This study examined the longitudinal relation between lifestyle coaching and changes in the different types of motivation in generally obese participants of a CLI. For physical activity, changes in motivational regulation were fully in line with the tenets of SDT and MI: participants showed a shift towards a more autonomous type of motivation (i.e. controlled types of motivation decreased and autonomous types increased). Moreover, if participants perceived an autonomy supportive coaching style, this was generally found to predict a larger shift in autonomous types of motivation. As regards healthy dietary behaviour, however, except for a small decrease in external motivation, no favourable changes in different types of motivation were observed. The level of perceived autonomy supportiveness of the Lifestyle Coaches appeared not to have induced any positive changes. An important predictor of favourable changes in autonomous motivation was the intensity of the BeweegKuur programmes.
Our finding of an improvement in autonomous types of motivation for physical activity confirms the findings of previous studies that investigated the relation between autonomous motivation and lifestyle changes in interventions for energy balance related behaviours ,. The findings of our study specifically demonstrate that autonomy supportive lifestyle coaching in a ‘real world’ primary care CLI contributes to a favourable shift in motivational regulation for physical activity and that, in contrast, more controlled lifestyle coaching is related to higher amotivation and to a decrease of autonomous motivation.
However, in contrast to some previous studies ,, we did not observe this favourable pattern of changes in quality of motivation for healthy dietary behaviour. It has been argued before that changing dietary behaviour may involve some physical and psychological discomfort, making it hard to be intrinsically motivated to do it . Participants in our study had a higher level of controlled types of motivation for healthy diet, compared to physical activity at baseline. Moreover, almost 75% had losing weight as their main goal compared to 34% who chose health improvement as a goal. Previous studies revealed that higher controlled regulation of eating behaviours is related to poorer body image, lower psychological well-being , to a quantity focused eating regulation  and avoidance food planning . Both strategies are negatively related to healthy eating behaviours . Moreover it has been shown that physical appearance-focused (e.g. lose weight or gain a better physical appearance) instead of health-focused weight loss goals are related to less successful eating regulation strategies . Furthermore, dietary behaviour may also include strong habits developed during childhood ,. Such findings demonstrate the complexity of eating behaviour regulation and indicate that a more autonomous motivation is indeed required for successful eating behaviour regulation. However, as we observed in our study, it also indicates that obtaining this autonomous motivation may be rather difficult.
Our finding that the influence of coaching on motivation for physical activity was virtually absent when programmes included physical activity guidance, seems in line with findings of van Hoecke et al. (2014) . Provision of a physical activity programme may facilitate the need for competence and may therefore be equally effective as need supportive coaching . Our findings in the samples that were stratified by programme type also indicate that the negative influence of perceived controlled coaching on autonomous motivation for healthy eating may be neutralised when an intervention includes a physical activity component. Previous studies have suggested a clustering of personal determinants of diet and activity  and it has been demonstrated that autonomous exercise motivation may ‘spill-over’ to facilitate improvements in eating self-regulation .
Except for the intrinsic motivation of participants in the Independent exercise programme, we found no relation between autonomy supportive lifestyle coaching and improvement of autonomous motivation for healthy diet. The absence of this relation may be caused by the LSC’s lack of knowledge about dietary behaviour and their insufficient skills to change it . Previous failed attempts by participants to change their dietary behaviour may have resulted in a struggle and frustration with this behavioural goal, indicated by a lack of change in almost all types of the motivational pattern (“I want, I need and I must”). Failed attempts can cause frustration  and feelings of lack of competence, and would thus undermine one of the basic needs for autonomous motivation , if failures have repeatedly occurred. As a consequence the LSC should pay sufficient attention to improvement of feelings of competence.
The autonomy supportive coaching style we measured may not have fully covered the true breadth of autonomy support. Autonomy supportive coaching should include support of autonomy, competence and relatedness ,. The items included in our autonomy supportive coaching style questionnaire however mainly concerned autonomy. It has been demonstrated before that primary care nurses find it difficult to apply autonomy-supportive coaching in lifestyle related behaviours – and with communication about nutritional behaviour in particular . Although participants in our study indicated that they were very satisfied with the performance of the LSC, they gave substantially lower scores for the LSC’s support to improve their dietary behaviour than the LSC’s support to improve their physical activity. Given the previously mentioned complex nature of unhealthy dietary behaviour , the LCSs must feature thorough knowledge of the problem and highly developed MI skills to favourably influence the autonomous motivation to improve this behaviour.
Although external regulation decreased on a group level, we observed a positive relation between autonomy-supportive coaching and an increase in external motivation for physical activity, which is not in accordance with SDT and MI. It has been observed before that people in treatment in general have a more external health locus of control  compared to those not in treatment. Moreover, our data revealed that participants judged the LSC to be very sympathetic and supportive. Participants with a higher external locus of control may have perceived the sympathy of the LSC as very rewarding, which may have induced an increase of their external motivation.
Strengths and limitations
Strengths of the current study include its theoretical foundation, longitudinal design, real-life intervention setting and use of validated questionnaires. The self-selected sample used in the study limits its external validity. We used a self-report questionnaire completed by participants as a proxy measure of LSC performance. This is a relatively cheap and manageable way to measure professional performance in primary care. Nevertheless, the validity of this measure would be served by direct observations, or in the ideal case by standardized patients, the gold standard in measurement of professional performance. However, both these alternatives are time-consuming and costly, especially in studies with larger numbers of participants ,.
Although this study was neither designed nor executed as an effectiveness trial, we performed an intention to treat analysis with the last observation carried forward as well as with the group mean imputation method  to replace missing data in order to address potential bias in the study results due to loss to follow up. The results of both approaches however did not change the key findings of our study regarding the predictive value of perceived coaching style on participants’ motivational regulation. However, it would require a randomized controlled trial to determine whether a change in coaching style would actually result in improved motivational regulation.