Exploring longitudinal associations between the food and physical activity environment and weight status: natural experiments around the residential address.

Overweight and obesity are major risk factors for various chronic diseases1. Overweight and obesity are defined as having a Body Mass Index (BMI, weight in kilograms divided by square height in metres) of 25.0 or higher and 30.0 or higher, respectively. Currently, approximately half of Dutch adults are overweight or obese2. In absence of effective health policy, it is expected that 62% of Dutch adults will be overweight or obese by 20403. 
Overweight and obesity have a multifactorial and complex aetiology. Fundamentally, these are caused by an excess of energy intake relative to energy expenditure. Previously, energy intake and energy expenditure were considered as the result of individual determinants, yet individually-focused policies have been ineffective in reducing overweight and obesity. Nowadays, it has become widely acknowledged that environmental determinants also play an important role in the development of overweight and obesity4. 
Abundant access to fast-food outlets and a lack of access to physical activity facilities are considered among the most important environmental determinants of overweight and obesity. Fast-food outlets (e.g., snack bars) typically offer cheap, quickly served and high-calorie meals, and are easily accessible. Abundant access to fast-food outlets may provoke a high-calorie diet, resulting in a higher weight status. Fast-food outlets have increased in number in the Netherlands: for example, the number of fast-food outlets in the province of Groningen has increased with 33% between 2013 and 20185. Also, a lack of physical activity facilities (e.g. indoor gyms and outdoor sport fields) in the living environment may make it more difficult for individuals to be physically active, thereby affecting their weight status. 
Yet, evidence on the association between the food and physical activity environment and weight status is inconsistent and mainly comes from the United States6. There are several reasons for the inconsistency of the evidence. 
First, accurate measurement of the environmental exposure is a challenge7: it requires linkages between costly and computationally challenging geographic information systems, individual participant health data, and privacy-sensitive participant addresses. Alternatively, researchers rely on group-level exposure estimates (e.g. the number of fast-food outlets per zip-code unit7). Yet, this introduces the risk of misclassification since individuals within the same zip-code unit may differ in their food and physical activity environment. For example, physical activity facilities may be mostly present in one specific area within a zip code. Furthermore, studies typically measure food and physical activity environment either in terms of density (e.g., the amount of fast-food outlets within a certain range around the residential address) or proximity (e.g., the distance towards the nearest fast-food outlet from the residential address)8. These measures are correlated but conceptually distinct9: for example, a fast-food outlet may be nearby (high proximity) but still be the only one in the area (low density). Therefore, taking into account both measures provides a more comprehensive assessment of the food and physical activity environment10. 
Second, there is a lack of longitudinal studies on the association between the food and physical activity environment and weight status. Ideally, longitudinal designs should be enriched by natural experimental approaches, examining the effect of changes in the environment (e.g. by moving to a different address, or the openings or closings of fast-food outlets or physical activity facilities when staying at the same address) on weight status over time11. As trials with randomization of individuals to different environments are unfeasible, natural experiments with adequate confounder adjustment represent the highest form of evidence in this field. The few longitudinal studies on this topic yielded inconsistent results, and suffer from generalizability issues12, only assessed environmental exposure at baseline13–18, or used self-reported weight status as an outcome19. Furthermore, they typically did not adequately address confounding (e.g., they did not use using propensity scores20,21). 
Third, potential effect modifiers in the association between the food and physical activity environment and weight status require further investigation. For example, the food and physical activity environment may be more influential in people who are genetically susceptible of developing an elevated weight status27. Investigating the modifying role of such genetic susceptibility in the association between food and physical activity environments and BMI is difficult, as it requires a linkage between environmental exposure data, genetic data, and individual health data. Such linkages are costly, computationally challenging, and rarely available worldwide. A UK Biobank study28 found that fast-food outlet exposure was only associated with BMI among individuals genetically susceptible of an elevated weight status29. However, this study was cross-sectional in nature and contained issues regarding misclassification of fast-food outlets. So, more rigorous research is needed to investigate the potentially modifying role of genetic susceptibility in the association between the food and physical activity environment and weight status. 
Fourth, it is unclear to what extent the association between the food environment and weight status is mediated through diet, and to what extent the association between the physical activity environment and weight status is mediated through physical activity. Although dietary intake and physical activity influence weight status31, few studies have investigated the potentially mediating role of diet and physical activity in the association between the food and physical activity environment and weights status. Such investigations may inform policies designed to reduce the potential effects of unfavourable food and physical activity environments (e.g., calorie-labelling). A European-wide study found that the association between fast-food outlet exposure and BMI was mediated through fast-food consumption32, but adopted a cross-sectional design and relied on self-reported BMI. So, more research is need to investigate the potentially mediating role of diet and physical activity in the association between the food and physical activity environment and weight status. 
Therefore, we aim to examine the association between changes in the food and physical activity environment around the residential address over time and changes in BMI in adults. 
Additionally, within these associations, we aim to investigate potential effect modifiers (i.e., genetic risk) and changes in potential mediating factors (e.g., physical activity, diet).