Effectiveness of a theory-based multicomponent intervention (Movement Coaching) on the promotion of total and domain-specific physical activity: a randomised controlled trial in low back pain patients

Physical activity is an independent risk factor for non-communicable diseases [1‐3]. Population-based surveys assessing physical activity behavior in Germany indicate that only around 25% to 56% of adult men and 16% to 38% of adult women meet the World Health Organization’s recommendations on physical activity [4‐6]. Data show that especially people with chronic health conditions are considered to be insufficiently physically active [7, 8]. In consequence, the promotion of physical activity is a major field in rehabilitation and health promotion.

The literature shows that comprehensive and high-quality interventions achieve the most significant long-term increases in physical activity behaviour [9]. The didactics of information delivery should be based on valid behavior change methods (e.g., social cognitive theory and the transtheoretical (or stages of behavior) model [10], setting clear and realistic goals and use simple and specific messages [11]. Though face-to-face interventions are considered the most effective approach [11], long-term interventions such as telephone-based interventions, internet-based interventions and mailed support are considered to increase sustainability [9, 11]. Current research discusses, whether individually tailored interventions are more effective than standard interventions [12, 13].

Assuming that comprehensive, theory based and tailored interventions are more effective at promoting physical activity than standard interventions [9], the aim of the present study was to compare two different interventions: A theory-based multicomponent intervention (Movement Coaching) comprising three different components (face-to-face contact, tailored telephone aftercare and internet-based aftercare) and a low-intensity and low cost intervention merely comprising two general presentations on physical activity without theoretical foundation. Owing to the high prevalence of chronic low back pain patients in inpatient orthopaedic rehabilitation in Germany this indication was chosen as a relevant sample for the study.

The research questions assessed were: (1) Is Movement Coaching more effective in promoting total physical activity than the low-intensity intervention? (2) Are there differences in leisure time, workplace and/or transportation physical activity between Movement Coaching and low-intensity intervention? (3) What are the predictors of an increase in physical activity?

Overall, 912 patients were assessed for eligibility, whereof 412 patients (44%) gave informed consent to the study participation and completed the baseline questionnaire. In the patients enrolled for eligibility, there were no differences in sex between participants and non-participants (p = 0.18) but the study participants were in average younger than the non-participants (p = 0.04). The most frequent reasons mentioned for not participating in the study were concerns in data protection (n = 99) and insufficient knowledge of the German language [27].

Overall, 35% of the Movement Coaching and 35% of the control group completed both, the six months and the twelve months follow-up questionnaire. Therefore, it was possible to analyse 144 questionnaires. Figure 2 shows the CONSORT flow diagram illustrating the progress through the phases of the present study.

Even though the results of the present study cannot contribute evidence on a superior strategy in physical activity promotion it provides novelty to the field of physical activity promotion.

The consideration of physical activity in different areas of life (workplace, leisure time, transportation) provides important implications for the future field of practice concerning physical activity promotion. In retrospect, for example, the comparatively low level of education in our sample would have needed further consideration for tailoring the intervention. As a low level of education tends to be associated with higher workplace activity [28] as well as a lower physical activity during leisure time [28‐30] the question arises, whether interventions promoting physical activity targeting persons with low level of education should rather focus on leisure time physical activity instead of total physical activity. The World Health Organization guidelines recommend at least 150 min of moderate-intensity or 75 min of vigorous-intensity aerobic physical activity throughout the week, what corresponds to 600 MET-min/week. The recommendations relate not only to sports, but also include explicitly other leisure time activities, workplace and transport activity [31]. Even though more than 50 % of the sample achieve the recommendations based on total physical activity (see Table 4), workplace physical activity contributed the largest share of physical activity performed. In contrast, the results on leisure time physical activity (see Table 4) indicated that less than 25% of the sample were engaged in leisure time physical activity during leisure time that was consistent with the World Health Organization’s recommendations to achieve health benefits. Ensuing from the highly different proportions of workplace and leisure time physical activity in total activity, a discussion about the same effects of workplace, leisure time and transportation physical activity on health should be started. Therefore, further studies on the association of physical activity in different areas of life and health status are of utmost importance.

The decline in physical activity in both groups is astonishing at first glance and seems to contradict current research showing moderate evidence for the increase of physical activity through different interventions [9, 25]. A possible explanation for the decline in physical activity might be the difficulties in measuring physical activity by a questionnaire. Literature shows a low correlation of objectively and subjectively measured physical activity [32, 33]. More particularly, a study from van Weering et al. (2011) showed that low back pain patients appear to have even more problems in estimating their physical activity levels than healthy people [34]. The assumption of an unrealistic self-assessment is supported by an additional study with the same sample comparing subjective and objective physical activity. Thereby, no significant correlations between subjective and objective results could be proved and self-reported data showed an overestimation of 46 min/day (vigorous intensity) to 78 min/day (moderate intensity) [35]. The second possible explanation could be the setting of the intervention, an inpatient rehabilitation center. As the facilitation of self-competence is an integral component of rehabilitation, it is reasonable to assume that patients might have improved their self-assessment during the intervention period. By implication, they probably overestimated their physical activity levels at the beginning of inpatient rehabilitation even more.

The results of our exploratory analyses might underline this assumption as we showed that higher baseline physical activity levels were associated with a decrease in physical activity during the twelve month period. Beyond, several studies supported the assumption that the overestimation of physical activity might be associated to a higher body mass index [36‐38] what might be supported by the association of body mass index and the increase of transport physical activity shown in our results. Hence, regarding the practical implications it needs to be considered, that our additional exploratory evaluation only gives information on factors associated with an increase in physical activity: no information on the amount of increase and its relevance on health can be given.

Overall, our study includes a number of important strengths. One such strength was the domain-specific measurement of physical activity. Besides this, choosing a low intensity intervention as a control group instead of a non-intervention control group was undoubtedly a strength of the present study. Since it was already assumed that the promotion of physical activity exerts an influence, the study went further and compared different approaches. Based on this, discussions on resource allocation in physical activity promotion might be opened. Yet, another methodological strength was the randomized controlled study design providing high internal validity of results and mostly regarded as the “golden standard” for evaluation in health care.

Nevertheless, the present study has several limitations. Certainly, the main limitation of the present study is the high drop-out. This has several consequences regarding the interpretation of the results. First, the findings should be interpreted with caution due to the fact that we did not achieve the calculated sample size and therefore the study is underpowered to detect the assumed between group differences at six and twelve month follow-up, respectively. Second, due to the associated problem of imputing missing values, especially the results regarding the decline of physical activity during the course of the study should be interpreted very cautiously. Third, a potential bias resulting from the non-reply to the postal six and twelve month follow-up questionnaires needs to be taken into account. We were not successful to increase the response rate during the conduction of the study, eeven though we integrated a repeated sending of the questionnaire if the participant did not reply within two weeks as well as incentives (all participants who sent back the postal six/twelve month follow-up questionnaires went into a draw to win a tablet computer and a voucher for a wellness-weekend). A further limitation is, although, the randomised controlled design assures a high internal validity this study design, that it is exposed to several problems in rehabilitation practice, like the patients recognising that they were receiving different types of treatment. Therefore, the authors cannot completely exclude the possibility of bias resulting from the exchange of information between the patients. To substantiate the robustness of the results missing values were imputed conservatively and an intention-to-treat analysis with last observation carried forward was calculated. This confirmed the robustness of the results in regard to the primary outcome total physical activity. However, the dropout rate in the present study is comparable to other studies evaluating multicomponent lifestyle interventions [39, 40]. A third limitation is the use of a self-reported physical activity measurement. Besides a potential bias due to social desirable answers and recall bias, validity of physical activity questionnaires needs to be questioned [33]. However, self-administered physical activity questionnaires are the most commonly used as this is the most inexpensive method to use in large-scale studies [41] and no differences in measurement or recall bias between the two groups in the present study were expected.

Our study brings up the important question if less is more in regard to physical activity promotion as our results could not statistically confirm the superiority of the comprehensive multicomponent intervention (Movement Coaching) compared to the low intensity control intervention. On the one hand, this contradicts the assumption that comprehensive and high-quality interventions achieve the most significant long-term increases in physical behaviour [9]. On the other hand, other studies comparing one intervention to another intervention instead of a non-intervention control group could not proof the superiority of a specific intervention strategy in physical activity promotion neither [12, 13, 42‐45]. As the importance of physical activity promotion in prevention and rehabilitation is beyond discussion further research not only on the effectiveness but also cost-effectiveness of different interventions promoting physical activity is needed.