Helping people to avoid becoming overweight (BMI ≥ 25 kg/m2) and obese (BMI ≥ 30 kg/m2) is an important public health priority, since the prevalence of excess body weight is high , and obesity is clearly associated with chronic diseases, such as cardiovascular diseases, type 2 diabetes, and certain types of cancers [2–4]. Excess weight originates from a long-term positive energy balance in which energy intake (from diet) exceeds energy expenditure (mainly physical activity). It has been estimated that adults gain on average 1.8 to 2.0 pounds per year, which is caused by an energy imbalance of only 100 kcal per day . If this energy imbalance of 100 kcal per day could be restored, weight gain could be prevented in approximately 90% of the population. The energy balance could be restored by, for instance, increasing life-style physical activity (e.g. walking) by only 15 to 20 minutes per day or approximately two hours per week .
Many people express positive intentions to increase their physical activity levels. However, many people do not act upon their positive intentions to change their lifestyles . Forming implementation intentions (II) has been proposed as a potentially effective and inexpensive intervention, particularly suited to help people to act upon their positive intentions [7, 8]. Implementation intentions are if-then plans specifying when, where and how one will act in order to achieve a goal ("If I encounter situation X, then I will perform behavior Y"; e.g. "If I arrive at work in the morning, then I will take the stairs instead of the elevator to my workplace"). By forming II individuals commit themselves to acting as soon as the specified situation is encountered. Usually II interventions consist of writing down when, where and how one will act to achieve an intended goal. Meta-analyses showed that II interventions may be a powerful tool in changing a range of health behaviors [6, 9, 10]. A large meta-analysis by Gollwitzer and Sheeran of k = 94 independent tests among a total of 8461 participants showed that II had a positive effect of medium-to-large magnitude (d = .65) on goal attainment . In the present study, we tested the impact of implementation intention interventions aimed at increasing physical activity by at least two hours per week in order to achieve weight maintenance or modest weight loss. If effective, this strategy could easily be implemented in large-scale population-based obesity-preventive interventions.
Although only a few studies examined the effect of II on physical activity, most of these studies revealed promising results of II [e.g. [11–14]]. However, several issues remain to be addressed before implementation intentions for physical activity can be applied in obesity prevention in the general public. First, the conducted trials targeted exercise or vigorous physical activity (i.e. specific categories of total physical activity) [e.g. ]. Weight maintenance, however, does not necessarily imply an increase in exercise, but can be achieved by an increase in all sorts of daily activities, including low-intensity activities [15, 16]. In fact, evidence suggests that increasing high-intensity exercise does not result in higher total energy expenditure, since such high-intensity exercise is often compensated by extra rest . Second, past studies tested II for specific and straightforward actions unlikely to result in clinically relevant changes in energy balance, such as one extra exercise session in the next week  or watching an exercise video once . Implementation intentions were thus formed for a very specific target assigned by the researchers. Total daily physical activity is a more complex category of behaviors, and includes a variety of actions (e.g. going to work by bike or foot, household work, exercise, walking, gardening). When forming II to increase total physical activity, people first have to decide what activity they chose to perform. A small meta-analysis suggested that implementation intentions were effective in pursuing self-generated goals, but that effect sizes of II were generally smaller for self-generated goals than for assigned goals . However, to date no studies directly compared the effect of II for an assigned versus a self-selected goal. We therefore tested two types of implementation intentions aimed at increasing physical activity, i.e., II for an assigned activity (walking) and II for a self-selected activity. We hypothesized that II for the assigned activity would be more effective than II for the self-selected activity. Third, people need to maintain physical activity for a longer period of time in order to prevent weight gain. Past research, however, has indicated that forming implementation intentions is more effective in studies with short-term follow-ups than for long-term goals . II are cognitive strategies that might be forgotten or become irrelevant over time . It might also be that anticipated situations and barriers may differ over time and become different from the situations and barriers anticipated in the original II. To date only effects of forming II at one point in time have been tested. It may well be that for longer-term goals, II should be formed more than once to establish clinically relevant changes in physical activity. In this study the effects of single and repeated formation of II was tested. We hypothesized that repeated formation of II would be more effective than single II formation for increasing physical activity. Fourth, previous trials were conducted in rather controlled settings (e.g. one-to-one sessions with consultants) and in selected groups (e.g. students or patients) [11, 20, 21], but not in the open, general adult population. For example, patients in rehabilitation from myocardial infarction formed implementation intentions for increasing moderate-intensity physical activity. An interviewer asked to provide more details about their plans (e.g. type of exercises, exact time, and exact circumstances) and gave supportive feedback about the plans . Although this procedure may be very useful in clinical settings and in high-risk approaches, it appears unfeasible for large-scale population-based interventions. We, therefore, examined II in a setting that could resemble that of a large-scale intervention. Finally, many past experimental studies encouraged respondents in an implementation intention condition to change behavior and to plan this change as precisely as possible, whereas respondents in the control condition only completed questionnaires without general encouragement to make behavior changes [e.g. ]. Such a design does not actually allow for assessing the unique effect of II, but assesses the effect of encouragement to change plus an II manipulation. To be able to assess the unique effect of II, in the present study also the control group was encouraged to increase their total physical activity.
In a randomized controlled trial, we tested whether forming implementation intentions to increase total physical activity by two hours per week had an impact on BMI and increased physical activity at two weeks, three and six months follow-up. Because II theory posits that forming II is a useful strategy for translating strong intentions into actual behavior change [7, 8], we hypothesized that II will be effective for individuals with strong intentions, but not for individuals with weak intentions to increase their PA by two hours per week.
More specifically, we hypothesized that:
1) Forming II leads to a greater decrease in BMI, and greater increases in physical activity (PA) than not forming II.
2) Forming II for an assigned activity (i.e. walking) is more effective than forming II for self-selected activities.
3) Repetition of forming II to increase PA is more effective than forming II once.
4) The effects of II will be moderated by the intention to increase physical activity.