Convergent and criterion-related validity of the short form of the International Physical Activity and the Incidental and Planned Physical Activity Questionnaires in people with multiple sclerosis

1 Introduction

Interventions that target physical activity and exercise represent an important area of research in neurodegenerative diseases NDDs (Spielman, Little, & Klegeris, 2016). Studies indicate that lifestyle factors, including physical activity, may change the course of NDDs and impact the severity of motor and non-motor symptoms (White & Castellano, 2008). In people with multiple sclerosis (PwMS), regular exercise and physical activity can improve function and quality of life, as well as have other positive behavioural and health-related benefits (Latimer-Cheung, Pilutti, et al., 2013). For example, level of physical activity in PwMS correlated with the number of relapses, sleep impairment, quality of life, depression and anxiety (Latimer-Cheung, Pilutti, et al., 2013). On the other hand, many trials showed that engaging PwMS in physical activity and exercise programmes helped to decrease the affected symptoms, such as: sleep; coordination, cognitive function and depression (Aburub, Khalil, Al-Sharman, Alomari, & Khabour, 2017; Al-Sharman, Khalil, El-Salem, Aldughmi, & Aburub, 2019; Briken et al., 2014). Therefore, there has been growing attention toward evaluating and measuring physical activity as an outcome in research and practice involving PwMS.

Generally, it is of importance to assess different levels of physical activity accurately. This is because people are required to perform specific level of physical activity intensities per week. For example, the World Health Organisation recommended performing 150 minutes of moderate intensity physical activity per week or at least 75 minutes of vigorous intensity aerobic physical activity per week to improve and/or maintain general health (World Health Organization, 2010). Accordingly, people might be able to do the required amount of time spent in physical exercise, but might not be able to measure the intensity of exercise accurately. Additionally, healthcare providers and researchers will be in need to assess if people are performing exercise and physical activity that meet the recommendations. Similarly, PwMS who have mild to moderate disability are recommended to perform at least 30 minutes of moderate intensity aerobic activity two times per week and strengthening exercises for major muscle groups two times per week or to perform 30 minutes of moderate to vigorous PA per day (Latimer-Cheung, Ginis, et al., 2013). These exercises are often performed, and recommended, as part of self-management plans in PwMS (Latimer-Cheung, Ginis, et al., 2013). Being unsupervised while performing exercise might lead people to underestimate or overestimate their performance. Thus, it is crucial to check if current tools of physical activity measurements are valid and reliable in PwMS, and to choose the best outcome, where possible.

In PwMS, both subjective and objective measures of physical activity are widely used in healthcare research (Motl, McAuley, & Snook, 2005).

Overall, physical activity questionnaires as subjective measure of physical activity, remain cheaper and in hand for most healthcare institutes. The International Physical Activity Questionnaire (IPAQ) is the most common used questionnaire to assess physical activity in the general population (Lee, Macfarlane, Lam, & Stewart, 2011). However, validity of the IPAQ in people with MS was assessed only in one study by comparing it with activity monitor data (Gosney, Scott, Snook, & Motl, 2007). In this later study, Gosney et al (2007) assessed the validity of the total score of IPAQ without assessing the validity of each level of physical activity including: sedentary; light; moderate; moderate to vigorous; vigorous and very vigorous physical activities. Only one study investigated validity of time spent in sedentary physical activity in PwMS using IPAQ, but did not explore validity of time spent in light, moderate, vigorous and moderate to vigorous physical activity (Motl, Sasaki, Cederberg, & Jeng, 2019). Although IPAQ scores for the different physical activity levels was not validated in MS yet, many studies depended on IPAQ as a tool of assessing levels of PA in PwMS (B. Sandroff et al., 2012). Similarly, the Incidental and Planned Exercise Questionnaire (IPEQ) has been widely used in other populations, and found to be valid and reliable tool for assessing physical activity (Merom et al., 2014). The IPEQ was used previously in one study in PwMS (Vister, Tijsma, Hoang, & Lord, 2017). As yet, its validity in this population is not established.

With the advances in technology, objective measures of physical activity, by mean of activity monitors, became an important method in managing and/or reducing sedentary lifestyle effects in the general population (Matthews, Hagströmer, Pober, & Bowles, 2012). Activity monitors are widely used in MS studies related to physiotherapy and general health research (Matthews et al., 2012). Although, the limitation of activity monitors that some monitors might not pick all the activities, such as stationary cycling, swimming or climbing stairs, data collected from physical activity monitors do not depend on memory and recall bias that could happen while using subjective questionnaires including the IPAQ and the IPEQ (Ekelund et al., 2006). Moreover, activity monitors such as Actigraphs have been validated for specific cut-points for different levels of physical activities, specifically in PwMS (B. M. Sandroff, Riskin, Agiovlasitis, & Motl, 2014). Therefore, this study aimed to assess the convergent and criterion validity of the self-administered, short form of the IPAQ and the IPEQ for measuring different levels of physical activity in a sample of PwMS using the activity monitor (Actigraph) as the criterion instrument. For the convergent validity, it was hypothesized that PA parameters recorded by the IPEQ and IPAQ would have moderate positive correlations with the corresponding PA parameters derived from the activity monitor. Additionally, for the criterion validity, it was hypothesized that PA parameters recorded by the IPEQ and IPAQ would agree with the corresponding PA parameters derived from the activity monitor.

2 Methodology

3 Results

The demographic characteristics of the participants are presented in Table 1. Data were collected from 50 participants (39 females, 11 males) with the mean (SD) of age was 36.7±10.04 years. Table 1 demonstrates the descriptive data obtained by the activity monitor (i.e. Actigraph) and self-reported PA from the IPAQ and the IPEQ. Overall, based on IPAQ data, 58% of participants reported no participation in vigorous- intensity PA. In total, 36% of participants according to IPAQ and 70% of participants according to IPEQ did not achieve the recommended amount of PA, as they reported MVPA for less than 30 minutes a day and therefore were regarded as insufficiently active. Regarding data derived from the Actigraph, 22% did not register PA at vigorous- intensity level and additionally, 74% of participants did not accumulate at least 30 min per day at MVPA intensity and therefore were regarded as insufficiently inactive (Table 2).

Regarding convergent validity, Pearson correlation analysis showed that there is a significant correlation between sitting time as reported by IPAQ and sedentary time as recorded by Actigraph (r = 0.41, p = 0.003), VPA as reported by IPAQ and VPA as recorded by Actigraph (r = 0.46, p = 0.001), and MVPA as reported by IPAQ and MVPA as recorded by Actigraph (r = 0.36, p = 0.011) (Table 3).

Regarding the criterion validity, significant differences were noted on paired t-test between the variables derived from the Actigraph and the corresponding variables recorded on IPAQ (p < 0.0001) (Table 1), which indicates the presence of systematic difference between the objectively recorded data and the self-reported variables. This is supported by the observation that the average score of the criterion and the self-reported measure can always predict the difference between the criterion and the self-reported measure (p < 0.001), which indicates a significant proportional bias in the variables obtained from the self-reported questionnaires (i.e. IPEQ and IPAQ) (Table 5). The calculated LOA were large for all variables obtained from the IPAQ or the IPEQ (Table 5). The Bland and Altman plots indicated that the IPAQ or the IPEQ variables are consistently smaller than the corresponding Actigraphy data at lower time values and vice versa (i.e. participants reported bigger IPAQ and IPEQ values thancorresponding Actigraphy data at higher times values) (Fig. 1). This applies to all physical activity intensities including sedentary time, moderate, vigorous, and moderate to vigorous physical activity levels.

OPEN IN VIEWER

Fig. 1

Bland Altman plots of physical activity parameters measured by Actigraphy and by IPAQ and IPEQ. Solid line indicated the mean of difference, and the dotted line indicated 95% limits of agreement.

Overall, the sensitivity of IPAQ to capture insufficiently active MS participants was 92.3%, whereas the specificity of the IPAQ to capture the sufficiently active people as determined by the Actigraph was 45.94%. However, the sensitivity of IPEQ to capture insufficiently active MS participants was 30.7%, whereas the specificity of the IPEQ to capture the sufficiently active people as determined by the Actigraph was 70% (Table 2).

Although many studies have used IPAQ and tested its validity in different populations, this is the first study that investigate validity of IPAQ and IPEQ when compared to Actigraphy in PwMS. Results of this study showed that there are correlations between IPAQ sitting with Actigraphy sedentary measurement, vigorous PA in IPAQ with vigorous PA in Actigraphy, and moderate to vigorous PA in IPAQ with moderate to vigorous PA in Actigraphy,indicating that IPAQ has convergent validity for assessing PA in PwMS. These findings are similar to previous published findings that assessed validity of IPAQ in other populations. For example, IPAQ was found to be valid tool to asses PA in the general population in 12 different countries with high concurrent and construct validity in adults (Craig et al., 2003). Moreover, IPAQ was found to be valid tool to assess PA in people with neurological disorders such as cerebral palsy (Lavelle et al., 2020). IPAQ version used in this study, Arabic version of IPAQ, was also previously validated in the general population (Helou et al., 2018).

Our study has also revealed that the IPEQ has Poor convergent validity to assess PA in PwMS. Although previous studies showed that the IPEQ is a valid and reliable tool when compared with accelerometery in older adults (Delbaere et al., 2010), the current study contradict these findings. The IPEQ was used in one study in PwMS (Vister et al., 2017) without assessing its validity in this population. Thus, this study is the first study that assessed validity of IPEQ in PwMS. IPAQ but not IPEQ might be a cheap and a valid tool to assess PA where Actigraphy are not available / not feasible to be used.

However, when assessing criterion validity, proportional bias between IPAQ and Actigraphy in term of moderate, vigorous, moderate to vigorous PA and sedentary PA in PwMS. These results indicate overestimating PA levels in IPAQ when compared to the Actigraph data, particularly at larger values. In consistent with that, the IPAQ showed high sensitivity level, but very poor specificity level in this study. In other words, when classifying people into sufficiently active and insufficiently active (sufficiently active ≥30 minutes moderate to vigorous PA, insufficiently active < 30 minutes of moderate to vigorous physical activity), IPAQ was accurate in identifying people who are active (sensitive: true positive), but not accurate in term of classifying people as active while they are sedentary in Actigraphy (not specific: false negative). Similarly, IPEQ, had low sensitivity and relatively low specificity too.

Within the realm of criterion validity for IPAQ and IPEQ, a similar finding was reported in a study that assessed validity of IPAQ in people with Chronic Fatigue Syndrome when compared with daily PA diary (Meeus, Van Eupen, Willems, Kos, & Nijs, 2011). In both populations, MS and Chronic Fatigue Syndrome, the IPAQ and IPEQ might augment (ove-restimate) PA level as it depends on memory. Polls et al. (1996) advised that assessments that demands on self-report are not accurate in patients with memory difficulties. Because PwMS might have memory problems, self-reported PA questionnaires might have some recall bias (Pols, Peeters, Kemper, & Collette, 1996). Additionally, PwMS might have difficulties in estimating number of minutes spent in each activity category per day, due to the fluctuating nature of their symptoms during the week or even the day (Powell, Liossi, Schlotz, & Moss-Morris, 2017). The description of moderate- intensity and vigorous- intensity PA in both IPAQ and IPEQ and howindividuals may perceive them may add to this bias of over estimating PA when using these self-reported questionnaires. Therefore, results of the current study suggest that IPAQ and IPEQ need to be considered with caution in research that involves follow up sessions to assess PA in PwMS.

This study is not without any limitations. For example, the sample needs to be larger in size, with different disease severities included. Also, reliability of the IPAQ and IPEQ in PwMS was not addressed. Therefore, future studies are recommended to address validity and reliability of IPAQ and IPEQ in PwMS. Additionally, we do not have information about physical activity in people who did not wear the Actigraph according to the study requirements. Furthermore, the low EDSS of the participants limit the external validity of the study.

5 Conclusion

Findings of this study are important as they suggest that IPAQ in particular had convergent validity when compared to Actigraph data, but still had the problem of overestimating PA levels. Thus, researchers need to use self-reported measures such as IPAQ and IPEQ with caution if they are going to depend on it as the main outcome measure for PA.