Population Health: Behavioral and Social Science Insights

Changing Population Behavior and Reducing Health Disparities

Full title
Changing Population Behavior and Reducing Health Disparities: Exploring the Potential of "Choice Architecture" Interventions

By Theresa M. Marteau, Gareth J. Hollands, and Michael P. Kelly

Abstract

Much cancer and most diabetes would be avoided if people did not smoke, moderated alcohol use, consumed less food, and became physically active. Achieving such change in these behaviors across populations would also reduce much of the gaps in health and life expectancy between the rich and the poor. In this chapter we first describe attempts to change these four aspects of behavior over the last century, focusing on the United States and the United Kingdom. We then elaborate on one particular and recent set of interventions, altering "choice architecture" or micro-environments, that can change behavior, often without conscious awareness. We argue that these show greater potential for achieving change across populations than hitherto predominant approaches that are delivered individually and/or rely on information and persuasion. We outline the conceptual and empirical research needed to estimate the contribution choice architecture interventions could make to changing behavior in populations at the scale and pattern needed to prevent noncommunicable chronic diseases in the poorest as well as the richest. In addition to this scientific challenge, we note the need to address the political challenge that stems from free market economies built on over-consumption, including what can be considered the main vectors of noncommunicable disease in the 21st Century: tobacco, alcohol, processed foods, and transport powered by fossil fuels.

Introduction: The Problem

The majority of deaths worldwide are due to four noncommunicable diseases: cancer, cardiovascular disease, type 2 diabetes, and respiratory diseases; most of these deaths are potentially preventable. Four sets of behaviors contribute to the high and growing burden of noncommunicable disease: consumption of tobacco, alcohol, and highly processed foods, as well as physical inactivity. In the United States and the United Kingdom, most of these behaviors are more common in those who are poor, thereby contributing to the large and growing gap between the rich and the poor in premature mortality and years lived without disease or disability.

We show the changes in tobacco and alcohol consumption in the United States and United Kingdom over 100+ years1,2 in Figure 1.

Figure 1. Sales of tobacco and alcohol in the United States and United Kingdom: 1900-2010

Figure comprises two charts; one depicts sales of alcohol and tobacco in the United States between 1900 and 2010, and the other shows the same information for the United Kingdom.  Both tobacco and alcohol sales have declined in the United States, while in the United Kingdom, tobacco sales have declined but alcohol sales have increased since the mid-1950s.

Figure comprises two charts; one depicts sales of alcohol and tobacco in the United States between 1900 and 2010, and the other shows the same information for the United Kingdom.  Both tobacco and alcohol sales have declined in the United States, while in the United Kingdom, tobacco sales have declined but alcohol sales have increased since the mid-1950s.

Source: Tobacco data, United States, Prof. Cristine Delnevo, Rutgers, School of Public Health (personal correspondence, 2014). Used with permission. Alcohol Data (Nephew, Williams, Hoy, et al, 2003); Tobacco Data (Forey, Hamling, Hamling, 2012); United Kingdom. Alcohol data, United Kingdom Beer and Pub Association (Personal Correspondence, 2014).

Although sales of tobacco have declined dramatically, the decline has been less for those who are more materially and socially deprived, making smoking the most important contributor to health inequalities (Figure 2). 3,4

Rates of physical activity in the United States and United Kingdom are shown in Figure 3.

Figure 2. Rates of smoking by socioeconomic status in the United States and United Kingdom

Figure comprises two charts, one for the United States and the other for the United Kingdom, that illustrate rates of smoking by occupation (United States) or employment grade (United Kingdom). The charts show that although sales of tobacco have declined dramatically, the decline has been less for those in lower paying occupations (for example, food services, retail) or employment grades (grade 1 or routine).

Figure comprises two charts, one for the United States and the other for the United Kingdom, that illustrate rates of smoking by occupation (United States) or employment grade (United Kingdom). The charts show that although sales of tobacco have declined dramatically, the decline has been less for those in lower paying occupations (for example, food services, retail) or employment grades (grade 1 or routine).

Sources: National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention; 2011. Action on Smoking and Health (ASH) factsheet; 2013.

Rapid changes in rates of obesity, as a consequence of changes in both diet and physical activity, are shown in Figure 4. This shows variation in the rates of obesity by geographical region in the United States and United Kingdom, reflecting and reinforcing disparities in health outcomes.5

 Figure 3. Trends over time in energy expenditure related to the activity domains of active leisure, occupation, travel, home, and sedentary behavior in the United States and United Kingdom

Figure comprises two graphs, one for adults in the United States and the other for adults in the United Kingdom, that show changing trends over time in hours spent per week in physical activity and sedentary behavior between 1965 and 2009 (United States) and 1961 and 2005 (United Kingdom).  Graphs show increasing periods of sedentary behavior and decreasing periods of physical activity for both the United States and the United Kingdom.

 

U.S. adults MET-hours per week of all physical activity, and hours/week of time in sedentary behavior: measured for 1965–2009, forecasted for 2010–2030.

Source: Multinational Time Use Studies (MTUS) v.5.52 (1965, 1975, 1998) v.5.8 (1985, 1992, 1995), and American Time Use Survey 2003–2009; Applying Compendium of Physical Activity MET-intensity values based on reported time spent across 41 MTUS-coded activities and by occupation. Forecasting for 2010–2030 based on 2003–2009 slopes.

Figure comprises two graphs, one for adults in the United States and the other for adults in the United Kingdom, that show changing trends over time in hours spent per week in physical activity and sedentary behavior between 1965 and 2009 (United States) and 1961 and 2005 (United Kingdom).  Graphs show increasing periods of sedentary behavior and decreasing periods of physical activity for both the United States and the United Kingdom.

U.K. adults MET-hours per week of all physical activity, and hours/week of time in sedentary behavior: measured for 1961–2005, forecasted for 2006–2030.

Source: Multinational Time Use Studies v.5.52 (1961, 1983, 1987), and v.5.8 (1974, 1995, 2000, 2005); Applying the Compendium of Physical Activity MET-intensity values based on reported time spent across 41 MTUS coded activities and by occupation. Forecasting for 2006–2030 based on 1961–2005 slopes.

Source: Ng & Popkin, 2012. Reproduced with permission from John Wiley and Sons.

Figure 4. Rates of obesity in adults in the United States and United Kingdom by region: 1990-2010

The top figure presents maps of the United States showing the prevalence of obesity by State at three time points: 1990, 2000, and 2010.

The bottom figure shows rates of obesity in regions of the United Kingdom for males and females during three time periods: 1993-1995, 1999-2001, and 2008-2010.

Source: Behavioral Risk Factor Surveillance System, CDC.

The prize for changing these health-related behaviors across populations is impressive: the majority of type 2 diabetes (75 percent) and two-fifths of cancers (40 percent) would be avoided, and the gap in life expectancy between the rich and the poor halved were people to stop smoking, consume alcohol and food in modest quantities, and become physically active.6

We start with a brief history of organized attempts to change health-related behavior in populations over the last 100 years. This provides some evidence concerning more and less effective approaches. We argue, however, that there is a relatively limited evidence base from which to meet 21st century health challenges posed by unhealthy behaviors. Altering the environments in which these behaviors occur, or "the choice architecture," seems to hold more promise than interventions based on information and persuasion. The theoretical and evidential basis for this proposition, together with future research needs, form the final portion of this chapter.

A Brief History of Organized Efforts to Change Population Behavior to Improve Health

Governments across time have had an interest in the health of their populations to ensure, among other things, a productive workforce and sufficient numbers of fit and able people to join the military. The health focus of the U.S. and U.K. governments differed in the first compared to the second half of the 20th century. This reflected changes in the pattern of diseases affecting their populations, as well as changes in employment and growing affluence. We outline the emerging policy focus through that century towards the four sets of behavior that today form the cornerstone of policies aimed at preventing chronic noncommunicable diseases (Table 1).

The Early Years: 1900–1950

In 1900, life expectancy in the United States (among the white population) was 50.8 years, and in England it was 48.2 years.2 The early 20th century was characterized by concerns about the general fitness of the population. The major threats to life, particularly for children, were those from infectious diseases, and so attempts to control their spread were the policy focus. Sanitary reforms in the 19th century to ensure clean water and safe sewage disposal had already done much to reduce infant mortality, along with the creation of public health institutions operating most often locally. These paved the way for a particular focus on the promotion of children's health (Schorb, 2013 for a review).10

More Recent Efforts: 1950-2014

By 1950, life expectancy among the white population in the United States and all populations in the United Kingdom had risen dramatically to 69.4 and 69.2 years, respectively.12 This period has been characterized as reflecting "an epidemiological transformation."8,13 The leading causes of disease and death were no longer the infectious diseases of cholera, pneumonia, influenza, tuberculosis, and diarrhea that characterized the 19th century and first part of the 20th century but instead were the noncommunicable diseases of cancer, heart disease, stroke, and diabetes.

Table 1. Threats posed by diet, alcohol, tobacco, physical activity and policy responses during the first and second halves of the 20th century

1900-1949   1950-2014
The main diet-related threats to health early in the 20th century were those of malnutrition and starvation, as well as food adulteration, exacerbated by the Great Depression in the 1930s. Policies to address these problems in the United States included the food stamp program (http://www.nga.gov/exhibitions/sarg1b.shtm) and the distribution of agricultural surplus through relief programs and school lunches.7 Britain established child welfare programs providing clean milk, vitamins, and health care.8 Diet The main threat to health during this period was from excessive consumption of food in relation to need. The 1980s onwards saw the publication of many policy documents targeting obesity largely through the provision of information about diet and physical activity.9,10 Additional interventions included nutritional labeling of food and reformulation of foods to reduce calories. While rates of obesity in the United Kingdom and United States are beginning to plateau, it is unclear which if any interventions have contributed to this change.9
Excessive consumption of alcohol was deemed a problem in the United States and United Kingdom in the first half of the 20th century. Prohibition in the United States resulted in a nationwide ban on the sale and consumption of alcohol in 1920, halving consumption but increasing organized crime, the latter contributing to the repeal of this legislation in 1933. Alcohol consumption was reduced in Britain during the First World War following legislation that increased the price of beer and reduced licensing hours.11 Alcohol Consumption of alcohol increased in the United States and United Kingdom during this period, reflecting reduced relative price as well as increased availability and marketing. Evidence accumulated to show that policies that targeted the first two of these could reduce consumption, with the body of evidence regarding the connection between marketing and consumption being small and weak.
Tobacco during this period was generally seen as harmless, with some claims even being made for its health promoting qualities as exemplified by the use of doctors advertising particular brands of cigarettes. Government interest was primarily in tobacco taxes as a source of revenue. During the two wars in this period (1914- 1918 and 1939-1945) tobacco companies vied with each other to provide troops with cigarettes as part of their rations. Tobacco In the second half of the 20th century, tobacco was recognized as a lethal product. Multiple tobacco control interventions in the United States and United Kingdom, often as part of the World Health Organization (WHO) Framework Convention for Tobacco Control, reduced rates of smoking. These included increasing price and restricting smoking in public. Nonetheless, tobacco consumption remains the leading cause of preventable death in both countries and is a key contributor to health disparities.
Many of the populations in the United States and United Kingdom in the first half of the 20th century were engaged in nonsedentary jobs, with few owning motor cars. Exercise, one form of physical activity, was largely viewed as a means of achieving and maintaining fitness for military action or sporting achievement. Physical Activity Levels of physical activity declined globally during the latter half of the 20th century including in the United States and United Kingdom, principally reflecting reductions in occupational and travel-related activity, and increases in time spent in sedentary behavior.5 Policy interventions that favor active travel, more evident in the United Kingdom than in the United States, had some impact on increasing physical activity, but the scale of progress is deemed insufficient to stem the tide of physical inactivity and its associated widespread health-harming effects.5

It was during this period that lifestyle or health-related behavior was increasingly recognized as a risk factor for disease. The Framingham study was the first to use the term "risk factor" in 1961 to describe the role of diet in heart disease.14 Smoking emerged as a key health threat in the United States and United Kingdom with the publication of landmark reports in both countries in the 1960s.15,16 The transformation of common behaviors into risks led directly to the increasing focus on health-related behaviors and efforts to change these during the latter part of the 20th century. This shift was reflected in the growing number of citations to "health behavior" in the American Journal of Public Health (Figure 5). There was also growing awareness of health disparities.17-20

Figure 5. Citations to "health behavior" in the American Journal of Public Health: 1912-2008

Chart shows the increasing number of citations to the term “health behavior” in the American Journal of Public Health from zero in 1912 to nearly 400 in 2008.

Source: Armstrong, 2009. Reproduced with permission from Sage Journals.

Intervening to Change Health Behavior to Prevent Disease

Interventions that characterized attempts to change behavior at the population level varied with the target behavior, but mainly involved one of two approaches: those that required regulation, by altering the price and availability of two of the three sets of products increasingly associated with health harms (tobacco and alcohol), and educational or information-based approaches in which information was provided to persuade individuals to change their behavior. A wide variety of materials have been used to convey information including messages on television, leaflets, billboards, and direct communication in health education or counseling contexts being aimed at a range of audiences. The target for these interventions was usually particular behaviors, such as smoking, but with a gradual shift towards interventions that targeted multiple behaviors using education, counseling, or both.

There is a growing recognition of the limited effectiveness of such information-based interventions, with a recent systematic review of the evidence concluding that one-to-one, family oriented or worksite-based interventions are ineffective in general populations in reducing cardiovascular mortality or clinical events.21 The authors stated: "Health protection through national fiscal and legislative changes that aim to reduce smoking, dietary consumption of fats, 'hidden' salt and calories, and increase facilities and opportunities for exercise, should have a higher priority than health promotion interventions applied to general and workforce populations."21 These findings need to be read against a divide which had by then opened up in the public health community between those who favored an approach based on individual behavior change and those who advocated changing the legal, fiscal, or regulatory climate.22 Nonetheless, they capture the broad trend away from educational and information-based approaches as effective means to change behavior in populations at the scale and with the reach needed. The reason for this is evident in the more recent scientific literature on behavior and behavior change. This highlights the potentially greater effectiveness for changing behavior across populations of interventions that involve changing environments (with the potential to influence many people) and that focus on nonconscious processes (which appear more potent in eliciting many health-related behaviors).

The theoretical perspectives that guided the development of interventions began to shift over the latter part of the 20th century. For much of this period, the dominant models of behavior and behavior change had their roots in subjective expected utility (SEU) models.2,3 In brief, these models predicted a choice based on the option with the highest expected utility, this being a function of the personal utility of an outcome and the likelihood of its occurrence. In keeping with this, communications about risks to health commonly included a message about the benefits of a targeted behavior (e.g. fruit and vegetable consumption to improve general health, or stopping smoking to reduce risk of heart disease) and the perceived probability of its occurrence. More elaborate models incorporated self-regulatory concepts of self-control and acknowledged that while these models may predict intentions to change behavior, their ability to predict actual behavior was less accurate. The dominant models included the Theory of Planned Behavior,24 Protection Motivation Theory,25 the Health Action Process Approach,26 and Stages of Change.27 Aspects of these dominant models were implicit—and occasionally explicit—in policy documents and academic writing that focused attention on behavior as a reflective, conscious choice to the exclusion of nonconscious behaviors cued by environments.10

By the end of the 20th century, it was apparent that there had been some successes in achieving healthier behavior in populations, most notably in relation to tobacco, likely a result of multiple interventions that included information as well as fiscal regulation and controls on advertising. But it was also clear that far more was needed to reverse the high and rising burden of noncommunicable diseases and health disparities.

Two distinct responses were discernible among behavioral researchers. The first comprised concerted efforts to continue developing interventions using existing models of behavior and behavior change but to apply these with more theoretical integrity and methodological rigor. The Behavior Change Consortium, set up in 1999 as a collective across 15 of the National Institutes of Health (NIH) institutions, encouraged evaluations that tested more than one model or involved attempts to change multiple behaviors.28 While acknowledging the importance of combining individual and environmental approaches, particularly for increasing physical activity, the dominant models driving interventions remained those that focused on goal-directed, self-regulatory processes central to the conscious reflective systems guiding behavior. Working across these models, a group of largely U.K.-based psychologists developed a Behavior Change Technique typology as a basis for more systematic, theoretically-coherent conduct and reporting of studies.29

These and other initiatives will doubtless provide much needed improvement to the evidence base, particularly based more on individuals and directed at conscious processes.

The second distinct response was a recognition that the dominant paradigm, with its reliance on models that emphasized the conscious, intentional control of behavior, produced results with, at best, small effect sizes and high degrees of variance. The failure was pinned on the absence of nonconscious routes to behavior within the models that dominated the behavior change literature.30,31 Nonconscious processes had played a central role in explanations of human behavior at the end of the 19th century and the first part of the 20th century as reflected in the writings of James, Wundt, and Freud and a little later in the work of animal learning theorists.32,33 The cognitive revolution of the 1960s shifted the focus towards more conscious processes as evident in the models that came to dominate health behavior research.34 Some of the most compelling evidence drawing attention to the need to embrace nonconscious processes was provided in a meta-analysis of the experimental evidence concerning the causal link between intentions and behavior.31 The authors concluded that "...intentional control of behavior is a great deal more limited than previous meta- analyses of correlational studies have indicated." They went on to argue that future attempts to change behavior should pay greater attention to automatic, nonconscious routes to action.

Further concerns about interventions targeting individuals and those targeting conscious processes stem from their potential to increase health disparities either through accessibility or differential effectiveness. In a systematic review assessing these effects, evidence was lacking on the potential for educational interventions to affect disparities, but mass media campaigns were identified as one type of intervention associated with increased disparities and fiscal interventions associated with a decrease.35 There also has been considerable unease about the values implicit in approaches directed to individuals, notably that a focus on individuals too readily elides their responsibility for change with blame for not changing.36

We focus here on the potential of one particular set of interventions that involve altering environments to cue healthier behavior, also known as "Choice Architecture." In addition to capitalizing on contemporary models of behavior that emphasize the nonconscious shaping of behavior by cues in the environment, this approach complements those based on sociologically informed understanding of health disparities, which highlights the importance of the social and material conditions that constrain the way people live.

Potential of Choice Architecture

Before we consider the potential of choice architecture to change population behavior, we first define the term and present a provisional typology and map of the evidence, as a means to identify the conceptual space inhabited by this set of interventions. "Choice architecture" is a term used to denote an approach designed to change behavior unobtrusively in predictable ways. It draws on the psychological and sociological understanding of the relationship between people and the environments they inhabit. Choice architecture interventions work by altering features of these environments to change the way in which the choices people could make are presented, thereby altering the likelihood of different behavioral responses.

Although only recently brought to the wider attention of the public and policymakers by the book Nudge,37 the approach is not new, reflecting historical themes across psychology and neuroscience, economics, sociology, and the population health sciences. Links can be drawn, for example, to psychological perspectives on situationism—a position that regards behavior as principally determined by external, environmental factors rather than internal characteristics or motivations—and recognition of the automatic or nonconscious bases of much of our behavior.30 It is also informed by behavioral economics with its focus on explaining why people's behavior deviates from classical economic theory.32 From a sociological perspective, choice architecture can be conceptualized as an interaction between individual human agency and both the immediate and broader environment that make up the social structure, following a long line of sociological thought going back to Durkheim and Marx.38 It is consistent with a population health sciences perspective because it is a population-level approach, given features of the environment that have the potential to exert influence on everyone who comes into contact with them.

In relation to changing health behavior, this approach has typically been applied to the physical aspects of small-scale or micro-environments—principally those within buildings, such as workplaces and shops—rather than macro-environments, such as city design or transport infrastructure. Given this context, we have defined choice architecture interventions in the following way:39

  • Choice architecture interventions are those that involve altering the properties or placement of objects or stimuli within micro-environments with the intention of changing health-related behavior. Such interventions:
    • Are implemented within the same micro-environment as that in which the target behavior is performed.
    • Typically require minimal conscious engagement.
    • Can in principle influence the behavior of many people simultaneously.
    • Are not targeted or tailored to specific individuals.

A Typology of Choice Architecture Interventions

Informed by this definition, we conducted a large-scale systematic scoping review of evidence of the effects of choice architecture interventions on diet, physical activity, and alcohol and tobacco use.39,40

From this we developed a typology of choice architecture interventions in micro-environments, displayed on the left side of Figure 6.41 This groups the available evidence into nine types of intervention that share common characteristics. These can be placed within two higher level classes of intervention that involve altering: (1) the properties of objects or stimuli and/or (2) the placement of objects or stimuli. As can be seen in Figure 6, some interventions involve both.

Figure 6. Provisional typology of choice architecture interventions in micro-environments (left side) and mapping of available evidence (right side)

Figure comprises two side-by-side charts that depict a provisional typology of choice architecture interventions in micro-environments (left side) and the mapping of available evidence (right side).  The typology chart groups the available evidence into nine types of intervention that share common characteristics. The evidence chart shows the number of study reports in four areas: diet, physical activity, alcohol use, and tobacco use.

Source: Hollands et al, 2013. Originally published by BioMed Central; Used with permission.

Mapping the Existing Evidence Base

We mapped the existing evidence base corresponding to these intervention types and the health behaviors to which they have been applied, represented on the right side of Figure 6. Although we did not exhaustively identify all of the available research, the systematic approach taken in generating the figure means that it is likely to broadly reflect the overall distribution of evidence.

Most of the evidence base comprises studies of interventions to change purchasing and consumption of food, with far fewer studies having focused on interventions to change physical activity or alcohol or tobacco consumption. Reflecting our earlier observation that persuasion through providing information has historically been the principal approach to behavior change, the most common types of choice architecture intervention are those that involve giving verbal and numeric information to be read, thus being farthest from the spirit of our definition of choice architecture. These include point-of-choice labeling, such as nutritional labeling on foods, and the use of prompts such as posters to encourage behaviors such as stair-climbing. In the Box below we provide a short summary of where the evidence is accumulating and some examples of the types of interventions that have been used to change each of four sets of behavior.

Summary of accumulating evidence for choice architecture interventions by behavior

Importantly this summary describes where the evidence is accumulating, as indicated by a scoping review,39 but it does not provide a synthesis of effects (for which in-depth systematic reviews are needed).

Diet

Interventions to change diet-related behaviors, including the purchasing and consumption of food, make up the great majority of the choice architecture evidence- base with multiple examples for every intervention type, including:

  • Proximity: altering layouts to increase or decrease the distance of products from routes of passage, such as choosing the area where unhealthy foods are dispensed or individual products within a salad bar are placed, nearer or farther away from customers.42
  • Sizing: increasing or decreasing the portion sizes of foods served.43
  • Priming: using conspicuously placed recipe posters to prime healthier eating behaviors.44 Labeling: using nutritional labeling on the packaging of food products to indicate nutrient composition.45

Physical Activity

Interventions to increase physical activity were mainly those using prompting, although some studies using other interventions were found:

  • Prompting: the use of signs or posters to encourage stair vs. lift or escalator use posted on surfaces proximal to stairwells and on stair risers.46,47
  • Availability: reducing the availability of lift or escalator options, by increasing the time taken for elevator doors to close48 or reducing the number of working escalators.49

Alcohol Use

There is relatively little evidence. Some examples include:

  • Ambience: altering the type or volume of music played in shop or bar environments.50
  • Functional design: changing the shape of drinking vessels.51,52

Tobacco Use

There is relatively little evidence. Some examples include:

  • Sizing: altering the size of cigarettes.53
  • Presentation: changing the design of cigarette packaging.54

 

Estimating the Potential of Choice Architecture Interventions

Both primary and secondary research are needed to assess the potential of choice architecture interventions to change behavior across populations. Although the typology we present in Figure 6 is provisional, primarily descriptive, and may benefit from additional analysis and development, it can, nonetheless, function to frame and inform research to populate this space with estimates of the effect sizes of choice architecture interventions. By providing a structure for knowledge of the nature and parameters of the evidence base, it can inform both primary research and the design of reviews aiming to identify and interpret the available evidence.

Figure 6 also highlights that the evidence relating to some interventions and behaviors is limited in quantity (in particular, the paucity of studies focused on physical activity and alcohol and tobacco use), indicating significant potential for primary research. In some areas, there appears to be sufficient evidence for systematic reviews to generate reliable estimates of the effects of a given intervention.

Developing Systematic Reviews of the Effects of Choice Architecture Interventions

Systematic reviews seek to collate all evidence that fits prespecified eligibility criteria in order to address a specific research question, aiming to minimize bias in estimating effects by using explicit, systematic methods.55 A review is part of an iterative cycle of evidence generation governed by the research question, in which findings have the potential to inform policy and practice decisions, as well as decisions about the primary research that could contribute most to addressing the given research question. Further primary research may then feed back into an updated systematic review, and the cycle continues.56

In the case of choice architecture interventions, the approach has gained traction in policy circles and been devalued in others, but its effectiveness remains largely unknown.57 It is therefore premature to make judgments on likely effectiveness, since systematic reviews featuring formal critical appraisal and synthesis would be necessary to produce reliable assessments of the likely direction and magnitude of intervention effects, as well as the factors that may moderate those effects.

We are, for example, conducting a systematic review of one choice architecture intervention identified within the typology, namely "sizing," for which we found substantial evidence that has yet to be rigorously synthesized.58 This review aims to estimate the effects of manipulating portion, package, or tableware sizes on the selection or consumption of food, alcohol, or tobacco products. In addition, the review aims to estimate the extent to which these effects may be modified by both the characteristics of the intervention—important when such interventions are likely to display significant heterogeneity in their characteristics—and of the participants. The latter is central to understanding the potential of interventions to reduce or increase existing health disparities and to determine whether there is evidence of differential effects across populations.

Choice Architecture and Health Disparities

Reducing health disparities is an important goal of health policies in the United States and United Kingdom. The social patterning of health- related behavior has been estimated to explain between 20 percent and 70 percent of the variation in life expectancy between the least and most materially and socially deprived in high-income countries.59 Choice architecture interventions have the potential to reduce health disparities by focusing on two sets of interacting influences on these disparities: (1) those concerning the environments in which people live (physical, social, and economic), and (2) those concerning the cognitive resources of individuals to resist these environments.

Environments that house those who are more deprived contain more cues to unhealthier behaviors and fewer cues to healthier behaviors than environments that contain more advantaged residents. For example, the density of fast food outlets is greater in deprived neighborhoods,60-61 as is the density of alcohol and tobacco retail outlets.62,63 In addition, walkability is a feature not usually found in more deprived neighborhoods.64

The ability to resist tempting environmental cues, often without awareness, depends in part on two sets of cognitive resources: the individual's executive functioning, and his or her immediate cognitive capacity. There is now growing evidence to suggest that both these resources may be reduced in those who experience greater degrees of deprivation. Executive functioning is a theorized behavioral control network linked to the prefrontal cortex. One of its core functions is the inhibition of impulsive responses. The strength of executive functioning predicts obesity and alcohol and tobacco consumption, as well as physical inactivity, and there is increasing evidence of its links to environment factors experienced in one's early years, particularly poverty.65-67 We can think of executive function as being a relatively fixed trait, albeit with some plasticity. Immediate cognitive capacity is also influenced by poverty as evident in decisionmaking related to economic and non-economic problems, hypothesized to arise from the cognitive demands associated with poverty.68-71

In sum, those who are most deprived face a double hit: living in environments that provide stronger cues to unhealthy behaviors, coupled with a reduced capacity to resist impulses to these cues. These two sets of contributors to a higher likelihood of unhealthy behaviors in those who are most materially and socially deprived provide several points for intervention. These include interventions designed to protect children from environments that undermine the development of executive functioning, such as early years parenting programs72 and income supplementation of the poorest families.62 The second point for intervention is the environment. As highlighted in our definition of choice architecture, these interventions typically require little or no conscious engagement, and so they do not draw upon executive functioning.30 An approach that involves the removal of cues for unhealthier behaviors and the addition of ones for healthier behaviors has the potential to shape the behavior of all those exposed, regardless of their executive functioning. Although plausible, this potential remains an untested hypothesis.

Implications for Practice

The evidence base for choice architecture interventions is at an early stage, so specific recommendations for practice are unreliable at this time. We would, however, encourage practitioners to continue to follow developments in what is a rapidly growing body of evidence. Where interventions are implemented, this should be within a rigorous evaluative framework so that it can enhance the existing evidence base.

Implications for Research

We offer the following priorities for research to build the evidence base on choice architecture interventions.

Primary research:
  • Estimating the effect sizes of choice architecture interventions to change each of the four sets of behavior across populations and in those who are most deprived. In particular, there is a general paucity of studies of the effects of interventions to change levels of physical activity and alcohol and tobacco use.
  • Testing the behavioral and neuroscience bases of choice architecture interventions by, for example, (1) assessing the extent to which their impact requires minimal conscious engagement with the intervention and (2) assessing the extent to which their impact is not moderated by strength of cognitive resources, including executive functioning.
  • Testing components of the micro-environments identified here and exploring their links to broader social structures.
Evidence synthesis:
  • Conducting systematic reviews of the effects of choice architecture interventions, where it is determined there is sufficient evidence.

Concluding Comment

The changing patterns of health- related behaviors over the last century provide examples of unhealthy behavior that has declined, albeit more slowly in those who are most deprived (smoking), unhealthy behavior that has increased and fallen in response to some effective but unpopular interventions (excessive alcohol consumption), unhealthy behavior that has increased enormously but is perhaps now plateauing (excessive consumption of food), and unhealthy behavior that is rising with few signs of abating (physical inactivity). While a wide range of interventions have contributed to the success of tobacco control, the narrower range and scale of interventions targeted at the latter two types of behavior (alcohol use and inactivity) have been inadequate due to the nature and scale of the problem. It is increasingly acknowledged that the physical, material, and social environments in which we live cue these behaviors, often without awareness; this acknowledgment is shifting focus away from largely ineffective interventions that focus on encouraging individuals to resist these environments towards interventions that focus on changing environments to cue healthier behavior. As part of this, choice architecture interventions are a promising addition to the range of interventions needed to change behavior at the scale needed, including those involving altering physical macro-environments, social environments, and economic environments.

We now face two key challenges, one scientific and one political. The scientific challenge is to generate high quality evidence to estimate the magnitude of effect of choice architecture interventions, singly and in combination, for sustained change in the four sets of behavior that contribute most to premature deaths, disability, and disparities arising from chronic, noncommunicable diseases. The political challenge stems from the tension between health and wealth creation, made more acute by the recent economic recession. The United States and United Kingdom have economies that are built on excessive consumption—including consumption of tobacco, alcohol, and food, and indeed, the use of transport powered by fossil fuels. Attempts to reduce consumption, particularly of the first three products, in the form of evidence-based policies that might include restricting their availability, reducing their appeal through changing their presentation, or increasing their prices are met with intense lobbying of governments and the public, and in some cases, lead to litigation. Meeting these challenges will require vision and commitment from research funders alongside an alignment of public and political wills to implement the growing evidence base on effective interventions that could start to turn the tide on the huge and growing global burden of potentially avoidable noncommunicable diseases.

Acknowledgments

The Behaviour and Health Research Unit, University of Cambridge, is funded as part of the Department of Health Policy Research Program (Policy Research Unit in Behaviour and Health [PR-UN-0409-10109]). The Department of Health had no role in the writing of this paper, the decision to publish, or preparation of the manuscript. Theresa Marteau was funded by the National Institutes of Health (NIH) to attend the NIH's Office of Behavioral and Social Sciences Research meeting (March 26-27, 2014), at which an earlier draft of this paper was presented. We thank Virginia Berridge for acquainting us with the historical context of current policy foci on health- related behaviors. We also thank Andrew Hoy who compiled the data on which Figures 1 and 2 are based.

Authors' Affilliations

Theresa M. Marteau, Gareth J. Hollands, and Michael P. Kelly, Behaviour and Health Research Unit, University of Cambridge, U.K.

Address correspondence to: Theresa M. Marteau, Behaviour and Health Research Unit, University of Cambridge School of Clinical Medicine, Box 113 Cambridge Biomedical Campus, Cambridge, CB2 OSR; email tm388@cam.ac.uk.

References

  1. Nephew TM, Williams GD, Yi H, et al. Surveillance report #59: Apparent per capita alcohol consumption: national, state, and regional trends, 1977–2000 (Vol. Alcohol Epidemiologic Data System). Rockville, MD: NIAAA Division of Biometry and Epidemiology; 2003.
  2. Forey B, Hamling J, Hamling J, et al. International smoking statistics (Web edition). A collection of worldwide historical data. Sutton, UK: P.N. Lee Statistics & Computing Ltd 2012.
  3. Syamlal G, Mazurek JM. Current cigarette smoking prevalence among working adults—United States, 2004-2010. National Center for Chronic Disease Prevention and Health Promotion, CDC. MMWR 2011; 60(38):1305-9. Available at http://www.cdc.gov/mmwr/preview/mmwrhtml/mm6038a2.htm#tab2. Accessed October 2, 2014.
  4. Action on Smoking and Health (ASH). ASH factsheet. Smoking statistics who smokes and how much. Action on Smoking and Health; 2013. Available at http://ash.org.uk/files/documents/ASH_106.pdf  [online]. Accessed October 9, 2014.
  5. Ng SW, Popkin BM. Time use and physical activity: a shift away from movement across the globe. Obes Rev 2012; 13(8), 659-80.
  6. Report by the Secretariat. Prevention and control of noncommunicable diseases. Outcomes of the high-level meeting of the General Assembly on the prevention and control of non-communicable diseases and the first global ministerial conference on healthy lifestyles and noncommunicable disease control. Sixty-Fifth World Health Assembly, 26th April 2012, Geneva.
  7. Rosen G. A history of public health. New York: MD Publications, Inc; 1958.
  8. Gorsky M. Public health in the West since 1800: the responses. In Berridge V, Gorsky M, Mold A (Eds.), Public health in history. New York: McGraw-Hill Education; 2011.
  9. Jebb SA, Aveyard PN, Hawkes C. The evolution of policy and actions to tackle obesity in England. Obes Rev 2013; 14(Suppl 2):42- 59.195
  10. Schorb F. Fat politics in Europe: theorizing on the premises and outcomes of European anti-"obesity-epidemic" policies. Fat Studies 2013; 2(1):3-16.
  11. House of Commons, Health Committee. Report on Alcohol. London: House of Commons; 2010. Available at http://www.publications.parliament.uk/pa/cm200910/cmselect/cmhealth/151/15102.htm  Accessed August 25, 2014.
  12. Livi-Bacci M. A concise history of world population (3rd Ed). Oxford: Blackwell; 2001.
  13. McLeroy KR, Crump CE. Health promotion and disease prevention: a historical perspective. Generations 1994; 18(1):9-17.
  14. Kannel WB, Dawber TR, Kagan A, et al. Factors of risk in the development of coronary heart disease—six year follow-up experience. The Framingham Study. Ann Intern Med 1961;55:33-50. PMID: 13751193.
  15. A report of the Royal College of Physicians on smoking in relation to cancer of the lung and other diseases. London: Royal College of Physicians; 1962.
  16. Smoking and health: Report of the Advisory Committee of the Surgeon General of the Public Health Service. Rockville, MD: U.S. Public Health Service, Office of the Surgeon General; 1964.
  17. Antonovsky A. Social class, life expectancy and overall mortality. Milbank Q 1967; 45(2):31- 73.
  18. Marmot M. Fair society, healthy lives: a strategic review of health inequalities in England post-2010. London: UCL Institute of Health Equity; 2010.
  19. Townsend P, Davidson N, Whitehead M. Into the Divide. Inequalities in health: the Black Report and the health divide. London: Penguin; 1992 (revised).
  20. Armstrong D. Origins of the problem of health-related behaviours: a genealogical study. Soc Stud Sci 2009; 39(6):909-26.
  21. Ebrahim S, Taylor F, Ward K, et al. Multiple risk factor interventions for primary prevention of coronary heart disease. Cochrane Database Syst Rev 2011; CD001561.
  22. Capewell S, Graham H. Will cardiovascular disease prevention widen health inequalities? PLoS Med 2010;7(8):e1000320.
  23. Savage L. The foundations of statistics. New York: Wiley; 1954.
  24. Fishbein M, Ajzen I. Belief, attitude, intention and behavior: an introduction to theory and research. Reading, MA: Addison Wesley; 1975.208
  25. Rogers RW, Prentice-Dunn S. Protection motivation theory. In Gochman DSE (Ed.), Handbook of health behavior research 1: personal and social determinants (pp. 113-32). New York: Plenum Press; 1997.
  26. Schwarzer R. Self-efficacy in the adoption and maintenance of health behaviors: theoretical approaches and a new model. In Schwarzer R (Ed.) Self-efficacy: thought control of action (pp. pp. 217- 43.). Washington, DC: Hemisphere; 1992.
  27. Prochaska JO, Velicer WF, Rossi JS, et al. Stages of change and decisional balance for 12 problem behaviors. Health Psychol 1994;13(1):39-46.
  28. Ory MG, Jordan PJ, Bazzarre T. The behavior change consortium: setting the stage for a new century of health behavior-change research. Health Educ Res 2002;17(5):500-11.
  29. Michie S, Richardson M, Johnston M, et al. The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: building an international consensus for the reporting of behavior change interventions. Ann Behav Med 2013;46(1):81-95.
  30. Marteau TM, Hollands GJ, Fletcher PC. Changing human behavior to prevent disease: the importance of targeting automatic processes. Science 2012;337 (6101):1492-5.
  31. Webb TL, Sheeran P. Does changing behavioral intentions engender behavior change? A meta- analysis of the experimental evidence. Psychol Bull 2006;132(2):249-68.
  32. Hull CL. Principles of behavior: an introduction to behavior theory. New York: Appleton-Century Company; 1943.
  33. Skinner BF. The behavior of organisms; an experimental analysis. New York: Appleton-Century Company; 1938.
  34. Conner M, Norman PE. Predicting health behaviour. Buckingham, UK: Open University Press; 1996.
  35. Lorenc T, Petticrew M, Welch V, et al. What types of interventions generate inequalities? Evidence from systematic reviews. J Epidemiol Community Health 2013;67(2):190-3.
  36. Crawford R. Healthism and the medicalization of everyday life. Int J Health Serv 1980;10(3), 365-88.
  37. Thaler RH, Sunstein C. Nudge: improving decisions about health, wealth, and happiness. New Haven: Yale University Press; 2008.
  38. Kelly MP. The axes of social differentiation and the evidence base on health equity. J Royal Soc Med 2010;103(7):266-72.
  39. Hollands G, Shemilt I, Marteau T, et al. Altering choice architecture to change population health behaviour: a large-scale conceptual and empirical scoping review of interventions within micro- environments Cambridge: University of Cambridge; 2013.
  40. Shemilt I, Simon A, Hollands G, et al. Pinpointing needles in giant haystacks: use of text mining to reduce impractical screening workload in extremely large scoping reviews. Res Synth Methods 2014;5(1):31-49.
  41. Hollands GJ, Shemilt I, Marteau TM, et al. Altering micro-environments to change population health behaviour: towards an evidence base for choice architecture interventions. BMC Public Health 2013;13(1):1218.
  42. Rozin P, Scott S, Dingley M, et al. Nudge to nobesity I: Minor changes in accessibility decrease food intake Judgm Decis Mak 2011;6 (4):323-32.
  43. Rolls BJ, Roe LS, Meengs JS. The effect of large portion sizes on energy intake is sustained for 11 days. Obesity 2007;15(6):1535-43.
  44. Papies EK, Hamstra P. Goal priming and eating behavior: enhancing self-regulation by environmental cues. Health Psychol 2010;29(4):384-8.
  45. Dumanovsky T, Huang CY, Nonas CA, et al. Changes in energy content of lunchtime purchases from fast food restaurants after introduction of calorie labelling: cross sectional customer surveys. BMJ 2011;343.
  46. Soler RE, Leeks KD, Buchanan LR, et al. Point-of-decision prompts to increase stair use. A systematic review update. Am J Prev Med 2010;38(2 Suppl):S292-300.
  47. Webb OJ, Eves FF, Kerr J. A statistical summary of mall-based stair-climbing interventions. J Phys Act Health 2011;8(4):558-65.
  48. van Houten R, Nau P, Merrigan M. Reducing elevator energy use: a comparison of posted feedback and reduced elevator convenience. J Appl Behav Anal 1981;14(4):377-87.
  49. Faskunger J, Poortvliet E, Nylund K, et al. Effect of an environmental barrier to physical activity on commuter stair use. Scand J Food Nutr 2003;47(1):26-8.
  50. Guéguen N, Jacob C, Le Guellec H, et al. Sound level of environmental music and drinking behavior: a field experiment with beer drinkers. Alcohol Clin Exp Res 2008;32(10):1795-8.
  51. Attwood AS, Scott-Samuel NE, Stothart G, et al. Glass shape influences consumption rate for alcoholic beverages. PLoS ONE 2012;7(8):e43007.
  52. Wansink B, van Ittersum K. (2003). Bottoms up! The influence of elongation on pouring and consumption volume. J Consum Res 2003;30(3):455-63.
  53. Russell MA, Sutton SR, Feyerabend C, et al. Smokers' response to shortened cigarettes: dose reduction without dilution of tobacco smoke. Clin Pharmacol Ther 1080;27(2):210-8.
  54. Stead M, Moodie C, Angus K, et al. Is consumer response to plain/standardised tobacco packaging consistent with Framework Convention on Tobacco Control Guidelines? A systematic review of quantitative studies. PLoS ONE 2013;8(10):e75919.
  55. Higgins JPT, Green SE. Cochrane handbook for systematic reviews of interventions, version 5.0.2 [updated September 2009]. The Cochrane Collaboration; 2009. Available from http://handbook.cochrane.org/.
  56. Wilson E, Abrams K. From evidence-based economics to economics-based evidence: using systematic review to inform the design of future research. In: Evidence-based decisions and economics: health care, social welfare, education and criminal justice. Oxford: Wiley-Blackwell; 2010.
  57. Marteau TM, Ogilvie D, Roland M, et al. Judging nudging: can nudging improve population health? BMJ 2011;342:228.
  58. Hollands GJ, Shemilt I, Marteau TM, et al. Portion, package or tableware size for changing selection and consumption of food, alcohol and tobacco. Cochrane Database Syst Rev 2014;(4):Art. No. CD011045.
  59. Stringhini S, Dugravot A, Shipley M, et al. Health behaviours, Socioeconomic status, and mortality: further analyses of the British Whitehall II and the French GAZEL prospective cohorts. PLoS Med 2011;8(2):e1000419.
  60. de Vogli R, Kouvonen A, Gimeno D. 'Globesization': ecological evidence on the relationship between fast food outlets and obesity among 26 advanced economies. Crit Public Health 2011;21(4):395-402.
  61. Obesity and the environment: fast food outlets. Public Health Observatories in England: National Obesity Observatory; 2013. Available at http://www.noo.org.uk/uploads/doc/vid_15683_FastFoodOutletMap2.pdf. Accessed October 2, 2014.237
  62. Cooper K, Stewart K. Does money affect children's outcomes? A systematic review. York, England: Joseph Rowntree Foundation; 2013. Available at www.jrf.org.uk. Accessed October 2, 2014.
  63. Shareck M, Frohlich KL, Poland B. Reducing social inequities in health through settings-related interventions—a conceptual framework. Glob Health Promot 2013;20(2):39-52.
  64. Walking and cycling: local measures to promote walking and cycling as forms of travel or recreation. London: National Institute for Health and Care Excellence; 2012. Available at http://guidance.nice.org.uk/PH41/Guidance/pdf/English. Accessed October 2, 2014.
  65. Marteau TM, Hall PA. Breadlines, brains, and behaviour: targeting executive functioning and environments may loosen the link between demography and destiny. BMJ 2013;347:f6750.
  66. Moffitt TE, Arseneault L, Belsky D, et al. A gradient of childhood self-control predicts health, wealth, and public safety. Proc Nat Acad Sci 2011;108(7):2693- 8.
  67. Raver CC, Blair C, Willoughby M. Poverty as a predictor of 4-year-olds' executive function: new perspectives on models of differential susceptibility. Dev Psychol 2013;49(2):292-304.
  68. Mani A, Mullainathan S, Shafir E, et al. Poverty impedes cognitive function. Science 2013;341(6149):976-80.
  69. Shah AK, Mullainathan S, Shafir E. Some consequences of having too little. Science 2012;338(6107):682-5.
  70. Spears D. Economic decision-making in poverty depletes behavioral control. B E J Econ Anal Policy 2011;11(1).
  71. Vohs KD. The poor's poor mental power. Science 2013;341(6149):969-70.
  72. Vogel C, Brooks-Gunn J, Martin A, et al. Impacts of early Head Start participation on child and parent outcomes at ages 2, 3, and 5. Monogr Soc Res Child Dev 2013;78(1):36-63.
Theresa M. Marteau Theresa M. Marteau, PhD, is Director of the Behaviour and Health Research Unit in the Clinical School at the University of Cambridge, and Fellow and Director of Studies in Psychological and Behavioural Sciences at Christ's College, Cambridge. She studied social psychology at the London School of Economics and abnormal psychology at the University of Oxford. Her research interests include the development and evaluation of interventions to change behavior (principally diet, physical activity, tobacco and alcohol consumption) to improve population health and reduce health inequalities; risk perception and communication particularly of biomarker-derived risks and their weak links with behavior change; and the role of evidence in policy. She is a Fellow of both the Academy of Medical Sciences and the Academy of Social Sciences.
Gareth Hollands Gareth Hollands, PhD, is a Senior Research Associate in the Behaviour and Health Research Unit at the University of Cambridge. He studied Experimental Psychology at the University of Oxford, followed by work in clinical and research settings in neurological rehabilitation, mental health, and HIV/AIDS. He then completed an MSc (University College London) and a PhD (Institute of Psychiatry, King's College London), both in Health Psychology. His research interests include the effects of communicating risk information on health behaviour, development and testing of interventions that target non- conscious processes, and systematic reviews of interventions to change health behaviour.
Michael P. Kelly Michael P. Kelly, PhD, is Honorary Senior Visiting Fellow in the General Practice and Primary Care Research Unit at the Institute of Public Health, the University of Cambridge and Honorary Professor at University College London. Professor Kelly is a public health practitioner, researcher, and academic. His research interests are in evidence-based approaches to health improvement, health inequalities, behavior change, methodological problems in public health research, evidence synthesis, coronary heart disease prevention, chronic illness, disability, physical activity, public dental health, and community involvement in health promotion. From 2005-2008, he was the co-leader of the Measurement and Evidence Knowledge Network of the World Health Organization's Commission on the Social Determinants of Health. From 2005-2014 he was Director of the Centre for Public Health at the National Institute for Health and Care Excellence (NICE).
Page last reviewed July 2015
Page originally created August 2015
Internet Citation: Changing Population Behavior and Reducing Health Disparities: Exploring the Potential of "Choice Architecture" Interventions. Content last reviewed July 2015. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/professionals/education/curriculum-tools/population-health/marteau.html