Reliability of the Swedish version of the multidimensional outcome expectations for exercise scale (MOEES-SW) in a cardiac population

ABSTRACT

Purpose

Translate and adapt the Multidimensional Outcome Expectations for Exercise Scale (MOEES) into Swedish language and to explore psychometric properties, in terms of test-retest reliability, internal consistency as well as factor structure and floor and ceiling effects, of the Swedish version of MOEES in cardiac patients.

Materials and methods

This prospective psychometrical study set in a Swedish cardiac rehabilitation outpatient clinic included 74 patients; age ³18 years after acute coronary syndrome or that had undergone cardiac surgery. Translation and adaptation procedure of the MOEES followed established international guidelines. To evaluate test-retest reliability, relative (intra-class correlation coefficient ICC, 2.1)and absolute reliability (standard error of measurement (SEM) standard error of measurement % (SEM%), percentage of absolute agreement and Kappa coefficient for each item were calculated. Internal consistency was assessed with Cronbach´s alpha. The original 3-factor structure was tested with a confirmatory factor analysis. Floor and ceiling effects were calculated.

Results

In total, 60 patients, mean age 65 years, were included in the study. The test-retest showed for the three outcome expectations subscales (Physical, Self-evaluative and Social) ICC-values of 0.40 (CI 95% 0.20-0.58), 0.57 (0.39-0.71) and 0.72 (0.57–0.83), respectively. In general, the Kappa coefficients were low and varied between 0.11 and 0.44. Two questions had low loadings in the confirmatory factor analysis (<0.5) , contributing to a weak fit of the model. There was no floor effect, but the subscales physical and self-evaluative outcome expectation showed ceiling effects.

Conclusion

This is the first study to analyse test-retest reliability of the translated version of MOEES into Swedish in cardiac patients and shows need for further development of the instrument before use in clinical practice and research.

KEYWORDS:

• Multidimensional outcome expectations for exercise
• test-retest
• ischemic heart disease

INTRODUCTION

Ischemic heart disease (IHD) is a major cause of disability and death among adults in the world. IHD has large effects on the patient’s quality of life and often requires adjustments in several areas of daily life (Brink, Karlson, and Hallberg, 2006; Mayou et al., 2000). Guidelines for secondary preventive cardiac rehabilitation (CR) after cardiovascular events recommend behavioral interventions targeting: smoking cessation, regular exercise training, maintaining a healthy body weight and adherence to cardioprotective drug therapies, with the aim to decrease mortality, morbidity and improve quality of life (Heran et al., 2011; Piepoli et al., 2017).

Regular exercise training is an important part of CR in patients with IHD and post-cardiac surgery. Exercise reduces the risk of premature death and hospitalization as well as improves psychological factors, including increased health-related quality of life (HRQoL) and reduced depression and anxiety (Heran et al., 2011; Marzolini, Oh, and Brooks, 2012; Sandercock, Cardoso, Almodhy, and Pepera, 2013; Smith et al., 2004; Whalley, Thompson, and Taylor, 2014).

Several guidance documents provide recommendations for exercise-based CR (Hansen et al., 2018; Price, Gordon, Bird, and Benson 2016). Nonetheless, there is a lack of detailed guidelines on how to structure the exercise for sustainable behavior changes (Abell, Glasziou, and Hoffmann, 2017). To better understand why some cardiac patients maintain regular physical exercise after completion of hospital-based CR and some do not, it is crucial to identify determinants relevant for individual´s decisions and determine what factors are facilitators and those that are barriers for adoption and maintenance of exercise.

Social cognitive theory (SCT) is one of the most frequently applied theories used to describe health behaviors such as psychological factors influencing performance of exercise. Self-efficacy expectations and outcome expectations (OE) are described as important determinants for behavior change as well as behavior maintenance and may predict health behavior (Anderson-Bill, Winett, and Wojcik, 2011; Lee et al., 2011; Rajati et al., 2013; Taymoori, Rhodes, and Berry, 2010; Wójcicki et al., 2014).

The Multidimensional Outcome Expectations for Exercise Scale (MOEES) by Wójcicki, White, and McAuley (2009) has a theoretical base in (SCT). In MOEES, “outcome expectation” is defined as a multidimensional factor, as proposed by Bandura (1997). Hence, unlike other measures of outcome expectations which are all one-dimensional (Conn, 1997; Resnick et al., 2000; Shaughnessy, Resnick, and Macko, 2004); MOEES measures three dimensions of exercise outcome expectations: 1) physical; 2) social; and 3) self-evaluative (Hall, Wójcicki, Phillips, and McAuley, 2012). The original English version of the instrument has been tested and has demonstrated good validity in middle-aged and older people (Wójcicki, White, and McAuley, 2009) as well as in persons with multiple sclerosis (McAuley, Motl, White, and Wójcicki, 2010). However, the stability of MOEES (i.e. test–retest reliability) has not been investigated before, and overall research relating to the significance of outcome expectations on exercise in cardiac patients is scarce. Furthermore, since mainly the validity of MOEES, besides internal consistency, has been tested (Hall, Wójcicki, Phillips, and McAuley, 2012; McAuley, Motl, White, and Wójcicki, 2010; Wójcicki, White, and McAuley, 2009), it is imperative to assess the reliability of the scale. To date, there is no instrument in Swedish to investigate outcome expectations on exercise in IHD. Thus, the aim of this study was to translate and adapt the MOEES into Swedish language and to explore psychometric properties, in terms of test–retest reliability, internal consistency, and factor structure as well as floor and ceiling effects, of the Swedish version of MOEES in cardiac patients.

METHODS

This study was completed in two parts: first, we carried out a Swedish translation and adaptation of MOEES according to established guidelines (Beaton, Bombardier, Guillemin, and Ferraz, 2000; Mokkink et al., 2010); and secondly, an evaluation was conducted of the test–retest reliability, internal consistency, and factor structure as well as test of floor and ceiling effects, of the Swedish version of MOEES (MOEES-SW) in patients with IHD.

Sample

A total of 74 patients were invited to participate in the study. The patients were recruited from the outpatient clinic for CR, both patients attending the CR exercise group and patients attending individual visits at the outpatient clinic, at a university hospital in Sweden between May and October 2017. Informed written consent was obtained from all patients, and the regional ethics committee in Uppsala (Ethic code number: 2016/447), in accordance with the Helsinki Declaration, approved the study. Inclusion criteria were: age 18 years or older; cardiac disease (i.e. prior acute coronary syndrome, coronary arterial bypass grafting (CABG) or cardiac valve surgery); and being able to write and speak in the Swedish language.

Data Collection

The patients completed the MOEES-SW questionnaire and the Frändin-Grimby physical activity scale (Frändin and Grimby, 1994) as baseline data. Frändin and Grimbys physical activity scale states six different scenarios of physical activity: 1) mainly sitting in one place, reading or watching TV; 2) light PAs such as easy household tasks, as well as going for an occasional walk or doing easy gardening; 3) moderate PA for about 3 h/week, such as dusting, ordinary gardening, walking longer distances, and cycling; 4) moderate PA over 4 h/week or intense PA up to 4 h/week, such as heavy gardening, home maintenance or heavy domestic activities involving some breathlessness, and sweating; 5) active sports at least 3 h/week such as tennis, swimming, jogging, or heavy gardening or heavy leisure-time activities; and 6) competitive sports, strenuous exercise several times a week involving considerable physical exertion (e.g. swimming or jogging a longer distance).

A test leader was available to answer questions during the procedure. Background data, including type of cardiac disease, time since diagnosis of cardiac event, occupational status and if they attend 10 or more group-sessions at the hospital-based exercise CR program, were obtained from the participant’s medical records. The hospital-based exercise CR program consisted of 60-minute group sessions, twice a week, during three to four months led by a registered physiotherapist. The exercise aimed to increase physical capacity and was set up as interval training at a level of 60–85% of VO₂ max and 40–80% of 1RM. Eight different exercise stations were used including push-ups, sit-ups, and step-up exercises. The specific exercises, at each station, were repeated three times; 40–45 seconds of work followed by 15–30 seconds’ rest.

At the end of the outpatient clinic visit, the participants received a blank translated version of the MOEES-SW questionnaire. They were instructed to complete the questionnaire two weeks later at home, and return it in person or by mail in a prepaid envelope.

Measurement

The MOEES is a self-report instrument that contains 15 statements measuring three conceptually distinct constructs of OE: physical, social, and self-evaluative. Six items explore the physical dimensions (e.g. muscle and bone strength, ability to perform daily activities); five items explore the social dimensions (e.g. acceptance and social standing); and four items explore the self-evaluative dimensions (e.g. mental alertness and psychological state). A 5-point scale is used to indicate how strongly the participants agree or disagree with the 15 statements (from 1 = strongly disagree to 5 = strongly agree). Responses from items on each scale are summed to provide a total subscale score (i.e. scale range from 6 to 30 for physical outcome expectations, 4 to 20 for social outcome expectations and 5 to 25 for self-evaluative outcome expectations). Evidence of validity and good internal consistency among aging adults and persons with multiple sclerosis has been reported previously (Arifin, 2018; Beaton, Bombardier, Guillemin, and Ferraz, 2000; McAuley, Motl, White, and Wójcicki, 2010).

Translation and Adaptation of MOEES

Permission to use and translate the MOEES was obtained from the originators (Wójcicki, White, and McAuley, 2009). The process to translate and adopt the questionnaire into Swedish followed international guidelines (Beaton, Bombardier, Guillemin, and Ferraz, 2000; Sousa and Rojjanasrirat, 2011) and involved the following steps: Step 1: The MOEES was forward translated from English into Swedish by two independent bilingual professional translators, as well as by one researcher with specific knowledge in the topic. During the translation process, the translators were instructed to record, in a separate document, words that were difficult to translate or had ambiguous meanings and to note what considerations they made to lead them to their translated version; Step 2: Three bilingual Swedish professionals, with experience in the field of CR physical exercise and behavioral change, compared the three versions of the questionnaire translated from English into Swedish. They discussed ambiguities and discrepancies, phrasing, and content until reaching consensus. This process generated a single preliminary Swedish version; Step 3: Two independent bilingual professional translators, blinded to the original version of the questionnaire, back-translated the questionnaire from Swedish into English. The back-translators noted words with ambiguous meanings and considerations made during the translation process; Step 4: An expert committee involved in the translating process (the three professionals in step 2 and one translator) reviewed all the translations in order to reach conceptual, idiomatic, and semantic consensus on the pre-final Swedish version; Step 5: Pilot testing of the pre-final version. Ten healthy persons from a local activity group, over 65-years-old answered the questionnaire, along with questions about the relevance and potential ambiguities in the wording of the statements and response alternatives in the questionnaire; and Step 6: Considering the results from the pilot study, the expert committee prepared the final version for psychometrical testing of MOEES-SW.

Statistics

A total of 56 participants were required in order to estimate an ICC of 0.5 with 95% confidence interval (precision ± 0.20). Allowing for 20% attrition rate, 70 subjects were needed (Sousa and Rojjanasrirat, 2011). All statistical calculations were performed using Statistical Package for the Social Sciences (SPSS Inc, Chicago, IL, USA) 24.0 software for windows and R 3.1.1. Descriptive statistics were used for demographic data. Age is presented as mean ± SD. Physical activity level, assessed by Physical activity level (Frändin and Grimby, 1994) is presented as median (IQR). For nominal data, absolute numbers and proportions (%) are described. The results for the three different dimensions of the MOEES-SW are presented as mean ± SD.

Test–Retest Reliability

To examine the test–retest reliability, the intra-class correlation coefficient (ICC) with 95% confident interval (CI) was used as a measure of relative reliability, and calculated comparing the results of the first and second test occasion for each subscale. The ICC_2,1 calculation was based on absolute agreement two-way random model (single fixed raters), and values between 0.50 and 0.75 indicate moderate reliability, between 0.75 and 0.90 indicate good reliability and greater than 0.90 indicate excellent reliability (Koo and Li, 2016; Portney and Watkins, 2009). Absolute reliability for each subscale was calculated using the standard error of measurement (SEM) with the formula: SEM = SD x (√1-ICC) (Function SE MEAS in the Package Psychometric in R), and SEM % was calculated by dividing the SEM with the mean from test points 1 and 2 and then multiplying by 100. To test the reliability of each item, the percentage of absolute agreement was calculated. The kappa coefficient for each item was also calculated.

Internal Consistency

Internal consistency was assessed with Cronbach’s alpha, which measures the strengths of inter-item homogeneity. Data from the first test were analyzed. Items on a scale designed to measure the same construct should measure different aspects of the same trait and should therefore correlate moderately with each other (Streiner and Norman, 2015). MOEES is a questionnaire with three subscales; consequently, the items within each subscale should correlate sufficiently with each other to be considered homogeneous. Values >0.70 demonstrate a sufficient correlation between the items in each specific subscale. Values >0.95 indicate that the questionnaire contains too many items assessing the same underlying construct (Scholtes, Terwee, and Poolman, 2011).

Factor Analysis

A confirmatory factor analysis of the 3-factor structure proposed by Wójcicki, White, and McAuley (2009) was conducted to examine the strength of the 15 individual items’ loadings on the three hypothesized categories of outcome expectations, as in the original English version. Data from the first test were analyzed. Several indices of model fit were applied. The Chi-square statistics assessed the absolute fit of the model to the data (Jöreskog and Sörbom, 1996). The Root Mean Square Error of Approximation (RMSEA) value of <0.06 was regarded as a good model fit (Hu and Bentler, 1999; Schreiber et al., 2006). Finally, for the Comparative Fit Index (CFI) and the Tucker-Lewis Index (TLI), respectively, values >0.95 were considered indicative of a good model-data fit (Hu and Bentler, 1999; Schreiber et al., 2006). Floor and ceiling effects were calculated for the three subscales at test one. Over 15% of responses on the highest or lowest scores were considered to represent ceiling or floor effect (Terwee et al., 2007).

RESULTS

Part 1

After the translation, the expert committee considered the MOEES-SW to be understandable and acceptable. No further adaptations were necessary according to the pilot group. The time required to complete the questionnaire did not exceed 10 min.

Part 2

One individual, out of the 74 invited participants, declined participation. In total, 73 participants were included and completed the questionnaires at the visit. Of these, 60 participants responded to the MOEES-SW questionnaire a second time thus 55 complete in Physical dimension, 57 complete in the Self-evaluative dimension and 56 complete in the Social dimension, forming the study group for the test–retest analysis. Characteristics of the participants and scores at MOEES-SW are presented in (Tables 1 and 2). The characteristics for respondents and the non-respondents were similar, except for differences in age (65 ± 9 years vs 55 ± 12, p = .009).

Table 1. Characteristics of participants. n = 73.

Variable
Age, years (mean ± SD)	63 ± 10
Male, n (%)	57 (78)
IHD/CABG or cardiac valve surgery, n/n (%)	54/19 (74/26)
Participate in hospital-based exercise CR programme ≥ 10 times, n (%)	21 (29)
Weeks between cardiovascular event and first test, median (min-max)	9 (2–64)
Occupational status, n (%)
Retired	35 (58)
Working	22 (37)
Sick leave	2 (3)
Frandin and Grimby Physical activity level (1–6), median (IQR)	3 (3–4)

n = number, SD = standard deviation, IQR = Interquartile range; IHD = ischemic heart disease; CABG = coronary artery bypass grafting; CR-program = Cardiac rehabilitation program; Frandin and Grimby Physical activity level. Higher score = more active. 1 missing value

Table 2. Scores of MOEES-SW.

MOEES-SW	Test	Retest
Physical dimension (n = 55)	28.1 ± 2.1	26.8 ± 4.1
Self-evaluative dimension (n = 57)	21.1 ± 3.1	20.4 ± 4.0
Social dimension (n = 56)	14.8 ± 3.3	13.7 ± 3.5

Mean ± SD.

Test–Retest Reliability

The best relative test–retest reliability was demonstrated in the Social outcome expectations subscale ICC_2,1 0.72 (CI 95% 0.57–0.83) and the absolute reliability analysis of the same subscale showed SEM 1.75 and SEM % 12. The ICC_2,1 values for Physical- and Self-evaluative outcome expectation subscales were 0.40 (CI 95% 0.20–0.58) and 0.57 (CI 95% 0.39–0.71), respectively. The absolute reliability analysis of the Physical subscale showed SEM 1.63 and SEM % 6 and for the Self-evaluative subscale, SEM was 2.03 and SEM % 10 (Table 3).

Table 3. Test–retest reliability of MOEE-SW.

Variable	ICC	95% CI	SEM	SEM %
Physical n = 55	0.40	0.20–0.58	1.63	6
Self-evaluative n = 57	0.57	0.39–0.71	2.03	10
Social n = 56	0.72	0.57–0.83	1.75	12

ICC_2,1 calculation was based on absolute agreement two-way random model (single fixed raters): 0.50–0.75 = moderate reliability; 0.75–0.90 = good reliability; and greater than 0.90 = excellent reliability

In Physical and Self-evaluative outcome expectations subscales the Kappa coefficients varied between 0.11–0.28 and 0.17–0.44, respectively. Absolute agreement for Physical subscale varied between 45% and 72% and for Self-evaluative subscale the absolute agreement varied between 42% and 67%. The Kappa coefficients varied between 0.24 and 0.34 and the absolute agreement varied between 45% and 50% in Social outcome expectations subscale (Table 4–6).

Table 4. Kappa and % absolute agreement for the subscale physical.

Question in MOEES-SW	Kappa	95% CI	% absolute agreement
1 (n = 60)	0.15	(−0.05–0.34)	60%
3 (n = 60)	0.21	(0.00–0.42)	70%
5 (n = 58)	0.24	(0.06–0.43)	55%
7 (n = 60)	0.11	(−0.08–0.30)	70%
9 (n = 60)	0.12	(−0.04–0.28)	45%
12 (n = 57)	0.28	(−0.05–0.51)	72%

Table 5. Kappa and % absolute agreement for the subscale self-evaluative.

Question in MOEES-SW	Kappa	95% CI	% absolute agreement
4 (n = 60)	0.33	(0.16–0.50)	53%
6 (n = 60)	0.44	(0.27–0.61)	67%
10 (n = 60)	0.17	(0.00–0.34)	47%
13 (n = 57)	0.39	(0.20–0.57)	65%
15 (n = 57)	0.23	(0.07–0.39)	42%

Table 6. Kappa and % absolute agreement for the subscale social.

Question in MOEES-SW	Kappa	95% CI	% absolute agreement
2 (n = 60)	0.26	(0.09–0.44)	50%
8 (n = 60)	0.24	(0.06–0.42)	48%
11 (n = 57)	0.34	(0.17–0.51)	49%
14 (n = 56)	0.29	(0.12–0.45)	45%

Internal Consistency

The internal consistency of the three OE subscales of the MOEES-SW was: physical α = 0.73 (CI: 0.64–0.83), social α = 0.84 (CI: 0.78–0.90) and self-evaluative α = 0.76 (CI: 0.68–0.85).

Confirming the Factor Structure of the MOEES-SW

The confirmatory factor analysis showed a weak fit to the data examining the strength of the 15 individual items’ loadings on the three hypothesized categories of OE, as proposed in the original MOEES (Chi² = 188.8, p < .001, RMSEA [96% CI] = 0.13 [0.10–015], CFI = 0.80, TLI = 0.83). The weak fit can be explained by the low loadings of item 5 in the physical subscale and item 15 in the self-evaluative subscale (<0.5) (Figure 1).

Figure 1. Items and confirmatory factor loadings in MOEES-SW, n = 73.

Floor and Ceiling Effect

Both the physical and self-evaluative subscales showed ceiling effects as 37% and 16% had maximum points on the physical, respectively, on the self-evaluative subscales. No respondent gave the lowest score on any of the items in the three subscales. Hence, there was no floor effect.

DISCUSSION

This is the first test–retest reliability test of the MOEES. This is also the first time MOEES has been tested in cardiac patients. The results showed moderate to poor ICC and Kappa values. SEM% varied between 6% and 12%, and the internal consistency varied between 0.73 and 0.84. The internal consistency can be considered sufficient, but two questions (5 and 15) had low loadings in the factor analysis. There was no floor effect, but the subscales physical and self-evaluative outcome expectation showed ceiling effects. Furthermore, the translation and the adaptation of the MOEES were considered understandable in the Swedish language for all people involved (i.e. patients, pilot group, as well as expert committee) and took a maximum of 10 minutes to complete.

The mostly low ICC values and Kappa values can be explained by a generally poor absolute agreement but also by the ceiling effect in two of the subscales. If there is a little variation among subjects, the ICC may suggest poor reliability (Jordan, 2000). Floor and ceiling effect on MOEES have not been reported earlier but are of importance if changes in outcome expectations as a function of exercise participation are to be examined effectively. MOEES-SW can be used in different ways and may not be the proper instrument to measure changes as an effect of interventions in this group of patients with IHD and post-cardiac surgery. In this group of patients, MOEES-SW is perhaps more likely an important assessment tool for determination of what type of outcome expectations the patient is harboring. Such knowledge allows health promotion specialist and healthcare personnel to plan appropriate strategies to change or enhance outcome expectations. Outcome expectations are important in the behavioral process as they act as potential motivators for engagement in new physical activity behaviors (McAuley, Motl, White, and Wójcicki, 2010). The positive effects of regular exercise for cardiac patients are documented as multiple positive effects on traditional risk factors for IHD (Martinez et al., 2011). Thus, a sufficiently high physical activity level is recommended by international secondary prevention guidelines, and the exercise program should be individualized according to the patient´s conditions (Piepoli et al., 2017). However, the maintenance of the exercise varies between patients, and interventions targeted toward maintenance behavior should be evaluated to a greater extent. Hospital-based exercises may have to be more focused on psychological interventions (e.g. outcome expectation, self-efficacy, and social support implemented in the exercise training program to increase the chances of maintaining the behavior of regular exercise afterward). The potential impact on outcome expectations from patients´ experience of participating in the group-based exercise was not investigated specifically in the present study. The impact of group dynamics would however be of interest to investigate in future studies.

Ekblom et al. (2018) showed that patients who increase their physical activity level during the first year following myocardial infraction have a much lower mortality, over a 4.2-year follow-up period, compared with those who remained inactive. The lowest risk was seen in patients who remained physically active over the first year. These results strengthen the importance of identifying factors that are crucial in maintaining a physically active lifestyle with exercise. Outcome expectation of physical activity can be one factor (Ferrier, Blanchard, Vallis, and Giacomantonio, 2011).

In contrast to ICC, the internal consistency can be considered sufficient. Cronbach´s alpha in our study is in line with reported measures of internal consistency in older adults, both in independent- or assisted living in USA (Hall, Wójcicki, Phillips, and McAuley, 2012). The confirmatory factor analysis also provided good support for the included items (loading over 0.5) except for item number 5: “Exercise will strengthen my bones” and item 15: “Exercise will give me a sense of personal accomplishment”. In the sample of older adults with comorbidities, three other items demonstrated factor loadings <0.50; the accepted criterion established for moderate to strong fit. These three items were: “Exercise will aid in weight control”, “Exercise will improve my mood” and “Exercise will provide companionship”. These differences could probably be explained, both by the dissimilarity in the samples and cultural disparity. In the pilot testing of the MOEES-SW in a group of elderly people, before the psychometric testing, no comments were made by the participants about any of the items. However, as a result of the factor analysis, question 5 “Exercise will strengthen my bones” and question 15 “Exercise will give me a sense of personal accomplishment” do not seem to be correctly formulated; hence, they need to be reformulated for the next version of MOEES-SW.

Outcome expectations have not been reported previously in patients with IHD and after cardiac surgery. The cardiac patients in our study had very positive physical and self-evaluating beliefs about exercise and moderately positive beliefs about the social benefits of exercise and are comparable to ratings on MOEES made by symptom-free persons with multiple sclerosis (McAuley, Motl, White, and Wójcicki, 2010). However, in comparison with ratings for older adults (>85-years-old) residing in independent- or assisted living in USA who rated their outcome expectation values (mean standard deviation), on the subscales; physical 26.3 (2.9), social 12.7 (2.8), and self-evaluative as 21.3 (2.6) (Hall, Wójcicki, Phillips, and McAuley, 2012)the corresponding figures in our cardiac patients showed stronger outcome expectations (i.e. physical 28.1 (2.1) and social 14.8 (3.3)). In self-evaluative, our participants rated almost the same as the older participants (21.3 (3.1) vs. 21.3 (2.6)) in the Hall, Wójcicki, Phillips, and McAuley (2012) study. The stronger outcome expectations could be due to their relatively short-term experience of cardiac-related disability and therefore higher expectations to regain full exercise level. On the other hand, it has been reported that a number of patients with IHD were afraid of physical activity and avoided physical activity and exercise; consequently, kinesiophobia has a negative influence on attendance during CR (Bäck et al., 2016).

The high positive physical and self-evaluated outcome expectations indicate that our sample believed that regular physical activity and exercise could improve physical, self-evaluative, and social outcomes. In total, 29% of the participants in our study had participated in the hospital-based exercise group for ≥10 times, and all the participants had an individual appointment with a physiotherapist at least once.

More active individuals, as well as more efficacious individuals, report higher expectations for positive outcomes relative to exercise participation (Ferrier, Blanchard, Vallis, and Giacomantonio, 2011), which is harmonized with the social cognitive perspective (Bandura, 1997). The recruitment of participants inthe present study was a convenience sample but at the same time, it reflects well the situation we see in our clinic, i.e., approximately only one-third of the patients choose to participate in the group rehabilitation. A limitation of the study was that we did not control for changes in the level of physical activity during the two tests points.

Reliability methods, based on correlation coefficients, such as ICC, provide an indication of “relative reliability” (Lee et al., 2011). However, relative reliability measures are influenced by the range of measured values, giving no indication of actual measurement values or any systematic variability in the measures, and cannot be interpreted clinically. Reliability studies should therefore always include assessments of measurement errors and analysis of systematic bias, commonly referred to as “absolute reliability” (Stratford and Goldsmith, 1997). Absolute reliability expresses measurement errors in the same units as the original measurements. This makes it more relevant for the clinicians to interpret: smaller SEM values indicate more reliable measurements (Atkinson and Nevill, 1998).

SEM%, on the other hand, is independent of the units of measurement and could be used, for example, to compare methods (Stratford and Goldsmith, 1997). In contrast to the relative reliability in the present study, the results of the absolute reliability analysis of each subscale of MOEES-SW showed a better reliability, which could be interpreted as being of more value in the clinical context.

The stability of MOEES over time, i.e. test–retest reliability, has not been investigated before. In the present study, two weeks passed between test and retest, which was based on the recommended interval of 2–14 days (Streiner and Norman, 2015). Hence, changes could have occurred during these two weeks in outcome expectations; however, it is a relatively short time to detect any changes caused by the exercise.

CONCLUSION

To understand why some cardiac patients maintain regular physical training and some do not, following a cardiac event, it is crucial to identify factors that are relevant for individuals to change and maintain their behavior. Outcome expectations may play an important role in this process. As the first study to analyze test–retest reliability of the MOEES and the first to test the translated instrument in cardiac patients, this study shows that the instrument still needs further developments before use in clinical practice and research to measure changes as an effect of interventions.

Disclosure of Interest

The authors report no conflict of interest.

Acknowledgments

We thank the language reviewers for contributing to the translation of MOEES.