The Relationship between Self-Efficacy, Functional Exercise Capacity and Physical Activity in People with COPD: A Systematic Review and Meta-Analyses

Abstract

The purpose of this study was to investigate the strength of the relationships between self-efficacy and (i) functional exercise capacity and (ii) physical activity in chronic obstructive pulmonary disease (COPD), and whether self-efficacy assessment type (i.e., COPD symptoms, exercise-task, exercise-barrier, general, falls) and physical activity assessment type (i.e., self-report vs. objective) are moderators. A systematic search of COPD and self-efficacy concepts was conducted using eight databases from inception to 23 January 2019. Studies were included if they provided correlation coefficients of the relationship between self-efficacy and functional exercise capacity or physical activity, were conducted in adults diagnosed with COPD, and were published in English-language journals. A total of 14 correlation coefficients were included in the self-efficacy and functional exercise capacity meta-analysis, and 16 in the self-efficacy and physical activity meta-analysis. Data were screened, reviewed, and extracted independently by two reviewers, with discrepancies resolved by a third reviewer. Stronger self-efficacy was associated with better functional exercise capacity (weighted r = 0.38, 95%CI [0.25, 0.50]), and greater physical activity (weighted r = 0.25, 95%CI [0.17, 0.34]). Exercise-task self-efficacy had the strongest relationship to functional exercise capacity (weighted r = 0.64, 95% CI [0.51, 0.73]). For physical activity, the type of self-efficacy most strongly related was inconclusive. In COPD, self-efficacy has a relationship to functional exercise capacity and physical activity, the strength of which is influenced by the choice of self-efficacy measure. An understanding of these relationships will assist clinicians in selecting the self-efficacy measure most closely related to the outcome of interest.

Keywords:

• Chronic obstructive pulmonary disease
• self-efficacy
• physical activity
• functional exercise capacity

Introduction

Chronic obstructive pulmonary disease (COPD) is a common respiratory disorder predominantly caused by smoking. According to the World Health Organization, it is estimated to be the third leading cause of death globally by 2030 [1]. The disease is characterized by progressive worsening of symptoms, including dyspnea on exertion, cough, fatigue, and infectious exacerbations [2]. COPD is also associated with secondary manifestations including skeletal muscle impairment, mood disorders, hormonal imbalances, and osteoporosis [3]. Pulmonary rehabilitation (PR) composed of exercise training, education, and self-management curricula is recommended as part of standard care for COPD [4]. The PR components target the secondary manifestations of COPD to improve dyspnea, functional exercise capacity, and health-related quality of life (HRQoL) [3].

Functional exercise capacity [4, 5] is one of the key outcomes of PR and is commonly measured by walking tests. Greater functional exercise capacity is associated with decreased mortality and hospitalizations in COPD [6]. While PR is consistently shown to increase functional exercise capacity in COPD [4], many people do not maintain the improvements they achieve post-PR [7]. To sustain increases in functional exercise capacity, regular physical activity is required.

Physical activity is defined as “any bodily movement produced by skeletal muscles that results in energy expenditure beyond resting energy expenditure” [8]. In a systematic review, low levels of physical activity were consistently related to mortality and hospitalizations in people with COPD [9]. Physical activity levels among individuals with COPD are lower than healthy age-matched counterparts [10–12] and lower than individuals with other chronic conditions, including cardiovascular disease and diabetes [13]. There are numerous social, environmental, and individual variables that can impact physical activity levels. In COPD, there is evidence that functional exercise capacity, health-related quality of life, dyspnea, previous exacerbations, lung hyperinflation, gas exchange, and self-efficacy are all related to physical activity [9].

Self-efficacy is defined as beliefs in one’s capability to exert control over personal and situation events required to produce behavioral accomplishments [14]. In theories of chronic disease management, self-efficacy influences adherence to health enhancing behaviors, which in turn improves physical functioning [15, 16]. People with stronger self-efficacy for a behavior are more likely to pursue it, dedicate effort to it, and persevere in the face of barriers than people with weaker self-efficacy [14]. However, the relationship between self-efficacy and behavior is likely bidirectional, as direct personal experiences can influence self-efficacy. Successful personal experiences enhance self-efficacy and failures reduce it [14]. Selzler et al. [17] noted that over a 12-month period, prospective relationships between self-efficacy and attendance at an exercise program were bidirectional at five measurement time points.

Among people with COPD, self-efficacy has been found to be positively related to functional exercise capacity [18–20] and physical activity [21–23]. There is some variability in the size of the relationship reported, and in most studies, the sample size has been small. A more robust estimate of the relationship between self-efficacy, functional exercise capacity, and physical activity, as well as the identification of factors that impact this relationship will inform intervention targets and health programs in COPD. For example, if self-efficacy has at least a small-to-moderate relationship to functional exercise capacity and/or physical activity, it will indicate that self-efficacy is a useful target for researchers and clinicians aiming to improve these outcomes. Factors that impact the relationship among self-efficacy, functional exercise capacity, and physical activity will provide further detail on how these outcomes can be improved.

According to Bandura [14], self-efficacy is not a global assessment of one’s capabilities, but rather a behavior-specific set of capability beliefs. For example, an individual can be very confident that they can walk 500 meters but simultaneously very unconfident that they can swim 500 meters. Bandura [14] also contends that end-goal behavioral performance is a result of mastering elemental aspects of the task in addition to performing the task under challenging circumstances. So, to walk 500 meters without stopping, an individual must develop self-efficacy for physically walking the distance, in addition to self-efficacy for coping with challenges of physical exertion and practicing the behavior, such as pacing, managing breathing, and maintaining motivation. Thus, the relationship between self-efficacy and a given behavior will depend on the type of self-efficacy assessed and the nature of the behavior, with the most salient self-efficacy type having the strongest association.

Several types of self-efficacy have been assessed in COPD, including self-efficacy for managing COPD symptoms (COPD self-efficacy), self-efficacy for exercise tasks, self-efficacy for exercise barriers, self-efficacy for falling, and self-efficacy for general problem solving. To achieve functional exercise capacity and physical activity outcomes, some degree of physical abilities for exercise tasks and corresponding self-efficacy is required. However, other types of self-efficacy may also be important, such as self-efficacy for managing symptoms and self-efficacy for coping with exercising barriers (e.g., weather, discomfort, not feeling like it). It is unclear which type of self-efficacy is most important to the performance of functional exercise capacity and physical activity in COPD.

The primary aim of this study was to synthesize the studies on and estimate the relationships between self-efficacy and (i) functional exercise capacity and (ii) physical activity in people with COPD. The secondary aim of this study was to determine the moderating role of different types of self-efficacy (i.e., exercise task, exercise barrier, COPD symptoms, general, and falls self-efficacy) on these relationships, and physical activity type (i.e., self-report vs. objective) in the physical activity meta-analysis.

Methods

The meta-analysis protocols were registered on PROSPERO (CRD42018114846, CRD42019137172) and the PRISMA framework guided this research.

Search details

The following electronic databases were used to search for research articles examining the use of self-efficacy scales in people with COPD: Ovid MEDLINE: Epub Ahead of Print, In-Process & Other Non-Indexed Citations, Ovid MEDLINE Daily and Ovid MEDLINE, OVID Embase, Ovid PsycINFO, EBSCO CINAHL Plus, and EBSCO SportDiscus. The search strategies were translated using each database platform’s command language, controlled vocabulary, and appropriate search fields. MeSH terms, EMTREE terms, APA thesauri terms, CINAHL headings, SIRC terms, and textwords were used for the search concepts of COPD and self-efficacy. No date or language limits were imposed. Final searches were completed 23 January 2019. For full search strategy, see supplementary material.

Inclusion and exclusion criteria

Studies were included in the review if: (a) they measured self-efficacy and functional exercise capacity or physical activity, (b) they provided correlation coefficients (Pearson’s r) of the association between self-efficacy and physical activity or functional exercise capacity or reported statistics that could be converted into Pearson’s r (e.g., odds ratios), (c) they were conducted among adults diagnosed with COPD, and (d) they were published in English-language journals. Studies in any setting were eligible for inclusion, including, but not limited to acute care, pulmonary rehabilitation (inpatient, outpatient, home-based), and community programs. If a study included patients with various chronic conditions, the study was only included if people with COPD were represented in a distinct group.

Studies were excluded from the review if they were (a) qualitative, (b) not published in peer-reviewed journals, (c) systematic or nonsystematic reviews, (d) editorials, (e) did not report any association between self-efficacy and physical activity or functional exercise capacity, or (f) contained samples that combined COPD and other respiratory conditions. No restrictions were made regarding the duration of illness.

Study selection and data extraction

The results of the database search were imported to EndNote and then Covidence, which was used to track and manage the review process. At least two researchers (R.H., V.M., and A.S.) were involved at every stage of data screening and reviewing, extraction, and evaluation. Any disagreements were resolved by a consensus method (discussion followed by third researcher resolution).

Titles and abstracts were screened independently by two reviewers to identify potentially relevant articles. Subsequently, the full-text articles were reviewed for inclusion. Studies that met the inclusion criteria in the full-text screening phase were retained for data extraction. The reference lists of studies that were retained were searched to identify additional potentially relevant studies.

All data were extracted twice, by two authors. Descriptive data were extracted from each study, including study design, country where the study was conducted, sample size, proportion of female participants, participants mean age, smoking history, spirometry, dyspnea rating (i.e., modified MRC score), main purpose of the study, main inclusion criteria of participants, characteristics of the self-efficacy, physical activity/functional exercise capacity measures, and indicators of the association between self-efficacy and physical activity/functional exercise capacity.

Quality assessment

To assess the risk of bias, the Standard Quality Assessment Criteria for Evaluating Primary Research Papers from a Variety of Fields [24] was used. Two reviewers independently scored the studies according to the criteria of this scale, and a mean quality score was calculated for each article by taking the total score divided by the possible score and multiplied by 100. Only studies that met the 65% threshold were included for analysis, as recommended by Kmet et al. [24].

Definitions of variables and coding

For the functional exercise capacity and physical activity meta-analyses, the following self-efficacy types were coded to examine the moderating effects of self-efficacy type on the relationships between self-efficacy and (i) functional exercise capacity and (ii) physical activity: general, COPD symptom, exercise task, exercise barrier, and falls. Table 1 lists the types of self-efficacy, definitions, coding parameters, and measurements included in the meta-analysis. For the physical activity meta-analysis, physical activity measures were coded as either self-reported or objective to examine the moderating effects of the physical activity measurement type.

Table 1. List of self-efficacy types, definitions, coding parameters, and measures.

Outcome	N correlation coefficients FEC/PA	Coding parameters	Measures
Self-efficacy Type: Definition
General: Confidence that applies to many different behaviors and situations [51].	½	Items in measure referred to a general sense of capability, and did not explicitly refer to managing symptoms or performing exercise-related behaviors.	General Self-efficacy Scale (GSES) [51] General subscale of the Pulmonary Rehabilitation Adapted Index of Self-efficacy (PRAISE) as conceptualized by Song et al. [30] Self-efficacy subscale of Self-Management Abilities Scale (SMAS) [52]
COPD symptom: Confidence for managing and avoiding COPD symptoms like breathlessness [53].	7/3	Items in measure referred to one’s capability to manage COPD symptoms.	COPD Self-efficacy Scale (CSES) [53] Coping with breathing subscale of Modified Multidimensional Self-efficacy for Exercise Scale (MSES) [23] Self-efficacy for Managing Chronic Disease [54]
Exercise task: Confidence for performing physical movements like walking [55].	3/7	Items referred to one’s capability to perform movements related to physical activity or exercise, and did not mention management of COPD symptoms.	Walking Self-efficacy by Atkins et al. [56] Scales for walking, general exertion, moving things, lifting, and climbing stairs developed by Kaplan et al. [48] Self-efficacy Scale by Rejeski [27] Task subscale of MSES [55] Perceived physical abilities subscale of the Physical Self-efficacy Scale [57]
Exercise barrier: Confidence for exercising under challenging situations [44].	2/2	Items referred to one’s capability to be physically active under challenging circumstances or maintain regular activities, and did not mention the management of COPD symptoms.	Exercise self-efficacy subscale of PRAISE as identified by Song et al. [30] Perceived Exercise Self-efficacy Scale [58] Coping with Exercise and Schedule subscales of the MSES [55] Exercise Self-Regulatory Efficacy Scale (Ex-SRES) [35]
Falls: Fear of falling [59]	1/2	Items referred to concern about fall during a variety of activities.	Falls Efficacy Scale – International [59]

Note. FEC = functional exercise capacity; PA = physical activity.

Analysis

MetaXL software was used for analysis. A random effects model of correlation coefficients was used to estimate the population effect size. Funnel plots were computed to examine publication bias, and forest plots were used to illustrate the results. Heterogeneity was assessed using the Q and I² statistic.

Separate subsamples of participants (e.g., males and females) were treated as different studies when reported individually. When multiple correlation coefficients were provided for self-efficacy and one of the outcomes (functional exercise capacity or physical activity), the correlation coefficients were combined for that outcome using the procedure outlined by Schmidt and Hunter [25], but only if combining correlation coefficients did not mix construct conceptualizations. For example, if multiple correlation coefficients were provided for physical activity and several exercise task self-efficacy scales, the correlation coefficients were combined. But if separate correlation coefficients were provided for exercise barrier self-efficacy and exercise task self-efficacy, the correlation coefficients were not combined. In this instance, the study would contribute multiple correlation coefficients to the meta-analysis, which would be evaluated in the subgroup analysis. Subgroup analyses were conducted for self-efficacy conceptualizations (i.e., COPD symptom, exercise task, exercise coping, general, and falls) and physical activity measurement type (i.e., self-report, objective).

Results

Description of studies and content analyzed

Functional exercise capacity

The selection process resulted in 13 studies that met the inclusion criteria (Figure 1). The studies contributed 14 correlation coefficients, totaling 1,330 participants, to the meta-analysis, with sample sizes between 30 and 242. The characteristics of each study are outlined in Supplementary Table 2 The majority of studies were conducted in the United States [20, 22, 26–28], with the remaining studies conducted in Canada [29], South Korea [30], Nigeria [18], China [31], the United Kingdom [19], Norway [32], Australia [33], and Turkey [34]. The studies were conducted between 1997 and 2017, with the most common study design cross sectional (k = 7), followed by longitudinal (k = 4), experimental (k = 1), and chart review (k = 1). The studies were conducted in outpatient PR (k = 6), outpatient clinics (k = 3), inpatient PR (k = 1), and hybrid settings of home and outpatient clinic (k = 2), and outpatient clinic and university facilities (k = 1). Across studies the mean age was 67.3 (SD = 4.0), comprised of 41% female participants. The mean FEV₁% predicted was 45.5 (SD = 7.8, range = 36%–57%).

Figure 1. Study flow diagram for meta-analyses; FEC = functional exercise capacity; PA = physical activity.

Physical activity

The selection process resulted in 12 studies that met the inclusion criteria (Figure 1). The studies contributed 16 correlation coefficients (6 objective measures and 10 self-reported measures). There were 1,575 participants in the meta-analysis, with sample sizes between 15 and 253. The characteristics of each study are outlined in Supplementary Table 3. The majority of studies were conducted in the United States [22, 35–37], followed by the Netherlands [21, 38], Canada [23], South Korea [39], China [40], Belgium [41], Norway [42], and Australia [33]. The studies were conducted between 2000 and 2019, with most studies (k = 7) conducted within the last 8 years. The most common study design was cross sectional (k = 9), followed by longitudinal (k = 2), and experimental (k = 1). The studies were conducted in outpatient PR (k = 6), outpatient clinics (k = 2), community programs (k = 1), and hybrid settings of home and clinic-based PR maintenance programs (k = 2), outpatient clinics and outpatient PR (k = 1), outpatient clinic and university facilities (k = 1), and outpatient clinic, outpatient PR, and community programs (k = 1). Across studies the mean age was 66.6 (SD = 3.2), comprised of 37% female participants. The mean FEV₁% predicted was 49.6 (SD = 12.2, range = 36%–78%).

Results of meta-analysis

Functional exercise capacity

The funnel plot (Figure 2(a)) was approximately symmetrical, suggesting no publication bias, although several studies fell outside the triangular region of the plot. The shape of the funnel plot along with a larger I² statistic suggests variability in the relationship across studies, which may indicate the presence of a moderator or bias (e.g., measurement).

Figure 2. Funnel plots of standard error by Fisher’s z-transformed correlation. (a) Functional exercise capacity. (b) Physical activity.

The results of the meta-analysis including the estimates of the average effect, heterogeneity statistics, and subgroup analyses are presented in Figure 3. The estimate of the overall relationship between self-efficacy and functional exercise capacity was moderate (weighted r = 0.38, 95% CI [0.25, 0.50]), indicating that greater self-efficacy is associated with better functional exercise capacity.

In the subgroup analysis of self-efficacy type, exercise task self-efficacy had the strongest association with functional exercise capacity (weighted r = 0.64, 95% CI [0.51, 0.73]), followed by COPD symptom self-efficacy (weighted r = 0.31, 95% CI [0.16, 0.44]), falls self-efficacy (weighted r = 0.31, 95% CI [0.11, 0.36]), exercise barrier self-efficacy (weighted r = 0.24, 95% CI [0.11, 0.36]), and general self-efficacy (weighted r = 0.16, 95% CI [–0.02, 0.33]). The forest plot for the subgroup analysis of self-efficacy type is displayed in Figure 3.

Figure 3. Forrest plot of self-efficacy and functional exercise capacity relationship by self-efficacy subgroup.

Physical activity

The funnel plot (Figure 2(b)) was approximately symmetrical, with both larger and smaller studies represented above and below the meta-analysis average suggesting no publication bias.

The results of the meta-analysis including the estimates of the average effect, heterogeneity statistics, and subgroup analyses are presented in Figure 4. The estimate of the overall relationship between self-efficacy and physical activity was weak-to-moderate (weighted r = 0.25, 95% CI [0.17, 0.34]), indicating that greater self-efficacy is associated with more physical activity.

In the subgroup analysis of self-efficacy type, falls self-efficacy had the strongest association with physical activity (weighted r = 0.35, 95% CI [0.08, 0.57]), followed by exercise task self-efficacy (weighted r = 0.29, 95% CI [0.19, 0.39]), exercise barrier self-efficacy (weighted r = 0.29, 95% CI [0.03, 0.51]), general self-efficacy (weighted r = 0.26, 95% CI [0.06, 0.44]), and COPD symptom self-efficacy (weighted r = 0.12, 95% CI [–0.19, 0.41]). The forest plot for the subgroup analysis of self-efficacy type is displayed in Figure 4. In the subgroup analysis of physical activity type, objectively measured physical activity (weighted r = 0.25, 95% CI [0.10, 0.38]) had a similar association to self-efficacy as self-reported measured physical activity (weighted r = 0.26, 95% CI [0.15, 0.36]).

Figure 4. Forrest plot of self-efficacy and physical activity relationship by self-efficacy subgroup.

Discussion

This is the first study to provide a systematic synthesis of existing evidence regarding the relationship between self-efficacy and two key outcomes, functional exercise capacity, and physical activity in people with COPD. For self-efficacy and functional exercise capacity, the average estimated relationship was positive and moderate, and for self-efficacy and physical activity, it was positive and weak to moderate. Exercise task self-efficacy was the type of self-efficacy most strongly associated with functional exercise capacity. The type of self-efficacy most strongly associated with physical activity was inconclusive.

The relationship between self-efficacy and functional exercise capacity was notably stronger for exercise task self-efficacy compared to the other self-efficacy types. Exercise task self-efficacy is about one’s confidence for performing the physical movements involved in exercising [43], including walking. Thus, exercise task self-efficacy is highly relevant to the requirements of walking tests. While the subgroup analysis of exercise task self-efficacy only contained three correlation coefficients (374 participants), the confidence intervals were narrower than in the COPD symptom self-efficacy subgroup analysis of seven correlation coefficients (555 participants), suggesting that exercise task self-efficacy should be considered as an intervention target when assessing and aiming to improve functional exercise capacity in COPD.

There are several ways to enhance exercise task self-efficacy in an intervention. Exercise training provides opportunities to practice and have successful experiences (i.e., mastery experiences) with walking and other physical exercise tasks and is the best way to increase exercise task self-efficacy [14] and functional exercise capacity. Observing others of similar abilities, age, and gender have successful exercise experiences (i.e., vicarious experience) is another way to increase exercise task self-efficacy [14, 23]. For example, observing others with COPD in exercise videos may be a particularly useful way to increase exercise task self-efficacy if direct personal experience is limited due to practical restrictions such as physical space or personal reasons such as feeling intimidated or anxiety about exercising. Selzler et al. [23] found that observing a COPD patient of the same gender perform an exercise test increased observer self-efficacy as much as their direct personal experience with the exercise test. Another way to enhance exercise task self-efficacy is to provide words of encouragement about one’s ability to perform the exercise task (i.e., verbal persuasion), although mastery and vicarious experiences will have a stronger impact on self-efficacy [14]. Such words of encouragement could be delivered by health care professionals in formal counseling sessions and routine interactions during exercise, as well as from the individual in the form of self-affirmations.

While there was a moderate relationship between self-efficacy and functional exercise capacity, there was only a small-to-moderate relationship between self-efficacy and physical activity. Physical activity is a more variable measure than functional exercise capacity and is influenced by more personal, social, and situational factors [9]. As such, the relationship between self-efficacy and physical activity is expected to be weaker than the relationship between self-efficacy and functional exercise capacity.

The relationship between self-efficacy and physical activity was similar for self-reported measures and objective measures of physical activity. This finding suggests that both measurement types may be appropriate when assessing the self-efficacy and physical activity relationship in people with COPD. Self-reported measures have the added benefit of being cost and time effective.

While self-efficacy was positively related to physical activity, firm conclusions cannot be drawn from the moderator analysis of self-efficacy type because of a relatively small number of participants and wide confidence intervals for the estimated coefficients. In cardiac rehabilitation, self-efficacy for coping with exercise and scheduling barriers were stronger predictors of physical activity than self-efficacy for exercise tasks [44–46]. Similarly, in a study of COPD patients [23], Selzler et al. noted that self-efficacy for coping with exercise barriers had a stronger association with physical activity when compared to self-efficacy for exercise tasks, coping with breathing during exertion, and walking. In the current meta-analysis, the size of the coefficients for exercise task self-efficacy and exercise barrier self-efficacy was the same, although the estimate for exercise task self-efficacy was more reliable given that it encompassed 570 participants and had relatively narrow confidence intervals. We operationalized exercise barrier self-efficacy to include both coping and scheduling self-efficacy, which were assessed separately by Selzler et al. [23]. By combining these coefficients, the size of the relationship between exercise barrier self-efficacy and physical activity was reduced. In COPD, scheduling exercise may not be an exercise barrier given that most patients are retired. Future research examining the parameters of exercise barrier self-efficacy in COPD is needed.

When examining the relationships of self-efficacy types to various outcomes in COPD, there are similarities and differences in these relationships. Confidence for performing physical movements related to exercise was relevant to outcomes of functional exercise capacity, physical activity, and HRQoL [47], emphasizing the importance of exercise training as a key component of COPD management [4]. For physical activity and HRQoL, confidence for performing behaviors under challenging conditions is also important. However, the cumulative findings also highlight that the salient abilities to achieve physical activity and HRQoL outcomes are slightly different, indicating that interventions need to be specific to the desired outcome. For physical activity participation, the best available evidence indicates targeting skills and confidence for overcoming motivational and environmental barriers to exercise [23, 35]. The above are also important for HRQL, along with skills and confidence for managing and avoiding breathing difficulties [47].

Further, knowledge of the self-efficacy type most strongly associated with an outcome such as functional exercise capacity will assist clinicians in their choice of outcome measure. Knowing that exercise task self-efficacy had the strongest relationship to functional exercise capacity would have made it a more frequently chosen outcome rather than COPD symptom self-efficacy. To examine the relationship between functional exercise capacity and self-efficacy, an exercise task self-efficacy measure, such as self-efficacy for walking scales [48–50] or the multidimensional self-efficacy scale (MSES) [43], would be useful. The MSES assesses exercise task self-efficacy, exercise coping self-efficacy, and exercise scheduling self-efficacy with three items each. A modified version of the MSES [23] includes an additional three-item scale that assesses coping with breathing during exercise. The MSES is also appropriate when assessing the relationships of self-efficacy, physical activity, and HRQoL. Another scale for assessing the relationship between physical activity and self-efficacy in COPD is the Exercise Self-Regulatory Efficacy Scale [35].

Study limitations

Most of the studies were cross sectional, which prevents determining causality and the direction of the relationship between self-efficacy and these outcomes. However, cumulative theoretical and empirical evidence suggests that the relationships are likely bidirectional. Firm conclusions cannot be drawn from the subgroup analysis, particularly for physical activity because of the small number of studies in each group. Given that functional exercise capacity was examined in terms of the lower extremities only, the results cannot be generalized to functional capacity of the upper extremities. Similarly, our results do not generalize to laboratory measures of exercise capacity. While this meta-analysis controlled for sampling error and assessed study quality, other factors that could potentially impact the estimated correlation coefficients, such as measurement error, were not controlled for. As the authors determined the operational definitions for the self-efficacy types, the observations could differ if other operational definitions are used.

Conclusions

In COPD, self-efficacy is moderately associated with functional exercise capacity and weakly-to-moderately associated with physical activity. The strength of the association depends on the type of self-efficacy assessed, with self-efficacy for exercise tasks having the strongest relationship to functional exercise capacity. When considering the relationship between self-efficacy and functional exercise capacity, interventions and measures specific to exercise task self-efficacy should be utilized. The type of self-efficacy most strongly associated with physical activity remains to be elucidated, although we tentatively suggest targeting and measuring self-efficacy for exercise tasks and exercise barriers.

Abbreviations
COPD	=	chronic obstructive pulmonary disease
PR	=	pulmonary rehabilitation

Declaration of interest

The authors report no conflict of interest.