Jean AdamspersName, Oliver MyttonpersName, Martin WhitepersName, Pablo MonsivaispersName (2016date). “Why Are Some Population Interventions for Diet and Obesity More Equitable and Effective Than Others? The Role of Individual Agency.” PLOS Medicinetitle·j.
Object of study — how the agencyterm demanded by population health interventions affects effectivenessterm and socioeconomic equityterm
Research question — why advice-centred diet and obesity policies may underperform and widen inequalities compared with low-agency structural interventions
Contribution — a two-axis framework separating intervention reach from the personal resourcesterm recipients must mobilize to benefit
phenomenon · claim-type · cross-domain-transfer · mechanism · claim-relation · replication-status · limitation — grounded in the full text| phenomenon | socially patterned attrition between intervention delivery and health benefit — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | |
| claim-type | critique — “Much current policy for improving diet and reducing obesity focuses on population interventions that require individuals to use a high level of individual agency to benefit. These are likely to be less effective, and less equitable, than interventions requiring less agency. Although we have focused here on diet and obesity, these arguments ring true across many areas of public health.” | |
| method — evidence | — “Our examples above focus on diet, including diet’s role in obesity, but all public health interventions could be placed on a continuum that describes the amount of agency individuals must use to benefit from an intervention. Similarly, while Rose dichotomised preventive interventions as high risk or population, it seems more likely that this too is a continuum. Local planning arrangements restricting proliferation of hot food takeaways near schools target “vulnerable populations” (i.e., children) [ 8 ]. This places such interventions somewhere between high-risk interventions targeted at those known to be at high baseline risk and population interventions available to all.” | |
| object of study — evidence | — “Jean Adams and colleagues argue that population interventions that require individuals to use a low level of agency to benefit are likely to be most effective and most equitable.” | |
| research question — evidence | — “The obvious assumption of these programmes is that advice, guidance, and encouragement will change the population’s diet and activity behaviours. We explore why this is unlikely to be the case and why such strategies are unlikely to reduce inequalities in diet and obesity. We propose a framework that captures a wider range of intervention strategies and discuss how these may be used to improve diet, reduce obesity, and tackle inequalities.” | |
| contribution — evidence | — “Public health interventions can be described according to where they lie on two continuums: the population and high-risk approaches anchor one continuum, while the other continuum captures the personal resources (or “agency”) individuals have to use to benefit from interventions.” | |
| mechanism | Each required act between delivery and outcome creates an additional point of attrition. — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | mechanisms[0] |
| mechanism | Unequal access to cognitive, psychological, time, and material resources magnifies small inequalities across intervention steps. — “There are a number of reasons why high-agency population interventions may reinforce socioeconomic inequalities. Exerting agency requires individuals to rally their cognitive, psychological, time, and material resources [ 7 ]—all of which tend to be socioeconomically patterned. More socioeconomically advantaged people, with better health literacy (a cognitive resource) [ 17 ], may find it easier to make sense of the information provided in public health messages. More affluent parents may have more strategies for resisting their children’s “pester power” for less healthy foods (a psychological resource) [ 18 ]. More affluent people are more likely to have the material resources to be able to afford more expensive but healthier foods [ 19 ] and the time resources to source and prepare them [ 20 ]. Small inequalities at each intervention step illustrated in Fig 1 may multiply to become large inequalities in outcomes [ 16 ].” | mechanisms[1] |
| mechanism | Changing environments can alter exposure or consumption without demanding conscious individual decisions. — “In contrast, population interventions that require recipients to use little or no agency to benefit may be more effective and equitable. When food manufacturers reduce the salt content of bread, decreased salt intake occurs without individuals having to consciously engage with any information or actively change their behaviour [ 9 ]. So-called “nudge” interventions are one class of lower agency interventions. In the context of changing health-related behaviours, “nudges” have been defined as changes to the placement or properties of objects at the microenvironmental level—for example, changing the placement of healthier options in a buffet [ 10 ]. Low-agency population interventions can operate at all levels and not just at the microenvironmental level.” | mechanisms[2] |
| sense | agency as recipient resource demand imposed by an intervention — “There are a number of reasons why high-agency population interventions may reinforce socioeconomic inequalities. Exerting agency requires individuals to rally their cognitive, psychological, time, and material resources [ 7 ]—all of which tend to be socioeconomically patterned. More socioeconomically advantaged people, with better health literacy (a cognitive resource) [ 17 ], may find it easier to make sense of the information provided in public health messages. More affluent parents may have more strategies for resisting their children’s “pester power” for less healthy foods (a psychological resource) [ 18 ]. More affluent people are more likely to have the material resources to be able to afford more expensive but healthier foods [ 19 ] and the time resources to source and prepare them [ 20 ]. Small inequalities at each intervention step illustrated in Fig 1 may multiply to become large inequalities in outcomes [ 16 ].” | senses[0] |
| sense | agency as a graded policy-design dimension rather than a unitary personal capacity — “Our examples above focus on diet, including diet’s role in obesity, but all public health interventions could be placed on a continuum that describes the amount of agency individuals must use to benefit from an intervention. Similarly, while Rose dichotomised preventive interventions as high risk or population, it seems more likely that this too is a continuum. Local planning arrangements restricting proliferation of hot food takeaways near schools target “vulnerable populations” (i.e., children) [ 8 ]. This places such interventions somewhere between high-risk interventions targeted at those known to be at high baseline risk and population interventions available to all.” | senses[1] |
| sense | choice as socially and materially structured behaviour rather than unconstrained autonomy — “Low-agency interventions are sometimes considered to be synonymous with limiting free choice [ 11 ]. However, it is unlikely that many people genuinely do make “free choices” about what they eat. Food “choices” are strongly influenced by habits, what food is available and affordable, and cultural norms [ 22 ]. Low-income parents often struggle to afford the fruit and vegetables they know to be important for their children’s health [ 23 ]. Using subsidies to make healthier food more affordable is a low-agency population intervention that may increase the choices available to these parents.” | senses[2] |
| sense | attention as deliberate engagement with health information required by high-agency policy — “Population interventions like Change4Life that focus on providing advice, guidance, and encouragement rely heavily on individuals being able and motivated to engage with this advice, guidance, and encouragement. These types of interventions have been described as highly “agentic” [ 7 , 8 ]: recipients must use their personal resources, or “agency,” to benefit. The effectiveness and equity of these interventions has been questioned [ 7 , 8 ].” | senses[3] |
| audit | high-agency intervention —causes→ attrition — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but attrition 'might be expected' at each step of a high-agency intervention; hedged. — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | concept_edges[0] |
| relation | high-agency intervention —causes→ attrition — More required decisions create more points at which benefit can be lost. — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | concept_edges[0] |
| weight | high-agency intervention —causes→ attrition — 0.9 — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | concept_edges[0] |
| audit | attrition —causes→ effectiveness — Audit 2026-09-28: DOUBTFUL - read against its span and kept, but attrition at each step makes high-agency interventions LESS effective; the edge lacks its direction (reduces). — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | concept_edges[1] |
| relation | attrition —causes→ effectiveness — Accumulated loss reduces overall intervention effect. — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | concept_edges[1] |
| weight | attrition —causes→ effectiveness — 0.9 — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | concept_edges[1] |
| relation | low-agency intervention —distinct-from→ free choice — Lower agency demand does not simply eliminate meaningful choice. — “Low-agency interventions are sometimes considered to be synonymous with limiting free choice [ 11 ]. However, it is unlikely that many people genuinely do make “free choices” about what they eat. Food “choices” are strongly influenced by habits, what food is available and affordable, and cultural norms [ 22 ]. Low-income parents often struggle to afford the fruit and vegetables they know to be important for their children’s health [ 23 ]. Using subsidies to make healthier food more affordable is a low-agency population intervention that may increase the choices available to these parents.” | concept_edges[2] |
| weight | low-agency intervention —distinct-from→ free choice — 0.9 — “Low-agency interventions are sometimes considered to be synonymous with limiting free choice [ 11 ]. However, it is unlikely that many people genuinely do make “free choices” about what they eat. Food “choices” are strongly influenced by habits, what food is available and affordable, and cultural norms [ 22 ]. Low-income parents often struggle to afford the fruit and vegetables they know to be important for their children’s health [ 23 ]. Using subsidies to make healthier food more affordable is a low-agency population intervention that may increase the choices available to these parents.” | concept_edges[2] |
| limitation | The effectiveness comparison is presented as some evidence rather than a definitive causal estimate. — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | limitations[0] |
| limitation | The framework does not attempt fine-grained placement of interventions along its continua. — “Two continuums describe all public health interventions, with examples related to diet and obesity. Interventions grouped together fall at similar points on the two continuums; we have not attempted to make fine-grained distinctions of where interventions fall on each continuum.” | limitations[1] |
| limitation | High-agency information may remain necessary and can act synergistically with structural measures. — “This is not to say that high-agency population interventions are not valuable. Whilst information and encouragement are often not sufficient to change diet [ 21 ], they may still be necessary in some cases. It is also possible that high- and low-agency population interventions act synergistically. For example, a hypothetical intervention involving fortification of flour with folic acid (a low-agency population intervention) that is highly publicised and discussed may raise awareness of the benefits of taking folic acid supplements during pregnancy and so prompt greater engagement with information leaflets about prenatal health (a high-agency population intervention).” | limitations[2] |
| limitation | Low-agency interventions can face political, commercial, and public-acceptability barriers. — “Public acceptability of low-agency population interventions is dependent not only on how much these interventions are felt to intrude on personal autonomy. Instead, how effective interventions are perceived to be, the importance of the public health “problem” targeted, and personal familiarity with interventions all influence intervention acceptability [ 24 ]. England’s experience with banning smoking in public places, in which acceptability increased after the ban was enforced [ 25 ], shows that public acceptability of low-agency population interventions is not immutable. The public health community needs to put more effort into learning about, and doing, effective “hearts and minds” work to build public acceptability of low-agency population interventions.” | limitations[3] |
finding · quantity · variable · definition · key-term — grounded in the full text| finding | Agency demand is an intervention property that helps determine effectiveness and distribution of benefit. — “The amount of agency individuals must use to benefit from an intervention is a fundamental determinant of how, and for whom, it will work.” | content_tags.findings[0] |
| finding | Advice and encouragement are high-agency interventions because recipients must mobilize personal resources. — “Population interventions like Change4Life that focus on providing advice, guidance, and encouragement rely heavily on individuals being able and motivated to engage with this advice, guidance, and encouragement. These types of interventions have been described as highly “agentic” [ 7 , 8 ]: recipients must use their personal resources, or “agency,” to benefit. The effectiveness and equity of these interventions has been questioned [ 7 , 8 ].” | content_tags.findings[1] |
| finding | Structural changes can lower agency demand by producing benefit without conscious engagement or active behaviour change. — “In contrast, population interventions that require recipients to use little or no agency to benefit may be more effective and equitable. When food manufacturers reduce the salt content of bread, decreased salt intake occurs without individuals having to consciously engage with any information or actively change their behaviour [ 9 ]. So-called “nudge” interventions are one class of lower agency interventions. In the context of changing health-related behaviours, “nudges” have been defined as changes to the placement or properties of objects at the microenvironmental level—for example, changing the placement of healthier options in a buffet [ 10 ]. Low-agency population interventions can operate at all levels and not just at the microenvironmental level.” | content_tags.findings[2] |
| finding | Repeated decision points create attrition that reduces the reach of high-agency interventions. — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | content_tags.findings[3] |
| finding | Socioeconomic patterning of cognitive, psychological, temporal, and material resources can convert agency demands into health inequality. — “There are a number of reasons why high-agency population interventions may reinforce socioeconomic inequalities. Exerting agency requires individuals to rally their cognitive, psychological, time, and material resources [ 7 ]—all of which tend to be socioeconomically patterned. More socioeconomically advantaged people, with better health literacy (a cognitive resource) [ 17 ], may find it easier to make sense of the information provided in public health messages. More affluent parents may have more strategies for resisting their children’s “pester power” for less healthy foods (a psychological resource) [ 18 ]. More affluent people are more likely to have the material resources to be able to afford more expensive but healthier foods [ 19 ] and the time resources to source and prepare them [ 20 ]. Small inequalities at each intervention step illustrated in Fig 1 may multiply to become large inequalities in outcomes [ 16 ].” | content_tags.findings[4] |
| finding | Low-agency interventions are predicted to improve both population effectiveness and equity. — “Low-agency population interventions are most likely to achieve the twin public health aims of preventing disease and minimising inequalities. These interventions should, therefore, form the backbone of public health strategies.” | content_tags.findings[5] |
| finding | High- and low-agency interventions can be complementary rather than mutually exclusive. — “This is not to say that high-agency population interventions are not valuable. Whilst information and encouragement are often not sufficient to change diet [ 21 ], they may still be necessary in some cases. It is also possible that high- and low-agency population interventions act synergistically. For example, a hypothetical intervention involving fortification of flour with folic acid (a low-agency population intervention) that is highly publicised and discussed may raise awareness of the benefits of taking folic acid supplements during pregnancy and so prompt greater engagement with information leaflets about prenatal health (a high-agency population intervention).” | content_tags.findings[6] |
| finding | Food choice is structured by habits, availability, affordability, and cultural norms rather than unrestricted individual autonomy. — “Low-agency interventions are sometimes considered to be synonymous with limiting free choice [ 11 ]. However, it is unlikely that many people genuinely do make “free choices” about what they eat. Food “choices” are strongly influenced by habits, what food is available and affordable, and cultural norms [ 22 ]. Low-income parents often struggle to afford the fruit and vegetables they know to be important for their children’s health [ 23 ]. Using subsidies to make healthier food more affordable is a low-agency population intervention that may increase the choices available to these parents.” | content_tags.findings[7] |
| quantity | one in three — adults described as overweight or obese — “We now find ourselves in a world in which one in three adults is overweight or obese [ 1 ]. Unhealthy diets are an important determinant of overweight and obesity and also contribute directly to a variety of other causes of disability and death. In fact, together with physical inactivity, dietary risk factors are responsible for 10% of disability-adjusted life years lost globally [ 2 ]. Like many risk factors, unhealthy diets are not evenly distributed in the population. In most developed countries, and increasingly in developing countries too, obesity and unhealthy diets are more common in more socioeconomically disadvantaged groups [ 3 , 4 ]. Socioeconomic inequalities in health and disease are at least partly due to these inequalities in diet and obesity.” | content_tags.quantities[0] |
| quantity | 10% — global disability-adjusted life years attributed to diet and physical inactivity — “We now find ourselves in a world in which one in three adults is overweight or obese [ 1 ]. Unhealthy diets are an important determinant of overweight and obesity and also contribute directly to a variety of other causes of disability and death. In fact, together with physical inactivity, dietary risk factors are responsible for 10% of disability-adjusted life years lost globally [ 2 ]. Like many risk factors, unhealthy diets are not evenly distributed in the population. In most developed countries, and increasingly in developing countries too, obesity and unhealthy diets are more common in more socioeconomically disadvantaged groups [ 3 , 4 ]. Socioeconomic inequalities in health and disease are at least partly due to these inequalities in diet and obesity.” | content_tags.quantities[1] |
| quantity | 20% — illustrative attrition at each intervention step — “Illustration of the intervention pathway in low- (top) and high-agency (bottom) population interventions. Both examples illustrate population interventions to increase folate in women trying to conceive. The top panel illustrates an information leaflet encouraging women to take folic acid supplements (a high-agency population intervention). The bottom panel illustrates the universal addition of folic acid to mainstream wheat flour (a low-agency intervention). Numbers are illustrative and indicate how many women might be in each pathway if it is hypothetically assumed that there is 20% attrition at each step. The steps shown in both cases are illustrative and not necessarily exhaustive.” | content_tags.quantities[2] |
| variable | intervention effectiveness (population outcome) — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | content_tags.variables[1] |
| variable | personal resources (socially patterned capacity to benefit) — “There are a number of reasons why high-agency population interventions may reinforce socioeconomic inequalities. Exerting agency requires individuals to rally their cognitive, psychological, time, and material resources [ 7 ]—all of which tend to be socioeconomically patterned. More socioeconomically advantaged people, with better health literacy (a cognitive resource) [ 17 ], may find it easier to make sense of the information provided in public health messages. More affluent parents may have more strategies for resisting their children’s “pester power” for less healthy foods (a psychological resource) [ 18 ]. More affluent people are more likely to have the material resources to be able to afford more expensive but healthier foods [ 19 ] and the time resources to source and prepare them [ 20 ]. Small inequalities at each intervention step illustrated in Fig 1 may multiply to become large inequalities in outcomes [ 16 ].” | content_tags.variables[3] |
| definition | agency — “There are a number of reasons why high-agency population interventions may reinforce socioeconomic inequalities. Exerting agency requires individuals to rally their cognitive, psychological, time, and material resources [ 7 ]—all of which tend to be socioeconomically patterned. More socioeconomically advantaged people, with better health literacy (a cognitive resource) [ 17 ], may find it easier to make sense of the information provided in public health messages. More affluent parents may have more strategies for resisting their children’s “pester power” for less healthy foods (a psychological resource) [ 18 ]. More affluent people are more likely to have the material resources to be able to afford more expensive but healthier foods [ 19 ] and the time resources to source and prepare them [ 20 ]. Small inequalities at each intervention step illustrated in Fig 1 may multiply to become large inequalities in outcomes [ 16 ].” | content_tags.definitions[0] |
| definition | high-agency intervention — “Population interventions like Change4Life that focus on providing advice, guidance, and encouragement rely heavily on individuals being able and motivated to engage with this advice, guidance, and encouragement. These types of interventions have been described as highly “agentic” [ 7 , 8 ]: recipients must use their personal resources, or “agency,” to benefit. The effectiveness and equity of these interventions has been questioned [ 7 , 8 ].” | content_tags.definitions[1] |
| definition | low-agency intervention — “In contrast, population interventions that require recipients to use little or no agency to benefit may be more effective and equitable. When food manufacturers reduce the salt content of bread, decreased salt intake occurs without individuals having to consciously engage with any information or actively change their behaviour [ 9 ]. So-called “nudge” interventions are one class of lower agency interventions. In the context of changing health-related behaviours, “nudges” have been defined as changes to the placement or properties of objects at the microenvironmental level—for example, changing the placement of healthier options in a buffet [ 10 ]. Low-agency population interventions can operate at all levels and not just at the microenvironmental level.” | content_tags.definitions[2] |
| definition | nudge — “In contrast, population interventions that require recipients to use little or no agency to benefit may be more effective and equitable. When food manufacturers reduce the salt content of bread, decreased salt intake occurs without individuals having to consciously engage with any information or actively change their behaviour [ 9 ]. So-called “nudge” interventions are one class of lower agency interventions. In the context of changing health-related behaviours, “nudges” have been defined as changes to the placement or properties of objects at the microenvironmental level—for example, changing the placement of healthier options in a buffet [ 10 ]. Low-agency population interventions can operate at all levels and not just at the microenvironmental level.” | content_tags.definitions[3] |
| definition | population approach — “Mass-media social marketing initiatives draw on Geoffrey Rose’s population approach to prevention [ 5 ]. This approach involves interventions that are delivered across the whole population, or whole population groups, without first identifying those at increased risk of disease. The aim of the population approach is to reduce risk factor levels by a small amount in everyone. In contrast, the high-risk approach identifies and targets those individuals at high risk of disease, aiming to reduce risk factors by a large amount in only these people. Rose argued that a greater overall benefit would be achieved by modest reductions in risk in many people than by large reductions in just a few people.” | content_tags.definitions[4] |
| concept | agency — not defined in this paper's own words; personal resources required to obtain benefit — “There are a number of reasons why high-agency population interventions may reinforce socioeconomic inequalities. Exerting agency requires individuals to rally their cognitive, psychological, time, and material resources [ 7 ]—all of which tend to be socioeconomically patterned. More socioeconomically advantaged people, with better health literacy (a cognitive resource) [ 17 ], may find it easier to make sense of the information provided in public health messages. More affluent parents may have more strategies for resisting their children’s “pester power” for less healthy foods (a psychological resource) [ 18 ]. More affluent people are more likely to have the material resources to be able to afford more expensive but healthier foods [ 19 ] and the time resources to source and prepare them [ 20 ]. Small inequalities at each intervention step illustrated in Fig 1 may multiply to become large inequalities in outcomes [ 16 ].” | concepts[0] |
| concept | high-agency intervention — not defined in this paper's own words; information-dependent action pathway — “Population interventions like Change4Life that focus on providing advice, guidance, and encouragement rely heavily on individuals being able and motivated to engage with this advice, guidance, and encouragement. These types of interventions have been described as highly “agentic” [ 7 , 8 ]: recipients must use their personal resources, or “agency,” to benefit. The effectiveness and equity of these interventions has been questioned [ 7 , 8 ].” | concepts[1] |
| concept | low-agency intervention — not defined in this paper's own words; structural benefit with little recipient action — “In contrast, population interventions that require recipients to use little or no agency to benefit may be more effective and equitable. When food manufacturers reduce the salt content of bread, decreased salt intake occurs without individuals having to consciously engage with any information or actively change their behaviour [ 9 ]. So-called “nudge” interventions are one class of lower agency interventions. In the context of changing health-related behaviours, “nudges” have been defined as changes to the placement or properties of objects at the microenvironmental level—for example, changing the placement of healthier options in a buffet [ 10 ]. Low-agency population interventions can operate at all levels and not just at the microenvironmental level.” | concepts[2] |
| concept | nudge — not defined in this paper's own words; microenvironmental choice-architecture change — “In contrast, population interventions that require recipients to use little or no agency to benefit may be more effective and equitable. When food manufacturers reduce the salt content of bread, decreased salt intake occurs without individuals having to consciously engage with any information or actively change their behaviour [ 9 ]. So-called “nudge” interventions are one class of lower agency interventions. In the context of changing health-related behaviours, “nudges” have been defined as changes to the placement or properties of objects at the microenvironmental level—for example, changing the placement of healthier options in a buffet [ 10 ]. Low-agency population interventions can operate at all levels and not just at the microenvironmental level.” | concepts[3] |
| concept | attrition — not defined in this paper's own words; loss of recipients at successive action steps — “There is some evidence that high-agency population interventions are less effective, overall, than those that require less agency [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventions for attrition at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attrition might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attrition can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency intervention) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency intervention) [ 15 ].” | concepts[4] |
| concept | equity — not defined in this paper's own words; socioeconomic distribution of intervention benefit — “Population interventions like Change4Life that focus on providing advice, guidance, and encouragement rely heavily on individuals being able and motivated to engage with this advice, guidance, and encouragement. These types of interventions have been described as highly “agentic” [ 7 , 8 ]: recipients must use their personal resources, or “agency,” to benefit. The effectiveness and equity of these interventions has been questioned [ 7 , 8 ].” | concepts[5] |
| concept | effectiveness — not defined in this paper's own words; population-level intervention benefit — “Population interventions like Change4Life that focus on providing advice, guidance, and encouragement rely heavily on individuals being able and motivated to engage with this advice, guidance, and encouragement. These types of interventions have been described as highly “agentic” [ 7 , 8 ]: recipients must use their personal resources, or “agency,” to benefit. The effectiveness and equity of these interventions has been questioned [ 7 , 8 ].” | concepts[6] |
| concept | personal resources — cognitive, psychological, time, and material resources — “There are a number of reasons why high-agency population interventions may reinforce socioeconomic inequalities. Exerting agency requires individuals to rally their cognitive, psychological, time, and material resources [ 7 ]—all of which tend to be socioeconomically patterned. More socioeconomically advantaged people, with better health literacy (a cognitive resource) [ 17 ], may find it easier to make sense of the information provided in public health messages. More affluent parents may have more strategies for resisting their children’s “pester power” for less healthy foods (a psychological resource) [ 18 ]. More affluent people are more likely to have the material resources to be able to afford more expensive but healthier foods [ 19 ] and the time resources to source and prepare them [ 20 ]. Small inequalities at each intervention step illustrated in Fig 1 may multiply to become large inequalities in outcomes [ 16 ].” | concepts[7] |
| concept | free choice — not defined in this paper's own words; purported autonomy in food selection — “Low-agency interventions are sometimes considered to be synonymous with limiting free choice [ 11 ]. However, it is unlikely that many people genuinely do make “free choices” about what they eat. Food “choices” are strongly influenced by habits, what food is available and affordable, and cultural norms [ 22 ]. Low-income parents often struggle to afford the fruit and vegetables they know to be important for their children’s health [ 23 ]. Using subsidies to make healthier food more affordable is a low-agency population intervention that may increase the choices available to these parents.” | concepts[8] |
| concept | population approach — not defined in this paper's own words; small risk reduction across many people — “Mass-media social marketing initiatives draw on Geoffrey Rose’s population approach to prevention [ 5 ]. This approach involves interventions that are delivered across the whole population, or whole population groups, without first identifying those at increased risk of disease. The aim of the population approach is to reduce risk factor levels by a small amount in everyone. In contrast, the high-risk approach identifies and targets those individuals at high risk of disease, aiming to reduce risk factors by a large amount in only these people. Rose argued that a greater overall benefit would be achieved by modest reductions in risk in many people than by large reductions in just a few people.” | concepts[9] |
Full text — tagged
Why Are Some Population Interventions for Diet and Obesity More Equitable and Effective Than Others? The Role of Individual Agencyterm Abstract Jean AdamspersName and colleagues argue that population interventions that require individuals to use a low level of agencyterm to benefit are likely to be most effective and most equitable.
Summary Points Public health interventions can be described according to where they lie on two continuums: the population and high-risk approaches anchor one continuum, while the other continuum captures the personal resourcesterm (or “agencyterm”) individuals have to use to benefit from interventions.
Population interventions that require individuals to use a high level of agencyterm to benefit tend to be favoured by governments around the world.
Population interventions that require individuals to use a low level of agencyterm to benefit are likely to be most effective and most equitable. More effort is required to develop, evaluate, and implement population interventions that require low levels of agencyterm for individuals to benefit.
We now find ourselves in a world in which one in three adults is overweight or obese [ 1 ]. Unhealthy diets are an important determinant of overweight and obesity and also contribute directly to a variety of other causes of disability and death. In fact, together with physical inactivity, dietary risk factors are responsible for 10% of disability-adjusted life years lost globally [ 2 ]. Like many risk factors, unhealthy diets are not evenly distributed in the population. In most developed countries, and increasingly in developing countries too, obesity and unhealthy diets are more common in more socioeconomically disadvantaged groups [ 3 , 4 ]. Socioeconomic inequalities in health and disease are at least partly due to these inequalities in diet and obesity.
Many governments have responded to the challenge of unhealthy diets and obesity with strategies that focus on advice, guidance, and encouragement to adopt healthier lifestyles. In England, the government’s current “flagship” obesity prevention programme is Change4Life. This social marketing campaign uses mass media and other avenues to inform and educate the public about the harms of excess weight and the benefits of being more active and eating more healthfully. Simple strategies for changing behaviour are also offered. Similar programmes operate elsewhere, including MangerBouger in France and Let’s Move in the United States.
The obvious assumption of these programmes is that advice, guidance, and encouragement will change the population’s diet and activity behaviours. We explore why this is unlikely to be the case and why such strategies are unlikely to reduce inequalities in diet and obesity. We propose a framework that captures a wider range of intervention strategies and discuss how these may be used to improve diet, reduce obesity, and tackle inequalities.
The Population and High-Risk Approaches to Prevention Mass-media social marketing initiatives draw on Geoffrey Rose’s population approachterm to prevention [ 5 ]. This approach involves interventions that are delivered across the whole population, or whole population groups, without first identifying those at increased risk of disease. The aim of the population approachterm is to reduce risk factor levels by a small amount in everyone. In contrast, the high-risk approach identifies and targets those individuals at high risk of disease, aiming to reduce risk factors by a large amount in only these people. Rose argued that a greater overall benefit would be achieved by modest reductions in risk in many people than by large reductions in just a few people.
Unfortunately, not all population approachesterm are the same [ 6 ]. Population approachesterm that rely on individuals engaging with information are unlikely to achieve the full potential of Rose’s insights.
Population Interventions Differ in the Degree of “Agencyterm” Individuals Must Use to Benefit Population interventions like Change4Life that focus on providing advice, guidance, and encouragement rely heavily on individuals being able and motivated to engage with this advice, guidance, and encouragement. These types of interventions have been described as highly “agentic” [ 7 , 8 ]: recipients must use their personal resourcesterm, or “agencyterm,” to benefit. The effectivenessterm and equityterm of these interventions has been questioned [ 7 , 8 ].
In contrast, population interventions that require recipients to use little or no agencyterm to benefit may be more effective and equitable. When food manufacturers reduce the salt content of bread, decreased salt intake occurs without individuals having to consciously engage with any information or actively change their behaviour [ 9 ]. So-called “nudgeterm” interventions are one class of lower agencyterm interventions. In the context of changing health-related behaviours, “nudgesterm” have been defined as changes to the placement or properties of objects at the microenvironmental level—for example, changing the placement of healthier options in a buffet [ 10 ]. Low-agency population interventions can operate at all levels and not just at the microenvironmental level.
The amount of agencyterm individuals must use to benefit from an intervention is a fundamental determinant of how, and for whom, it will work.
Intervention “Agencyterm” Influences Intervention Effectivenessterm·var and Equityterm There is some evidence that high-agency population interventions are less effective, overall, than those that require less agencyterm [ 7 , 11 , 12 ]. By reducing the need for individual decisions, there is less room in low-agency interventionsterm for attritionterm at each of the many steps from intervention delivery to health outcome. For example, for an information leaflet encouraging women trying to conceive to take folic acid supplements (a high-agency population intervention) to have an effect, women must see and read the leaflet, understand the information presented, and then purchase and take supplements. Attritionterm might be expected at each of these steps. In contrast, if all commercial wheat flour is fortified with folic acid (a low-agency population intervention), one of the few, if only, steps at which attritionterm can occur is the decision to keep consuming products made with commercial wheat flour. This is illustrated schematically in Fig 1 . There is also accumulating evidence that high-agency population interventions are likely to reinforce, or even worsen, socioeconomic inequalities in health [ 13 – 16 ]. For example, information campaigns about the importance of folic acid intake during pregnancy (a high-agency interventionterm) exacerbate socioeconomic inequalities in folate status more than supplementing food with folic acid (a low-agency interventionterm) [ 15 ].
Illustration of the intervention pathway in low- (top) and high-agency (bottom) population interventions. Both examples illustrate population interventions to increase folate in women trying to conceive. The top panel illustrates an information leaflet encouraging women to take folic acid supplements (a high-agency population intervention). The bottom panel illustrates the universal addition of folic acid to mainstream wheat flour (a low-agency interventionterm). Numbers are illustrative and indicate how many women might be in each pathway if it is hypothetically assumed that there is 20% attritionterm at each step. The steps shown in both cases are illustrative and not necessarily exhaustive.
There are a number of reasons why high-agency population interventions may reinforce socioeconomic inequalities. Exerting agencyterm requires individuals to rally their cognitive, psychological, time, and material resources [ 7 ]—all of which tend to be socioeconomically patterned. More socioeconomically advantaged people, with better health literacy (a cognitive resource) [ 17 ], may find it easier to make sense of the information provided in public health messages. More affluent parents may have more strategies for resisting their children’s “pester power” for less healthy foods (a psychological resource) [ 18 ]. More affluent people are more likely to have the material resources to be able to afford more expensive but healthier foods [ 19 ] and the time resources to source and prepare them [ 20 ]. Small inequalities at each intervention step illustrated in Fig 1 may multiply to become large inequalities in outcomes [ 16 ].
A New Framework for Public Health Interventions Our examples above focus on diet, including diet’s role in obesity, but all public health interventions could be placed on a continuum that describes the amount of agencyterm individuals must use to benefit from an intervention. Similarly, while Rose dichotomised preventive interventions as high risk or population, it seems more likely that this too is a continuum. Local planning arrangements restricting proliferation of hot food takeaways near schools target “vulnerable populations” (i.e., children) [ 8 ]. This places such interventions somewhere between high-risk interventions targeted at those known to be at high baseline risk and population interventions available to all.
When combined, these two continuums provide a framework for describing the whole range of public health interventions. Fig 2 illustrates this with examples of diet and obesity interventions.
Two continuums describe all public health interventions, with examples related to diet and obesity. Interventions grouped together fall at similar points on the two continuums; we have not attempted to make fine-grained distinctions of where interventions fall on each continuum.
Low-Agency Population Interventions Should Be the Backbone of Public Health Strategies Low-agency population interventions are most likely to achieve the twin public health aims of preventing disease and minimising inequalities. These interventions should, therefore, form the backbone of public health strategies.
This is not to say that high-agency population interventions are not valuable. Whilst information and encouragement are often not sufficient to change diet [ 21 ], they may still be necessary in some cases. It is also possible that high- and low-agency population interventions act synergistically. For example, a hypothetical intervention involving fortification of flour with folic acid (a low-agency population intervention) that is highly publicised and discussed may raise awareness of the benefits of taking folic acid supplements during pregnancy and so prompt greater engagement with information leaflets about prenatal health (a high-agency population intervention).
Why Are Low-Agency Population Interventions Underused?
Given the clear value of low-agency population interventions, why do high-agency ones still dominate? One reason may be a perception that low-agency population interventions are less acceptable to various interested parties. In the case of diet and obesity interventions, these parties include politicians, who may enact interventions; the public, who are the recipients of interventions; and food companies, whose commercial interests may be affected by interventions.
Low-agency interventionsterm are sometimes considered to be synonymous with limiting free choiceterm [ 11 ]. However, it is unlikely that many people genuinely do make “free choicesterm” about what they eat. Food “choices” are strongly influenced by habits, what food is available and affordable, and cultural norms [ 22 ]. Low-income parents often struggle to afford the fruit and vegetables they know to be important for their children’s health [ 23 ]. Using subsidies to make healthier food more affordable is a low-agency population intervention that may increase the choices available to these parents.
Public acceptability of low-agency population interventions is dependent not only on how much these interventions are felt to intrude on personal autonomy. Instead, how effective interventions are perceived to be, the importance of the public health “problem” targeted, and personal familiarity with interventions all influence intervention acceptability [ 24 ]. England’s experience with banning smoking in public places, in which acceptability increased after the ban was enforced [ 25 ], shows that public acceptability of low-agency population interventions is not immutable. The public health community needs to put more effort into learning about, and doing, effective “hearts and minds” work to build public acceptability of low-agency population interventions.
It may seem obvious that low-agency population interventions would be particularly unacceptable to food suppliers, manufacturers, and retailers with vested commercial interests in maintaining current dietary patterns. There is now growing scientific evidence confirming the efforts that those with commercial vested interests can make to avoid low-agency population interventions and the impact these efforts can have on policy [ 26 ]. Organisations with vested commercial interests may also have substantial resources at their disposal to circumvent, or undermine, any low-agency public health interventions that are implemented. There is often public cynicism about the actions of food companies [ 27 ]. This cynicism could be exploited to give politicians and policy makers the “space” they need to enact low-agency population interventions.
Conclusion Much current policy for improving diet and reducing obesity focuses on population interventions that require individuals to use a high level of individual agencyterm to benefit. These are likely to be less effective, and less equitable, than interventions requiring less agencyterm. Although we have focused here on diet and obesity, these arguments ring true across many areas of public health.
Whilst action across the full spectrum of the framework described in Fig 2 is required, more attention should be given to the development and implementation of low-agency population interventions. Many before us have made similar arguments (e.g., [ 7 , 21 , 28 ]), yet action to implement low-agency population interventions remains limited. Politicians, civil servants, researchers, and others may shy away from low-agency interventionsterm, as they see these as more politically difficult. We all need to have more courage to argue the case that these interventions can be publically acceptable, support people to live healthier lives, and reduce inequalities in health. Low-agency population interventions should be central to public health action on diet and obesity.
Acknowledgments We thank David Ogilvie, Jenna Panter, and Nick Wareham for providing helpful comments on a previous version of this manuscript.
Funding Statement This work was undertaken by the Centre for Diet and Activity Research (CEDAR), a UKCRC Public Health Research Centre of Excellence. Funding for CEDAR from the British Heart Foundation, Cancer Research UK, Economic and Social Research Council, Medical Research Council, the National Institute for Health Research, and the Wellcome Trust, under the auspices of the UK Clinical Research Collaboration, is gratefully acknowledged. The CEDAR grant is managed by the Medical Research Council (grant code MR/K023187/1) and the principal applicant is Prof NJ Wareham (who is not an author on this paper). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Footnotes References