Regular consumption of fruit and vegetables (F&V) has been associated with a decreased risk of developing many forms of cancerReference Riboli and Norat1 and coronary heart diseaseReference Joshipura, Hu, Manson, Stampfer, Rimm and Speizer2. Yet despite these well-established health benefits, both childrenReference Brady, Lindquist, Herd and Goran3 and adultsReference Stables, Subar, Patterson, Dodd, Heimendinger and Van Duyn4 in the USA and many other Western countries, such as the UKReference Wardle, Jarvis, Steggles, Sutton, Williamson and Farrimond5, Reference Hoare, Henderson, Bates, Prentice, Birch and Swan6, are not meeting national guidelines for F&V consumption. Consequently there is a need to develop effective local, regional and national strategies to increase F&V consumption.
The mediating variable model suggests that changes in a behaviour are a function of changes in the variables that directly impact the behaviourReference Baranowski, Anderson and Carmack7, Reference Baranowski, Lin, Wetter, Resnicow and Hearn8. Thus, in order to change F&V consumption the factors that impact F&V consumption need to be changed. While many variables such as preferencesReference Blanchette and Brug9 and self-efficacyReference Steptoe, Perkins-Porras, McKay, Rink, Hilton and Cappuccio10 have been associated with F&V consumption, perhaps the strongest predictor has been home availabilityReference Blanchette and Brug9, Reference Bere and Klepp11, Reference Neumark-Sztainer, Wall, Perry and Story12. Home availability of F&V is particularly interesting because of its proximal nature to consumption. Availability should be reasonably easy to manipulate and small changes in availability may yield changes in consumption that could provide important health benefits.
The model of home food management processes proposed by Campbell and DesjardinsReference Campbell and Desjardins13 indicates that food consumption is influenced by a series of hierarchical and interlinked processes that begin with the wider environment, move through household resources (F&V can only be purchased if there is sufficient money), family provision strategies for food, food acquisition (where the F&V are purchased), food supply management (how the food is stored, i.e. fresh, frozen or canned), food preparation techniques (how the food is prepared, i.e. steamed or deep fried) and end in consumption. Thus home availability plays a central role in this model, which highlights that F&V are only consumed if purchased, stored and prepared. In order to impact consumption, the food needs to be ‘available’ to eat.
Different studies have interpreted ‘availability’ in slightly different waysReference Bryant and Stevens14. US investigators have interpreted availability as the presence of foods in an environment, while Norwegian researchers have included availability items (such as the presence of F&V in the fridge) in an accessibility measure. This difference can be confusing, as US researchers interpret accessibility as not just whether a food is available in the home but whether it is in a form, location and time that makes consumption easyReference Cullen, Baranowski, Owens, Marsh, Rittenberry and de Moor15. In earlier public health research alcohol and tobacco availability measures assessed the extent to which these two drugs could be obtained from different sources such as vending machines and storesReference Klepp, Jones-Webb, Wagenaar, Short, Murray and Forster16. Extrapolation suggests that availability should address the extent to which F&V are available in a defined location. Consequently, we adopted the definition of Cullen and colleaguesReference Cullen, Baranowski, Owens, Marsh, Rittenberry and de Moor15 that ‘availability concerns whether foods of interest (fruit and vegetables) are present in an environment’ (p. 616) such as in the home, school or work cafeteria. With this definition in mind the aims of the present paper are to: (1) illustrate associations between food availability and consumption; (2) highlight how increasing availability has been used to increase consumption; (3) propose the mechanisms by which availability is associated with consumption; and (4) highlight the areas in which further research is required.
Methods
A literature search was conducted for studies examining F&V availability and consumption. Computerised searches of PubMed and PsycINFO were conducted in December 2005 using ‘availability, accessibility, fruit and vegetables’ as keywords. These searches were supplemented by additional follow-up manual searches and examination of the authors’ extensive personal records. All of the papers were reviewed to ensure that articles assessed aspects of F&V availability. Articles that met this criterion were summarised and placed into tables that were grouped by study design.
Results
Qualitative studies
Studies that included qualitative reports of associations between F&V availability and intake are summarised in Table 1Reference Baranowski, Domel, Gould, Baranowski, Leonard and Treiber17–Reference Wind, Bobelijn, De Bourdeaudhuij and Klepp24. Home F&V availability was perceived to be associated with intake among 4th–6th grade childrenReference Baranowski, Domel, Gould, Baranowski, Leonard and Treiber17, Reference Cullen, Baranowski, Rittenberry and Olvera21. Similar perceptions were reported among 7th and 10th grade students in relation to healthy foods (including F&V)Reference Neumark-Sztainer, Story, Perry and Casey20. Among Native American adults, qualitative research showed that the limited availability of fresh F&V on the reservation was perceived to hinder consumptionReference Vastine, Gittelsohn, Ethelbah, Anliker and Caballero22. The availability of F&V also appears to be influenced by broader social and demographic characteristics. Although F&V were available in most children's homes, the degree of availability differed by socio-economic status (SES) but not ethnicityReference Kirby, Baranowski, Reynolds, Taylor and Binkley18, Reference Cullen, Baranowski, Baranowski, Warnecke, de Moor and Nwachokor19, Reference Cullen, Baranowski, Rittenberry and Olvera21. Children from higher-income homes were more likely to have a greater variety of fresh F&V available while children from lower-income homes were likely to have less choice and more canned and frozen foodsReference Kirby, Baranowski, Reynolds, Taylor and Binkley18. The types of foods available were also different in urban and rural homesReference Kirby, Baranowski, Reynolds, Taylor and Binkley18. Children reported that F&V were not sold as alternative items during school lunch, and hence their lack of availability at this location likely hindered consumptionReference Cullen, Baranowski, Baranowski, Warnecke, de Moor and Nwachokor19. Collectively these studies provide qualitative support for an association between F&V availability and consumption, but suggest that availability may be affected by broader sociodemographic characteristics such as income and location (urban, rural or reservation).
SES – socio-economic status.
Cross-sectional studies
Cross-sectional assessments of F&V availability are summarised in Table 2Reference Hearn, Baranowski, Baranowski, Doyle, Smith and Lin25–Reference Befort, Kaur, Nollen, Sullivan, Nazir and Choi32. Hearn and colleagues first reported a positive association between home availability and consumption among 3rd grade studentsReference Hearn, Baranowski, Baranowski, Doyle, Smith and Lin25. The same study also showed that the availability of F&V at school lunch predicted intake at this specific eating venue. Further, although home availability did not differ by SES or ethnicity, schools that provided a higher number of subsidised lunches (an indicator of SES) tended to offer fewer servings of fruitReference Hearn, Baranowski, Baranowski, Doyle, Smith and Lin25. F&V availability also predicted intake among 218 low-income mothersReference Quan, Saloman, Nitzke and Reicks26. Thus, these studies and the qualitative research (above) suggest that the associations among availability at home and school and SES are not clear and more research is needed in this area.
SES – socio-economic status.
Home availability of F&V was associated with self-reported intake among girls, but not boysReference Cullen, Baranowski, Owens, Marsh, Rittenberry and de Moor15, Reference Hanson, Neumark-Sztainer, Eisenberg, Story and Wall29. The cause of this gender discrepancy is not clear, but the results suggest that targeting increased availability may be less likely to be effective at increasing the consumption of boys and therefore gender-specific strategies may be necessary. A recent study has reported that parent-reported home availability was associated with fruit intake among white adolescents but not among African-American adolescents, while vegetable availability was not associated with intake in either ethnic groupReference Befort, Kaur, Nollen, Sullivan, Nazir and Choi32. This finding conflicts with the earlier qualitative work and suggests that further examination of ethnic differences is also warranted.
Home F&V availability has been associated with psychosocial predictors of intake. Among 4th grade students participating in the High 5 Alabama programme, mean consumption of F&V increased with availability, as did knowledge about F&V and F&V self-efficacyReference Kratt, Reynolds and Shewchuk27. Since knowledge and self-efficacy were strong predictors of intake in homes with high availability but not low availability, availability appears to have moderated the associations between self-efficacy, knowledge and F&V intakeReference Kratt, Reynolds and Shewchuk27. Availability was the strongest predictor of F&V intake among middle-school students and was also associated with parental control, parental modelling, parental support and self-efficacyReference Young, Fors and Hayes28. The association between parental modelling and support of F&V consumption and childhood intake was moderated by home availability. Thus, parents were more likely to consume F&V and encourage their children to do so if F&V were available in the home. These two studies demonstrate that availability is not only associated with intake, but it also moderates the association between other psychosocial variables and intake. Home availability of F&V is therefore both directly and indirectly associated with children's consumption.
One of the most consistent psychosocial predictors of F&V intake has been preferencesReference Blanchette and Brug9, Reference Bere and Klepp33, but the association between preferences and availability was not clear. Among 225 5th and 6th grade students, availability alone was a significant predictor in a model that accounted for 11% of the variance in F&V intake among participants with high preferences for F&VReference Cullen, Baranowski, Owens, Marsh, Rittenberry and de Moor15. However, among those with low preferences, both availability and accessibility were significant predictors in a model that accounted for 23% of the variance. Thus, the association between intake and availability was moderated by preferences. F&V availability was associated with taste preferences among middle-school studentsReference Neumark-Sztainer, Wall, Perry and Story12; when availability was low intake did not differ by preferences, but when preferences were low, intake increased among all groups if F&V were available. Thus, while both studies support associations between preferences, intake and availability, the nature of the relationships is not clear and requires further clarification.
Cross-sectional associations between dietary fat availability and intake
Support for the association between availability and intake was provided by similar findings in relation to the availability and consumption of dietary fat. In a random digital dial survey the presence of high-fat foods in the pantry correlated with the percentage of energy consumed from fatReference Patterson, Kristal, Shannon, Hunt and White34. A similar observation was reported among North American Chinese womenReference Satia, Patterson, Kristal, Hislop and Pineda35. Interestingly, the availability of high-fat foods was associated with both intake and preferences, and preferences moderated the relationship between availability and intakeReference Raynor, Polley, Wing and Jeffery36. These studies therefore provide support for the findings of Cullen et al. Reference Cullen, Baranowski, Owens, Marsh, Rittenberry and de Moor15 and suggest that further examination of these two key variables and their interaction to predict intake of foods is warranted.
Associations between grocery store and restaurant availability and consumption
Studies that assessed associations between grocery store or restaurant F&V availability and intake are summarised in Table 3Reference McKie, Clark, Maclellan and Skerratt37–Reference Rose and Richards41. In a qualitative study of participants residing in a Scottish Island community, the availability of fresh F&V in this extreme location was limited; the foods were expensive and of poor quality; and the participants felt that limited availability was a major factor in their consumption of an undesirable dietReference McKie, Clark, Maclellan and Skerratt37. Availability of fruit, juice and vegetables at restaurants in the census tract within which boy scouts resided was associated with fruit, juice and vegetable consumptionReference Edmonds, Baranowski, Baranowski, Cullen and Myres39. Since Scottish researchers have shown that foods such as orange juice and tinned tomatoes were less available in lower-income postcodes than higher-income placesReference Cummins and Macintyre40, it is possible that census tract differences are a function of SES. Alternatively, it is possible that differences could be a function of broader environmental features such as location and town planning. This hypothesis is supported by a secondary analysis of the US national Food Stamp Program, in which the distance to the nearest supermarket was inversely associated with F&V intakeReference Rose and Richards41. Participants who lived more than 5 miles from the nearest supermarket consumed 62 g of fruit and 36 g of vegetables per day less than those residing closer to the store. This equated to approximately 1 serving of fruit and half a serving of vegetables, and thus greater distance was associated with a combined 1.5 fewer servings per day. It is important to note that the association between the availability of F&V in the local neighbourhood and consumption is consistent with the hierarchical model of Campbell and DesjardinsReference Campbell and Desjardins13 and provides support for wider environmental features influencing home availability which in turn influences consumption. However, the pathways from local environment availability to home availability to consumption still need to be elucidated.
FJV – fruit, juice and vegetables.
Intervention studies
Studies that have either examined how to modify availability to increase consumption or conducted interventions that have focused on increasing availability are summarised in Table 4Reference Baranowski, Domel, Gould, Baranowski, Leonard and Treiber17Reference Bere and Klepp33Reference Baranowski, Davis, Resnicow, Baranowski, Doyle and Lin42–Reference Bere, Veierod, Bjelland and Klepp46. Focus groups conducted with 4th and 5th grade students and teachersReference Baranowski, Domel, Gould, Baranowski, Leonard and Treiber17 indicated that children influence the types and amounts of F&V that parents purchase and keep within the home, and thus enhancing children's ability to ask for F&V (in a polite manner at an appropriate time) may result in increased availability.
Incorporating the development of asking skills (via role play) into an intervention for elementary-school children resulted in a significant 0.2 serving difference in the F&V intake of intervention over control schoolsReference Baranowski, Davis, Resnicow, Baranowski, Doyle and Lin42, Reference Davis, Baranowski, Resnicow, Baranowski, Doyle and Smith43. Parental interviews also indicated that F&V home availability increased as a function of participating in the interventionReference Davis, Baranowski, Resnicow, Baranowski, Doyle and Smith43. Although the small increase in F&V servings and the relatively small number of schools in the study (the school was the unit of analysis) prevented a mediation analysis, it seems likely that the change in availability mediated the effect of this intervention. A similar finding was reported in a subsequent intervention that focused on developing asking skills among urban boy scouts, which resulted in a 0.8 serving increase in F&V intakeReference Baranowski, Baranowski, Cullen, deMoor, Rittenberry and Hebert44. The study also reported a significant intervention effect for home availability, suggesting that home availability mediated change in F&V intake among adolescents. Finally, the authors of a recent Norwegian studyReference Bere, Veierod, Bjelland and Klepp46 that attempted to increase availability via a curriculum and parent intervention in part attributed their failure to impact consumption to an inability to change home F&V availability.
Longitudinal association
Longitudinal support for the association between the availability of F&V and intake was obtained from a study (Table 4) which showed that a measure including F&V availability predicted intake at baseline and 18 months laterReference Bere and Klepp33. The study also found that changes in home and school availability and accessibility (combined measures) correlated with changes in intake and that baseline availability moderated the relationship between change in F&V preferences and intake. Thus, the associations between availability and intake were maintained over time, availability functioned as a mediator of change, and the interactions between preferences and availability were important in understanding and developing strategies to change children's F&V consumption.
Influences on home availability
A recent study of adult food shoppers showed that three measures of the perceived benefits of purchasing F&V (outcome expectancies) were correlated with home F&V availabilityReference Baranowski, Watson, Missaghian, Broadfoot, Baranowski and Cullen30. Measures of social supportReference Baranowski, Missaghian, Watson, Broadfoot, Cullen and Nicklas31, pantry management practicesReference Baranowski, Missaghian, Watson, Broadfoot, Cullen and Nicklas31 and outcome expectanciesReference Baranowski, Watson, Missaghian, Broadfoot, Baranowski and Cullen30 were also correlated with home F&V availability. Frequent food shopping would enable fresh F&V to be readily available within the home, while less frequent shopping is likely to result in either less fresh F&V or more canned and frozen F&VReference Yoo, Baranowski, Missaghian, Baranowski, Cullen and Fisher47. The most common food shopping pattern was one big weekly trip to the grocery store with a few additional small trips. However, while 35% of the participants adopted this pattern, a small but potentially very important 8% of participants shopped once a month, while 6% only shopped biweeklyReference Yoo, Baranowski, Missaghian, Baranowski, Cullen and Fisher47. While no significant socio-economic differences were found for overall frequency of food shopping, lower-income respondents were more likely to obtain F&V from convenience stores while higher-income participants were more likely to purchase F&V from restaurants for home consumption. These findings suggest that strategies to increase home availability of F&V could be tailored based on home food management practices. An improved understanding of purchasing patterns and associations with the availability for different types of F&V, as well as participants’ preferences for canned, fresh or frozen foods, would facilitate this tailoring as it would allow messages to be created that focus on increasing the availability of the foods that families are more likely to purchase. This improved understanding could also be used to further develop children's asking skills by either developing techniques to encourage their parents to purchase more fresh F&V (i.e. change their shopping practices) or by encouraging their parents to purchase more canned or frozen foods.
Discussion
This review demonstrated that the home availability of F&V was associated with intake among children, adolescents and adults, and the association between intake and consumption was maintained over time. The association between availability and intake was complex and likely to differ by demographic and psychosocial variables. A greater understanding of the mechanisms by which availability may impact consumption would aid the design of more effective programmes that attempt to increase consumption via availability.
A key research issue is to clarify the direction of causality. Do children eat more F&V because they are available, or do parents make F&V available because they know their children will eat them? This issue could be addressed by simply providing F&V to families, assessing what happens to them (e.g. consumed by whom or left spoiled) and evaluating the overall diet (e.g. whether family members eat less other F&V to compensate for the free food). The issue of the role of preferences could be addressed by systematically varying the F&V offered to reflect those known to be preferred by the child. Addressing the causality issue will determine whether home availability should be included as a mediating variable in intervention studies.
While there are many factors that could explain the association between availability and consumption there appear to be three likely explanations. First, assuming the direction of causality is from availability to consumption, availability may simply be a facilitating factor: if it is there, people may consume it; it if is not there, they cannot consume it. Wansink and ParkReference Wansink and Park48 examined the association between container size, taste and popcorn consumption and found that there was 53% greater consumption from a large container than a small container. There was no significant effect for taste and no taste by container interaction. Thus, the container size functioned as a facilitator for consumption. The same study group also reported that participants used more cooking oil when it was provided in a large container than a small container, but that this relationship was mediated by costReference Wansink49. Thus, use increased with container size only if the user perceived a cost benefit (e.g. it is ok to use more because it has reduced cost). Applying this finding to F&V suggests that selling F&V in larger, cost-saving containers may result in increased consumption.
Second, external cues such as the sight of freely available F&V may trigger increased consumption. Consumption of chocolate was influenced by availability and convenience, suggesting that the external cue of seeing the food increased consumptionReference Painter, Wansink and Hieggelke50. Extension of this finding to F&V suggests that the external cues of viewing conveniently available F&V may at least partially explain the association between availability and consumption. Moreover, storing F&V in prominent locations, in either the home or place of work, may result in increased consumption.
Third, an alternative (and yet complementary) theory is that the association between availability and consumption is a function of increased exposure. This is consistent with the desensitisation and exposure therapies used in clinical psychologyReference Tryon51. Although there is some debate as to the mechanisms by which exposure/desensitisation therapy works, repeated exposure to a phobia and anxiety-causing events has resulted in desensitisation to the event. Appetite for pizza was affected by prior exposure to pizzaReference Marcelino, Adam, Couronne, Koster and Sieffermann52, suggesting that pizza exposure influenced pizza appetite. Thus, if individuals are more regularly exposed to F&V via increased availability they may be more likely to develop an appetite and taste preference for these foods, which would result in higher intakes. This possibility is supported by research which has shown that increased exposure to a perceived unappetising vegetable resulted in an increase in the preference for that vegetable and increased consumption among 2–6-year-old childrenReference Wardle, Cooke, Gibson, Sapochnik, Sheiham and Lawson53. Similar results have also been reported for baby food among 4- to 7-month-old infantsReference Birch, Gunder, Grimm-Thomas and Laing54. It is important to note, however, that repeated monotonous exposure was associated with lower consumption of a meat sauce than repeated exposure that included varietyReference Zandstra, de Graaf and van Trijp55. Thus, while frequent exposure in the form of F&V vegetable availability may be important for intake, a strategy that promotes increasing the availability of a wide variety F&V may be more likely to yield sustained increases in consumption.
In light of this evidence, the social ecological framework may prove useful in furthering the understanding of mechanisms by which availability influences consumption and incorporating increased availability into the design of an intervention. The social ecological framework suggests that a person's behaviour is impacted by the environment in which the behaviour occurs, factors unique to the individual, interpersonal factors whereby the individual interacts with his or her peers, and the interactions between these three broad categoriesReference McLeroy, Bibeau, Steckler and Glanz56, Reference Stokols57. In terms of F&V, increased availability is likely to interact with personal factors, such as taste and preferences, and interpersonal factors, such as other family members’ or peers’ consumption and beliefs, to result in increased consumption. Thus, addressing these associations may be an effective strategy to increase consumption. A recent intervention included increased availability of F&V at school lunch (environment), taste sessions to increase preferences (personal factors) and events (interpersonal factors) such as challenge weeks to promote F&V consumption in the school. The study resulted in a 0.14 serving per day (P = 0.03) difference in the F&V consumption of intervention and control schools. While this difference was small, the study provided insights into how it is possible to effectively incorporate increased availability into a social ecological framework intervention, which warrants further examination.
Future research directions
While highlighting associations between F&V availability and consumption this review raises a number of issues that require further consideration. A key task is to gain a greater understanding of the mechanisms by which availability impacts consumption. While we have proposed some mechanisms, they are untested in relation to F&V. Empirical testing of these associations is necessary to develop more effective strategies to transform increased availability into increased consumption.
The biggest gap in this literature is the assessment of policy-level interventions. Norwegian research has shown that elementary students attending a school that provided a free piece of fruit per day consumed an extra 0.7 serving of fruit compared with students in a paying school fruit programme and 0.9 of a serving more than participants without a school fruit programmeReference Bere, Veierod and Klepp58. US Senator Tom Harkin instituted the US Department of Agriculture's Fruit and Vegetable Pilot Program (FVPP) to distribute free F&V to schools in a number of statesReference Buzby, Guthrie and Kantor59. The FVPP evaluation was positive, but included mostly qualitative testimonial data. It is not clear if the FVPP led to increased F&V intake among children or displaced F&V that would have otherwise been eaten at home. However, there is a shortage of studies that have examined policy-level interventions to increase the availability of F&V within the wider community. It is also not clear whether governmental sponsoring of reduced-price F&V will increase home availability and consumption of these foods. Moreover, while it is possible that some of the costs could be offset by increased demand, the economics of such approaches are not clear. At a more local level, more research is needed to understand how to increase availability in work locations, what has to be done to increase availability in work canteens, how work refreshment areas can most effectively promote F&V, what the barriers to promotion are and how they can be overcome.
Conclusions
This review has highlighted that the availability of F&V is associated with intake, that the relationship is sustained over time, and that changes in availability are likely to mediate changes in consumption. The review has also placed availability in the context of home food shopping practices to demonstrate the broader impacts on home availability. Thus, although the manipulation of availability is likely a key method for obtaining increases in consumption, a greater understanding of the mechanisms involved in the association and the effect of policy-level interventions is needed to appreciate the feasibility of public health interventions that focus on increasing F&V availability.
Acknowledgements
Sources of funding: This paper is a publication of the United States Department of Agriculture/Agricultural Research Service (USDA/ARS) Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, Texas, and has been funded in part with federal funds from the USDA/ARS under Cooperative Agreement No. 58-6250-6001. The contents of this publication do not necessarily reflect the views or policies of the USDA, nor does mention of trade names, commercial products or organisations imply endorsement from the US government.
Conflict of interest declaration: None declared.
Authorship responsibilities: All authors participated in the conception of the paper, the search of journals and personal records, abstracting of papers, and drafting and editing of the manuscript.