering these two questions, patients were told that physical activity refers to moving about the house or yard on a frequent basis, traveling to places outside of the home, engaging in organized sports activities, or exercising. These 2 items were rated on a 10-point scale (1=not at all confident; 10=totally confident). Items were averaged to create a mean score of patients’ self-efficacy for physical activity during the day ahead (α= 0.77, Spearman-Brown Coefficient: 0.80).

Pain

Patients reported the pain or tenderness in ten sets of joints over the past 30 minutes during the morning assessment (Mason et al., 1992). Separate ratings were made for multiple joints or joint groups (e.g., knees, hips, hands) on a scale from 0 to 3 (0=no pain/tenderness; 3=severe pain/tenderness). Items were averaged to create a mean score of patients’ pain in the morning (α= 0.90).

Positive and negative affect

Patients reported to what extent they felt positive or negative affect over the past 30 minutes during the morning assessment (Thomas & Diener, 1990). Nine items were rated on a 7-point scale (0=not at all; 6=extremely). Five items (e.g., depressed or blue, frustrated, angry or hostile, unhappy, worried or anxious) were averaged to create a mean score of negative affect (α= 0.91), and four items (e.g., happy, joyful, pleased, enjoyment) were averaged to create a mean score of positive affect (α= 0.98).

Activity-related spouse support and control

At the end of day assessment, patients reported the extent to which the spouse had provided support or control regarding their physical activity that day. A total of nine questions were used to assess spouse autonomy support, persuasion, and pressure (Martire et al., 2013). All items had a three-point response option (1=not at all, 2=somewhat, 3=very much). The autonomy support scale was the average of three items (α= 0.90, e.g., showed understanding for how physically active I wanted to be). The persuasion scale was the average of four items (α= 0.88, e.g., tried to persuade me to be more physically active). The pressure scale was the average of two items (α= 0.89, e.g., criticized, made fun of, or teased me for not being more physically active).

Daily physical activity measures

Accelerometers were used to assess amount of time spent in moderate-intensity activity and daily steps taken during the 22-day assessment period. Accelerometers are motion-sensitive monitors that count the number of movements or steps taken per pre-specified time interval. Participants wore the GT1M or GT3X model of the CSA/MTItri-axial ActiGraph, with placement on the hip in order to best capture ambulatory activities. Data were collected in 1-minute epochs. Patients were instructed to wear the monitor during the day and remove it at night; a reminder to put the monitor on in the morning was provided electronically via the handheld computers. Participants used a written log to record any periods during which they did not wear the accelerometer. All times when the monitor was not worn were removed from data analysis. Data were then screened for anomalous values (activity counts >6,000 at any given minute), which affected <1 % of the activity data. Remaining data were coded to determine the total number of hours of wear time. A valid day having sufficient accelerometer data was defined as 10 or more hours of accelerometer wear. Data were downloaded every week from the accelerometers and these devices provided no feedback to participants.

Intensity of physical activity is commonly defined using an activity count cut point that corresponds to a range of metabolic equivalents. Previous work confirms that the common cut point for moderate intensity activity (1952 or more activity counts per minute) underestimates walking intensity in older adults as measured by oxygen consumption during treadmill tests (Brach, Wert, Van Swearingen, & Studenski, 2009). Thus, consistent with our previous research (Martire, et al., 2013; Martire, et al., 2016), we relied on work by Matthews and colleagues (Matthews, 2005; Matthews, Ainsworth, & Hanby, 2005) which combined information from laboratory and field studies to determine a cut point that may better capture moderate-intensity activities of daily living. Specifically, we defined 760 activity counts per minute or greater as moderate-intensity activity (i.e., moderate-to-vigorous, hereafter referred to as moderate for simplicity)1. In addition, daily steps were measured by the accelerometers as another indicator of patients’ daily physical activity. Moderate activity and steps have each been linked with different health outcomes (e.g., Brach et al., 2003; Moreau et al., 2001) and both describe physical activity in metrics that are readily understood and have real-world referents.

Data Analysis

Multilevel modeling was used to examine the associations between patients’ self-efficacy and subsequent physical activity (Singer & Willett, 2003). The data were structured hierarchically, with daily assessments (level 1) nested within persons (level 2). Thus, self-efficacy and physical activity could vary over days within a person as well as across persons. All analyses were conducted using SAS PROC MIXED with restricted maximum likelihood (REML) and robust standard errors (Maas & Hox, 2004) to adjust against the slight skewness of the outcomes (e.g., skewness of level 1 and level 2 residuals of steps: 0.26 and 0.50; skewness of level 1 and level 2 residuals of minutes in moderate-intensity activity: 0.96 and 1.05). The within-person interdependence of the repeated outcome assessments was accounted for by using autoregressive, AR (1), covariance structure for residual errors.

Analyses were conducted in a series of steps. In the first step, an empty model with a random intercept but no predictor was tested for each key variable to estimate the variance at different levels. In the second step, the within-day prospective associations between patients’ morning self-efficacy and subsequent physical activity throughout that day were examined. In this model (Model 1), predictors included a level 1 person-mean centered self-efficacy score which captures each person’s daily deviation from his or her own mean score of self-efficacy (within-person effect, WP), a level 2 mean score of self-efficacy which captures the individual differences in the average score of self-efficacy (between-person effect, BP), as well as the day of study. In each model, a random intercept was included to allow the mean score of physical activity to vary across individuals, and a random slope of the within-person effect of self-efficacy was also included if it was statistically significant to allow the association between daily self-efficacy and physical activity to vary between individuals.

In the third step of analysis, bivariate simple regression was conducted to identify the level 2 covariates (e.g., demographics) that were significantly linked to each physical activity outcome. The significant level 2 covariates were added to the model to examine the effects of self-efficacy on physical activity controlling for individual differences in background variables (Model 2). In the next model (Model 3), level 1 covariates that were suggested by previous research as physical activity antecedents (e.g., positive and negative affect, pain, spouses’ support and control) were added to the model to provide more robust tests of the unique predictive effects of self-efficacy on physical activity above and beyond the effects of other social psychological variables. In these models, all level 1 covariates were person-mean centered while level 2 covariates were grand-mean centered (Singer & Willett, 2003).

In addition to within-day analysis, we examined the cross-day lagged effect of self-efficacy by predicting day t’s physical activity from day (t-1)’s self-efficacy. In order to differentiate and account for the concurrent effect, day t’s self-efficacy was also included in the model. The lagged measure of the outcome (i.e., day (t-1)’s physical activity) was not included because it was highly likely to be correlated with the random components of the multilevel model, thereby violating the key model assumption (Spencer, 2002). Finally, the potential bidirectional lagged effect was tested by predicting day t’s morning self-efficacy from day (t-1)’s physical activity.

Results

Compliance Rates and Missing Data

Out of 3190 potential morning or end-of-day assessments (145 participants × 22 days), patients completed a total of 2948 morning assessments (92%) and 2960 end-of-day assessments (93%). At the person level, the average number of completed assessments for each participant was 20.33 (SD = 2.63, Range = 2–22) for morning assessments and 20.41 (SD = 2.45, Range = 8–22) for end-of-day assessments, respectively. Out of 145 patients, 135 provided sufficient accelerometer data (i.e., having 10 or more hours of wearing time per day) for 2624 out of 3190 potential daily accelerometer assessments (82%). The average number of days that the accelerometer was worn by patients was 19.44 (SD = 3.31) and the average daily accelerometer wear time was 14.67 hours (SD=1.68).

Missing data analyses suggested that the percentage of missing daily self-report or accelerometer data did not vary with person level covariates such as gender, age, race/ethnicity, education level, employment status, income, general health status, or the average scores of self-efficacy, positive affect, negative affect, pain, as well as spouse support and control across the 22 days. The day in the study had a significant effect on the likelihood of missing a daily morning assessment (b = 0.07, p < 0.001), or evening assessment (b = 0.06, p < 0.001), but not accelerometer data. Other level 1 covariates such as pain, positive and negative affect reported in the morning did not significantly impact the likelihood of missing daily accelerometer data.

Preliminary Analyses

For the purposes of preliminary analysis, mean scores were calculated for each of the daily measures across the 22-day period for each person. As shown in Table 2, on average, patients took 4178.75 steps (SD=2002.11) and spent 61.83 minutes (SD=38.30) in moderate-intensity activity each day. The intraclass correlations (ICCs) for daily steps and moderate-intensity activity were both 0.52, indicating that 52% of the total variance in these two variables occurred between persons while 48% of the total variance occurred from day to day within a person. Therefore, there was significant variability in daily steps and moderate-intensity activity at both between- and within-person levels. The ICC of the morning self-efficacy score was 0.79, suggesting that self-efficacy for physical activity varied more at the between-person level (79%) than at the within-person level (21%), but there was still considerable variance in self-efficacy from day to day for each person. Furthermore, the correlations show that patents who reported higher self-efficacy on average also took more daily steps and engaged in more moderate-intensity activity.

Table 2.

Descriptive Information and Intercorrelations of Aggregated Daily Variables

1. Self-Efficacy 2. Steps 3. Moderate Minutes a 4. Pain 5. Positive Affect 6. Negative Affect 7. Autonomy Support 8. Persuasion Control 9. Pressure Control
M. 7.14 4178.75 61.83 0.58 2.46 0.54 2.08 1.19 1.05
S.D. 1.87 2002.11 38.30 0.47 1.36 0.63 0.51 0.26 0.15
ICC   .79         .52     .52   .85   .69   .48   .68   .61   .50
α   .77         –     –   .90   .98   .91   .90   .88   .89
1. Self-Efficacy   –
2. Steps   .25**       –
3. Moderate Minutes a   .15+         .84***     –
4. Pain −.42***       −.20*   −.11   –
5. Positive Affect   .28***         .09     .14+ −.04   –
6. Negative Affect −.37***       −.11   −.05   .24** −.35***   –
7. Autonomy Support   .13         .03     .04   .03   .31**   .01   –
8. Persuasion Control −.10       −.06   −.00   .00 −.01   .18* −.11   –
9. Pressure Control −.22**       −.21**   −.12*   .21*   .03   .21*   .10   .67***   –
Open in a new tab

Note. ICC=Intraclass Correlation.

a

Minutes in moderate-intensity activity.

+

p <. 10.

*

p < .05.

**

p < .01.

***

p < .001.

Within-Day Prospective Analyses

The within-day prospective effect of patients’ morning self-efficacy on physical activity throughout that day was first examined in a random intercept, random slope model that controlled for average self-efficacy and day of study (Table 3, Model 1). As hypothesized, morning self-efficacy significantly predicted both steps and minutes in moderate-intensity activity within the same day, suggesting that on days when patients were one unit more confident than usual in their ability to be active during the day (on a 10-point scale), they took approximately 167 more steps and spent 2.4 more minutes in moderate-intensity activity that day. The between-person effect of self-efficacy was also significant for both steps and moderate-intensity activity, indicating that people who reported one point higher self-efficacy than other people on average took 275.25 more steps and spent 3.25 more minutes in moderate-intensity activity every day than their counterparts.

Table 3.

Within-Day Prospective Models Predicting Daily Steps and Moderate Activity from Morning Self-Efficacy

Fixed Effects Model 1
Model 2
Model 3
Estimate SE Estimate SE Estimate SE
Daily Steps
Intercept 2230.35*** (642.12) 2159.73       (1835.46) 2020.29       (1838.62)
Daily self-efficacy(WP) 166.97*** (38.84) 166.71*** (38.80) 155.34*** (42.37)
Mean self-efficacy(BP) 275.25**   (91.90) 173.42*     (74.10) 178.73*     (73.50)
Day of study −4.65       (6.73) −4.75       (6.68) −4.34       (6.83)
Age −60.97*** (16.99) −60.13*** (16.94)
Education 160.94*     (80.61) 163.95*     (80.87)
Employment status 835.03*     (347.31) 842.70*     (348.72)
Health status 321.20*     (144.16) 319.21*     (144.35)
Daily positive affect 4.71       (56.99)
Daily negative affect 18.11       (80.43)
Daily pain −375.73*     (171.50)
Daily spouse support 213.65       (134.46)
Daily spouse persuasion −264.78       (189.34)
Daily spouse pressure 96.79       (226.54)
Minutes in Moderate-Intensity Activity
Intercept 38.80**   (12.04) 117.45*** (32.78) 114.01*** (32.69)
Daily self-efficacy(WP) 2.40**   (0.73) 2.40*** (0.73) 2.52**   (0.82)
Mean self-efficacy(BP) 3.25*     (1.64) 2.00       (1.37) 2.23       (1.34)
Day of study −0.08       (0.14) −0.09       (0.14) −0.12       (0.15)
Age −1.93*** (0.28) −1.93*** (0.28)
Education 3.59*     (1.43) 3.65*     (1.44)
Daily positive affect 0.79       (0.99)
Daily negative affect 1.08       (1.62)
Daily pain −4.72       (4.51)
Daily spouse support 3.83       (2.87)
Daily spouse persuasion −5.34       (3.64)
Daily spouse pressure −1.65       (6.29)
Open in a new tab

Note. WP=within-person effect; BP=between-person effect.

*

p < .05.

**

p < .01.

***

p < .001.

In the second model (Model 2), background characteristics that were significantly related to steps or moderate-intensity activity in preliminary bivariate analyses were added into the model as level 2 covariates. Specifically, patients’ age, education level, employment status and self-reported health status were added into the model predicting steps whereas patients’ age and education level were added into the model predicting minutes in moderate-intensity activity. As shown in Table 3, the positive within-person prospective effects of morning self-efficacy remained significant in the model, supporting the unique predictive effects of patients’ morning self-efficacy on their daily physical activity above and beyond the effects of patients’ demographics. The between-person effects of self-efficacy remained significant for daily steps but not for minutes in moderate-intensity activity. Coefficients for the level 2 covariates showed that participants who were younger, employed, more educated and healthier took more steps every day compared with their counterparts. In addition, participants who were younger and more educated spent more time in moderate-intensity activity on average than their counterparts.

The final model (Model 3) added the within-person effects of patients’ daily pain, positive and negative affect, as well as spouses’ autonomy support, persuasion and pressure control. As shown, the positive within-person effects of morning self-efficacy on daily steps and minutes in moderate-intensity activity remained significant, providing the strongest support to the unique contribution of patients’ morning self-efficacy. The between-person effect of self-efficacy remained significant for daily steps but not for moderate-intensity activity after controlling for level 1 and level 2 covariates. Among all level 1 covariates, only patients’ pain reported in the morning had a negative impact on the steps taken the same day, suggesting that on days when patients felt one more unit of pain than usual (on a 4-point scale), they took 376 fewer steps.

Cross-Day Lagged Analyses

In order to explore whether patients’ self-efficacy would influence their physical activity the next day, lagged analyses were conducted in which patients’ self-efficacy on day t-1 predicted steps and moderate-intensity activity on day t. In each lagged model, day t’s self-efficacy and day of study were controlled. The results of these models suggested that morning self-efficacy did not significantly predict the next day’s steps (b=14.05, p=0.764) or minutes in moderate-intensity activity (b=0.46, p=0.657). We also conducted lagged analyses in which the within-person centered steps and minutes in moderate-intensity activity on day t-1 predicted day t morning’s self-efficacy. The results suggested that the predictive effect of day (t-1)’s physical activity was not significant for day t morning’s self-efficacy (b=−0.0003, p=0.151 for steps, b=0.001, p=0.373 for minutes in moderate-intensity activity).

Given the fact that both morning self-efficacy and pain had significant effects on subsequent physical activity within the same day, further analyses were conducted to examine within-person associations between self-efficacy, pain and physical activity. The results of the multilevel model suggested that the interaction between morning self-efficacy and pain did not significantly predict subsequent physical activity within the same day or the next day. These findings support the interpretation that on days when patients felt more confident in their ability to be active, they indeed took more steps and spent more time in moderate-intensity activity despite their level of pain that day.

Discussion

Physical activity is critical for the management of knee osteoarthritis. A major goal of this daily dairy study was to examine how OA patients’ morning self-efficacy to engage in physical activity despite pain impacted their physical activity throughout the day and across days. Our findings revealed that patients’ morning self-efficacy for physical activity had a significant positive effect on their steps and moderate-intensity activity throughout that day, above and beyond the effects of patients’ demographics, health, and other psychosocial factors such as daily pain, positive and negative affect as well as spouses’ activity-related autonomy support, persuasion and pressure control. However, the cross-day results suggested that the impact of patients’ self-efficacy did not extend to their next day physical activity. Similarly, patients’ physical activity did not have a cross-day lagged effect on their self-efficacy the next day. Finally, consistent with previous research, we found positive between-person effect of self-efficacy on physical activity, suggesting that OA patients who reported higher average level of self-efficacy over 22-day study period than others also engaged in more physical activity than their counterparts over the same period.

Theoretical and Practical Implications

To our knowledge, this is the first diary study to examine the role of self-efficacy in influencing OA patients’ daily physical activity assessed objectively in their natural environment. One of the most important findings of this study was that, the within-day prospective effect of morning self-efficacy on physical activity persisted even after adjustment for patients’ demographics, baseline health status, daily pain, affect and spouses’ support and control. This finding suggests that patients’ self-efficacy to engage in physical activity despite their pain is a particular important proximal antecedent of their actual physical activity on a daily basis. The clear temporal order established between self-efficacy and physical activity in the current study not only provides robust empirical evidence for Social-Cognitive Theory regarding the role of self-efficacy in enacting positive health behavior changes (Bandura, 1997), but also suggests that future physical activity interventions for people with OA should target their self-efficacy as a proximal psychological mechanism. Previous intervention studies have shown that manipulation of self-efficacy results in increased physical activity among healthy adults (Williams & French, 2011) and older adults (Allison & Keller, 2004). However, the effect of self-efficacy on promoting physical activity has not been established for people with OA, especially older adults with OA.

Recent reviews that assessed effectiveness of OA interventions to promote physical activity (e.g., Oliveira et al., 2016; Williamson et al., 2015) found small or no significant effects. As noted by Williamson et al (2015) in their review, even though some interventions included self-management elements that aimed to improve self-efficacy, the majority of them failed to assess changes in participants’ self-efficacy, as well as their impact on physical activity. Thus, it is difficult to evaluate the role of self-efficacy as an active ingredient within these interventions. One exception of which we are aware is Hughes et al.’s (2006) intervention study which was explicitly designed to enhance self-efficacy for exercise and maintenance of physical activity among older adults with OA and found significant treatment effect. Consistent with this study, our findings also support that self-efficacy should be an essential component of future physical activity interventions for older adults with OA. In our study, on days when patients felt slightly more confident than usual in the morning (only one unit higher on a 10-point scale), they took approximately 167 more steps and spent 2.4 more minutes in moderate activity that day. Although the increases in physical activity per day may seem to be of low magnitude, if accumulated over time, such daily changes would significantly improve the health for older adults with OA given the importance of physical activity for OA management and health.

The results of our study also have implications for optimal communication strategies of future physical activity interventions. Consistent with previous research (Basen-Engquist et al., 2013), the cross-day lagged analysis results from our study suggested that the positive effect of OA patients’ morning self-efficacy on subsequent physical activity did not extend to the next day. The lack of cross-day “carry-over” effect of self-efficacy on physical activity highlights the importance of sustaining the level of self-efficacy on a daily basis as a way to promote physical activity. In other words, timely bolstering of OA patients’ confidence in their capability to be active may be the key strategy to successfully promote patients’ daily physical activity. The development of mobile technology (e.g., mobile phone, tablet, PDA) provides more opportunities to deliver such just-in-time feedback and intervention to patients in their daily lives. The high compliance rate for daily assessments (>92%) in our study indicates the success of using personal computer and accelerometer to capture patients’ psychological status and physical activity in their natural settings. Recent studies used an ecological momentary intervention (EMI) that combined accelerometer with text messages delivered to people in their everyday life also found evidence to support the feasibility, acceptability and effectiveness of real-time interventions to promote physical activity among healthy and overweight populations (Berli, Stadler, Inauen, & Scholz, 2016; O’Reilly & Spruijt-Metz, 2013). Therefore, more research is needed to explicitly evaluate whether EMIs could increase OA patients’ self-efficacy and as a result promote physical activity.

Our findings have high public health significance, as it is estimated that nearly half of U.S. adults will develop symptomatic knee OA in their lifetime (Murphy et al., 2008). Physical activity is an especially important self-management tool for arthritis, as well as common comorbid conditions such as diabetes and heart disease, as it reduces the chance of recurrent health events and improves long-term functioning (Feinglass et al., 2005; Kodama et al., 2013). Older adults who maintain their physical function by staying active are better able to carry out activities of daily living independently and delay institutionalization. Moreover, improved self-management of chronic illnesses such as knee OA can reduce health care costs (Panagioti et al., 2014).

Limitations and Future Directions

Despite the important theoretical and practical implications and methodological strengths, the current study also has several limitations. First, although the current study established a temporal link between self-efficacy and subsequent physical activity, such evidence is still correlational in nature. To establish causal evidence for the role of self-efficacy as an underlying psychological mechanism for physical activity, future research should employ an experimental manipulation of self-efficacy. Second, due to the characteristics of our sample (older adults with OA), the generalizability of the results to other populations may be limited. Third, self-efficacy for physical activity was only assessed once per day in our study, and it is important for future research to better understand the fluctuations in self-efficacy within a day, and to identify factors that may influence such fluctuations. For example, the OA patients’ morning self-efficacy for physical activity was significantly correlated with morning pain (r = −0.16), positive affect (r = 0.21) and negative affect (r = −0.22) at the within-person level, suggesting that OA patients’ self-efficacy for physical activity was linked with their current physical and emotional states. Fourth, future studies should further explore the interplay between self-efficacy and other psychosocial and environmental factors, and how they impact patients’ physical activity jointly over time. Even though our study highlighted the unique effect of self-efficacy on subsequent physical activity above and beyond the effects of other factors by evaluating them simultaneously in the same model, additional factors are likely to influence physical activity. Partner’s autonomy support, for example, has been shown to have significant influence on OA patients’ physical activity when examined alone (Martire et al., 2013). Moreover, self-efficacy and partner’s autonomy support may influence physical activity jointly, so that patients may not be physically active on days when their self-efficacy was higher than usual but partner’s autonomy support was lower than usual.

Conclusions

The current study used a daily assessment approach and an objective behavioral measure to examine within- and between-person effects of self-efficacy for physical activity in the morning on subsequent physical activity throughout the day, and demonstrated this temporal link between self-efficacy and physical activity in an ecologically valid setting. This study makes a novel contribution to our understanding of a key psychological factor that could help to improve physical activity among people with OA, and is a crucial step for the development of future interventions that target this modifiable psychological mechanism.

Acknowledgments

This research was supported by NIH grants R01 AG026010 and K02 AG039412.

Footnotes

1

We also tested all models by using 1566 activity counts per minute or greater as the cut point of moderate-intensity activity (Miller, Strath, Swartz, & Cashin, 2010), and the effects of daily morning self-efficacy on the minutes spent in moderate-intensity activity (within-day and cross-day) remained statistically significant.

References

  1. Allison MJ, Keller C. Self-efficacy intervention effect on physical activity in older adults. Western Journal of Nursing Research. 2004;26:31–46. doi: 10.1177/0193945903259350. [DOI] [PubMed] [Google Scholar]
  2. Bandura A. Self-Efficacy: The Exercise of Control. New York: W. H. Freeman and Company; 1997. [Google Scholar]
  3. Basen-Engquist K, Carmack C, Li Y, Brown J, Jhingran A, Highes DC, Waters A. Social-Cognitive Theory Predictors of Exercise Behavior in Endometrial Cancer Survivors. Health Psychology. 2013;32:1137–1148. doi: 10.1037/a0031712. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Berli C, Stadler G, Inauen J, Scholz U. Action control in dyads: A randomized controlled trial to promote physical activity in everyday life. Social Science & Medicine. 2016;163:89–97. doi: 10.1016/j.socscimed.2016.07.003. [DOI] [PubMed] [Google Scholar]
  5. Bieler T, Siersma V, Magnusson SP, Kjaer M, Christensen HE, Beyer N. In hip osteoarthritis, Nordic Walking is superior to strength training and home-based exercise for improving function. Scandinavian Journal of Medicine & Science in Sports. 2016 doi: 10.1111/sms.12694. Advance online publication. [DOI] [PubMed] [Google Scholar]
  6. Brach JS, FitzGerald S, Newman AB, Kelsey S, Kuller L, VanSwearingen JM, Kriska AM. Physical activity and functional status in community-dwelling older women: A 14-year prospective study. Archives of Internal Medicine. 2003;163:2565–2571. doi: 10.1001/archinte.163.21.2565. [DOI] [PubMed] [Google Scholar]
  7. Brach JS, Wert D, Van Swearingen JM, Studenski SA. The Compendium of Physical Activity underestimates walking intensity in old more so than in young. Journal of the American Geriatrics Society. 2009;57:S110. [Google Scholar]
  8. Centers for Disease Control and Prevention (CDC) Osteoarthritis (OA) 2016 Retrieved from http://www.cdc.gov/arthritis/basics/osteoarthritis.htm.
  9. David P, Pennell ML, Foraker RE, Katz ML, Buckworth J, Paskett ED. How are previous physical activity and self-efficacy related to future physical activity and self-efficacy? Health Education & Behavior. 2014;41:573–576. doi: 10.1177/1090198114543004. [DOI] [PubMed] [Google Scholar]
  10. Downs A, Van Hoomissen J, Lafrenz A, Julka DL. Accelerometer-measured versus self-reported physical activity in college students: Implications for research and practice. Journal of American College Health. 2014;62:204–212. doi: 10.1080/07448481.2013.877018. [DOI] [PubMed] [Google Scholar]
  11. Dunton GF, Atienza AA, Castro CM, King AC. Using ecological momentary assessment to examine antecedents and correlates of physical activity bouts in adults age 50+ years: a pilot study. Annals of Behavioral Medicine. 2009;38(3):249–255. doi: 10.1007/s12160-009-9141-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Feinglass J, Thompson JA, He XZ, Witt W, Chang RW, Baker DW. Effect of physical activity on functional status among older middle-age adults with arthritis. Arthritis Care & Research. 2005;53:879–885. doi: 10.1002/art.21579. [DOI] [PubMed] [Google Scholar]
  13. Focht BC. Effectiveness of exercise interventions in reducing pain symptoms among older adults with knee osteoarthritis: A review. Journal of Aging and Physical Activity. 2006;14:212–235. doi: 10.1123/japa.14.2.212. [DOI] [PubMed] [Google Scholar]
  14. Gay C, Chabaud A, Guilley E, Coudeyre E. Educating patients about the benefits of physical activity and exercise for their hip and knee osteoarthritis. Systematic literature review. Annals of Physical and Rehabilitation Medicine. 2016;59:174–183. doi: 10.1016/j.rehab.2016.02.005. [DOI] [PubMed] [Google Scholar]
  15. Hughes SL, Seymour RB, Campbell RT, Huber G, Pollak N, Sharma L, Desai P. Long-term impact of Fit and Strong! on older adults with osteoarthritis. The Gerontologist. 2006;46:801–814. doi: 10.1093/geront/46.6.801. [DOI] [PubMed] [Google Scholar]
  16. Kodama S, Tanaka S, Heianza Y, Fujihara K, Horikawa C, Shimano H, Sone H. Association between physical activity and risk of all-cause mortality and cardiovascular disease in patients with diabetes: a meta-analysis. Diabetes Care. 2013;36:471–479. doi: 10.2337/dc12-0783. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Liao Y, Shonkoff ET, Dunton GF. The acute relationships between affect, physical feeling states, and physical activity in daily life: A review of current evidence. Frontiers in Psychology. 2015;6:1975. doi: 10.3389/fpsyg.2015.01975. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Maas CJM, Hox JJ. Robustness issues in multilevel regression analysis. Statistica Neerlandica. 2004;58:127–137. [Google Scholar]
  19. Martire LM, Stephens MAP, Mogle J, Schulz R, Brach J, Keefe FJ. Daily spousal influence on physical activity in knee osteoarthritis. Annals of Behavioral Medicine. 2013;45:213–223. doi: 10.1007/s12160-012-9442-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Martire LM, Wilson SJ, Small BJ, Conley YP, Janicki PK, Sliwinskia MJ. COMT and OPRM1 genotype associations with daily knee pain variability and activity induced pain. Scandinavian Journal of Pain. 2016;10:6–12. doi: 10.1016/j.sjpain.2015.07.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Mason JH, Anderson J, Meenan R, Haralson K, Lewis-Stevens D, Kaine J. The rapid assessment of disease activity in rheumatology (RADAR) questionnaire. Arthritis & Rheumatism. 1992;35:156–162. doi: 10.1002/art.1780350206. [DOI] [PubMed] [Google Scholar]
  22. Matthews CE. Calibration of accelerometer output for adults. Medicine and Science in Sports and Exercise. 2005;37(11):S512–S522. doi: 10.1249/01.mss.0000185659.11982.3d. [DOI] [PubMed] [Google Scholar]
  23. Matthews CE, Ainsworth BE, Hanby C, Pate RR, Addy C, Freedson PS, Macera CA. Development and testing of a short physical activity recall questionnaire. Medicine and Science in Sports and Exercise. 2005;37(6):986–994. [PubMed] [Google Scholar]
  24. McAuley E. The role of efficacy cognitions in the prediction of exercise behavior in middle-aged adults. Journal of Behavioral Medicine. 1992;15:65–88. doi: 10.1007/BF00848378. [DOI] [PubMed] [Google Scholar]
  25. McAuley E, Mullen SP, Szabo AN, White SM, Wójcicki TR, Mailey EL, Kramer AF. Self-regulatory processes and exercise adherence in older adults. American Journal of Preventive Medicine. 2011;41:284–290. doi: 10.1016/j.amepre.2011.04.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Miller NE, Strath SJ, Swartz AM, Cashin SE. Estimating Absolute and Relative Physical Activity Intensity Across Age via Accelerometry in Adults. Journal of Aging and Physical Activity. 2010;18(2):158–170. doi: 10.1123/japa.18.2.158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Moreau K, Degarmo R, Langley J, Mcmahon C, Howley ET, Bassett DR, Thompson DL. Increasing daily walking lowers blood pressure in postmenopausal women. Medicine & Science in Sports and Exercise. 2001;11:1825–1831. doi: 10.1097/00005768-200111000-00005. [DOI] [PubMed] [Google Scholar]
  28. Murphy SL, Niemiec SS, Lyden AK, Kratz AL. Pain, fatigue, and physical activity in osteoarthritis: The moderating effects of pain- and fatigue-related activity interference. Archives of Physical Medicine and Rehabilitation. 2016;9:201–209. doi: 10.1016/j.apmr.2015.05.025. [DOI] [PubMed] [Google Scholar]
  29. Murphy SL, Smith DM, Clauw DJ, Alexander NB. The impact of momentary pain and fatigue on physical activity in women with osteoarthritis. Arthritis & Rheumatism. 2008;59:849–856. doi: 10.1002/art.23710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Oliveira CB, Franco MR, Maher CG, Lin CC, Morelhão PK, Araujo AC, Pinto RZ. Physical activity interventions for increasing objectively measured physical activity levels in chronic musculoskeletal pain: systematic review. Arthritis Care & Research. 2016 doi: 10.1002/acr.22919. Advance online publication. [DOI] [PubMed] [Google Scholar]
  31. O’Reilly GA, Spruijt-Metz D. Current mHealth technologies for physical activity assessment and promotion. American Journal of Preventive Medicine. 2013;45:501–507. doi: 10.1016/j.amepre.2013.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Panagioti M, Richardson G, Small N, Murray E, Rogers A, Kennedy A, Bower P. Self-management support interventions to reduce health care utilisation without compromising outcomes: a systematic review and meta-analysis. BMC Health Services Research. 2014;14:356. doi: 10.1186/1472-6963-14-356. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Peeters G, Brown WJ, Burton NW. Psychosocial factors associated with increased physical activity in insufficiently active adults with arthritis. Journal of Science and Medicine in Sport. 2015;18:558–564. doi: 10.1016/j.jsams.2014.08.003. [DOI] [PubMed] [Google Scholar]
  34. Pisters MF, Veenhof C, Schellevis FG, Twisk JW, Dekker J, De Bakker DH. Exercise adherence improving long-term patient outcome in patients with osteoarthritis of the hip and/or knee. Arthritis Care and Research. 2010;62:1087–94. doi: 10.1002/acr.20182. [DOI] [PubMed] [Google Scholar]
  35. Pisters MF, Veenhof C, van Dijk GM, Dekker J, CARPA Study Group Avoidance of activity and limitations in activities in patients with osteoarthritis of the hip or knee: a 5-year follow-up study on the mediating role of reduced muscle strength. Osteoarthritis Cartilage. 2014;22(2):171–177. doi: 10.1016/j.joca.2013.12.007. [DOI] [PubMed] [Google Scholar]
  36. Singer JD, Willett JB. Applied longitudinal data analysis: Modeling change and event occurrence. Oxford, England: Oxford University Press; 2003. [Google Scholar]
  37. Spencer NH. Combining modelling strategies to analyse teaching styles data. Quality and Quantity. 2002;36:113–127. [Google Scholar]
  38. Stubbs B, Hurley M, Smith T. What are the factors that influence physical activity participation in adults with knee and hip osteoarthritis? A systematic review of physical activity correlates. Clinical Rehabilitation. 2015;29:80–94. doi: 10.1177/0269215514538069. [DOI] [PubMed] [Google Scholar]
  39. Teixeira PJ, Carraca EV, Marques MM, Rutter H, Oppert JM, De Bourdeaudhuij I, Brug J. Successful behavior change in obesity interventions in adults: a systematic review of self-regulation mediators. BMC Medicine. 2015;13:1–16. doi: 10.1186/s12916-015-0323-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Thomas DL, Diener E. Memory accuracy in the recall of emotions. Journal of Personality and Social Psychology. 1990;59:291–297. [Google Scholar]
  41. US Department of Health and Human Services. Physical Activity Guidelines Advisory Committee Report. Washington, DC: 2008. [Google Scholar]
  42. Veenhof C, Huisman PA, Barten JA, Takken T, Pisters MF. Factors associated with physical activity in patients with osteoarthritis of the hip or knee: a systematic review. Osteoarthritis Cartilage. 2012;20:6–12. doi: 10.1016/j.joca.2011.10.006. [DOI] [PubMed] [Google Scholar]
  43. Vancampfort D, Sienaert P, Wyckaert S, De Hert M, Stubbs B, Soundy A, Probst M. Health-related physical fitness in patients with bipolar disorder versus healthy controls: an exploratory study. Journal of Affective Disorders. 2015;177:22–27. doi: 10.1016/j.jad.2014.12.058. [DOI] [PubMed] [Google Scholar]
  44. Wallis JA, Webster KE, Levinger P, Taylor NF. What proportion of people with hip and knee osteoarthritis meet physical activity guidelines? A systematic review and meta-analysis. Osteoarthritis and Cartilage. 2013;21:1648–1659. doi: 10.1016/j.joca.2013.08.003. [DOI] [PubMed] [Google Scholar]
  45. Williams SL, French DP. What are the most effective intervention techniques for changing physical activity self-efficacy and physical activity behavior—and are they the same? Health Education Research. 2011;26:308–322. doi: 10.1093/her/cyr005. [DOI] [PubMed] [Google Scholar]
  46. Williamson W, Kluzek S, Roberts N, Richards J, Arden N, Leeson P, Foster C. Behavioural physical activity interventions in participants with lower-limb osteoarthritis: a systematic review with meta-analysis. BMJ Open. 2015;5:e007642. doi: 10.1136/bmjopen-2015-007642. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES