Fruit and Vegetable Availability and Selection

Fruit and Vegetable Availability and Selection

Fruit and Vegetable Availability and Selection Federal Food Package Revisions, 2009 Shannon N. Zenk, PhD, MPH, RN, Angela Odoms-Young, PhD, Lisa M. Po...

232KB Sizes 6 Downloads 80 Views

Fruit and Vegetable Availability and Selection Federal Food Package Revisions, 2009 Shannon N. Zenk, PhD, MPH, RN, Angela Odoms-Young, PhD, Lisa M. Powell, PhD, Richard T. Campbell, PhD, Daniel Block, PhD, Noel Chavez, PhD, RD, LDN, Ramona C. Krauss, MA, Steven Strode, James Armbruster

Background: With nearly 49,000 authorized retailers nationwide, a policy change that added fruits and vegetables (FV) to the U.S. Department of Agriculture’s Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) food packages in 2009 had the potential to expand neighborhood FV availability.

Purpose: This study examined changes in availability and selection of commonly consumed and culturally specifıc FV at authorized retailers (WIC vendors) before and after implementation of the revised WIC food packages. Methods: Quasi-experimental, one-group design with two pre-policy observations and one post-policy observation. Trained observers assessed a list of fresh, frozen, and canned FV at each vendor in seven northern Illinois counties. Eight indices of FV availability and selection were derived. Multiple regression estimated relationships. Data were collected in 2008 –2010 and analyzed in 2011.

Results: Overall, availability and selection of commonly consumed fresh FV and availability of African-American culturally specifıc fresh FV improved after implementation of the new policy. Modest improvements in the overall availability of canned low-sodium vegetables and frozen FV were observed. Changes differed by vendor type (large vendor, small vendor, and pharmacy). Changes in availability or selection did not differ by neighborhood characteristics (population density, median household income, racial/ethnic composition). Conclusions: Expansion of WIC foods was associated with small positive externalities on the food environment. Larger subsidies to create more demand and more-substantial stocking requirements for retailers may yield signifıcantly larger improvements and thus warrant further investigation. Approaches targeting rural, low-income, and racial/ethnic minority neighborhoods also may be needed. (Am J Prev Med 2012;43(4):423– 428) © 2012 American Journal of Preventive Medicine



ncreasing access to healthy foods through federal food and nutrition assistance programs, such as the U.S. Department of Agriculture’s Special Supple-

From the Department of Health Systems Science (Zenk), Department of Kinesiology and Nutrition (Odoms-Young), Institute for Health Research and Policy (Powell, Campbell), Community Health Sciences Division (Chavez), Department of Economics (Krauss), University of Illinois at Chicago; Department of Geography, Sociology, History, African-American Studies, and Anthropology (Block), Chicago State University, Chicago; and Bureau of Family Nutrition (Armbruster), Illinois Department of Human Services, Springfıeld, Illinois Address correspondence to: Shannon N. Zenk, PhD, MPH, RN, University of Illinois at Chicago, 845 S. Damen Ave., Chicago IL 60612. E-mail: [email protected] 0749-3797/$36.00

mental Nutrition Program for Women, Infants, and Children (WIC), may benefıt low-income populations, improving their diet and health outcomes, and thereby reduce disparities.1 In 2009, WIC aligned its food packages with the 2005 Dietary Guidelines for Americans and other guidelines.2– 4 Typically redeemed at authorized retailers (“WIC vendors”), the revised WIC food packages added a cashvalue voucher of $6 –$10 monthly per participant to purchase fruits and vegetables (FV). Although the primary policy aim was to improve WIC participant diet and health, with nearly 49,000 WIC vendors nationwide, the new requirement for these vendors to carry FV had the potential to expand availability, especially at small retailers and in rural, low-income, and minority neighborhoods that previously offered few high-quality FV.5–11

© 2012 American Journal of Preventive Medicine • Published by Elsevier Inc.

Am J Prev Med 2012;43(4):423– 428 423


Zenk et al / Am J Prev Med 2012;43(4):423– 428

Estimated probability (availability outcome) or proportion (selection outcome)

analytic sample was all WIC vendors in seven northern Illinois counties that were authorized in 2008 and that joined WIC in 2009 and 2010. It was an unbalanced panel consisting of stores that were ---------Pre-policy---------------Post-policy-------WIC authorized in a given year. change change Trained observers assessed FV at vendors using an instrument adapted from prior work.7,13–15 Each year, two observers independently assessed 10% of vendors in order to evaluate inter-rater reliability, which was high. Eight FV indices were derived: commonly consumed fresh FV, culturally specifıc African-American FV, culturally specifıc Latino FV, frozen vegetables, frozen fruits, canned vegetables; canned vegetables, no added salt; and canned fruits. “Availability” was defıned as the presence of one or more varieties; “selection” was defıned as the proportion of assessed varieties present. Vendors were classifıed as pharmacies, small vendors, and large vendors. Vendors were geocoded to 2000 census block groups O O X O (neighborhoods) and characterized based on population density, (T1; 2008) (T2; 2009) (T3; 2010) median household income, and racial/ethnic composition.16 Seasonality was controlled based on days from the date of data collection to July 1 of the data collection year. O = observation (data collection) point Analyses were conducted in 2011 in Stata, version 11. Binomial X = intervention point (policy change; August 3, 2009) T = time point logit models were estimated for dichotomous availability outcomes; fractional logit models were used for proportional selection Figure 1. Graphical depiction of study design outcomes.17 Each outcome was fırst regressed on data collection Note: Although there is no control group, this design permits an evaluation of year, vendor type, neighborhood characteristics, and covariates a short time series in which post-policy changes from 2009 to 2010 are (Model 1). Interactions between vendor type and year were then examined relative to pre-policy changes occurring between 2008 and 2009. added (Model 2). Follow-up tests of model coeffıcients were conThus, it is possible to evaluate the post-policy difference relative to “backducted to determine if pre- to post-policy change in the outcome ground noise” prior to the policy change. The 2008 data collection was was greater than the change in the previous year overall (Question 1; completed April–October, with 38% of stores observed in June–July and Model 1) and by vendor type (Question 2; Model 2). The comparanother 47% in August–September. The 2009 and 2010 data collections were ison of interest was (2010 –2009) ⫺ (2009 –2008). To test whether both completed in June–July. the pre- to post-policy outcome change differed by neighborhood characteristics (Question 2), The current study examined changes in Model 2 was re-estimated, including multiplicaFV availability and selection at WIC venSee tive interactions between year and each neighbordors before and after implementation of the related hood characteristic in separate models.

revised WIC food packages in Illinois. It Commentary by addressed the following questions: (1) Did Laraia in this Results issue. FV availability and selection improve preTable 1 shows descriptive statistics of to post-policy, including FV commonly vendors by type and year. Table 2 shows consumed in the general U.S. population, as well as among African Americans and Latioutcomes at baseline by vendor type. Apnos? (2) Did any pre- to post-policy changes in FV availpendix A (available online at ability and selection differ by vendor type or by neighborshows demographics of WIC vendor neighborhoods. hood characteristics? Overall, availability of fıve FV types improved pre- to post-policy (p⬍0.05; Table 3). Selection of two FV types Methods improved pre- to post-policy (p⬍0.05). By vendor type, The study used a quasi-experimental, one-group design with two availability of commonly consumed fresh FV and Afriobservations prior to the policy change in 2008 and 2009 and one can-American fresh FV improved pre- to post-policy at post-change observation in 2010 (Figure 1).12 Beginning August 3, large vendors (p⬍0.05; Table 3). Small vendors had gains 2009, Illinois WIC vendors (outside Chicago) were required to in the availability of low-sodium canned vegetables (p⬍0.05; stock $20 worth of FV. Small vendors (those with one to four cash Appendix B, available online at www.ajpmonline. registers) and pharmacies were required to carry at least two frozen/ canned vegetable varieties and two frozen/canned fruit varieties. org). A different pattern emerged for availability of AfriLarge vendors (fıve or more registers) were required to carry at least can-American fresh FV and frozen vegetables (p⬍0.05): two fresh vegetable varieties, two fresh vegetable varieties, two Availability declined pre-policy and then stabilized postfrozen/canned fruit varieties, and two frozen/canned vegetable policy (Appendix C, available online at www.ajpmonline. varieties. org). Availability of African-American fresh FV inThe sampling frame was WIC vendor lists from the Illinois Department of Human Services in 2008, 2009, and 2010. The creased at pharmacies (p⬍0.05).

Zenk et al / Am J Prev Med 2012;43(4):423– 428 a

Table 1. WIC vendors in northern Illinois by vendor type and year,b n (%)

2008 (baseline)c Total WIC vendors


Pharmacies Small vendors g

Large vendors

Table 2. Descriptive statistics for FV availability and selection by vendor type at baseline (2008) Selectionc

2009 2010 (immediately pre-policy)d (post-policy)e,f 346


Availability, n (%)

126 (36.4)

145 (39.8)

52 (15.8)

55 (15.9)

55 (15.1)

159 (48.3)

165 (47.7)

164 (45.1)


M (SD)

Commonly consumed fresh FV (n⫽21)


118 (35.9)

425 a

Large vendors

143 (89.9)


93.3 (0.2)

Small vendors

34 (65.4)


54.3 (0.4)

2 (0.3)


0.01 (0.0)


65.7 (0.2)


African-American fresh FV (n⫽11)


Counties included DeKalb, DuPage, Kane, Lee, Ogle, Winnebago, and west suburban Cook, purposefully sampled based on contiguity, urbanicity, income, and racial/ethnic composition, as well as proximity to the research team. Although Chicago is located in northern Illinois (Cook County), it was not included because the majority of WIC participants in Chicago are served by WIC food centers rather than authorized retailers. b Three vendors refused data collection in 2008 and one vendor refused data collection in 2009 and 2010. c Fourteen vendors were not observed in 2008 because of data collection errors. d In 2009, a total of 30 vendors joined (9.4% increase) and 13 dropped out (four closed, nine no longer authorized; 4.0% decrease) of the WIC program, for a net increase over 2008 of 17 vendors (5.2% increase). e In 2010, a total of 38 vendors joined (11.0% increase) and 20 dropped out (seven closed, 13 no longer authorized; 5.8% decrease) of the WIC program, for a net increase over 2009 of 18 vendors (5.2% increase). f There were no differences over time in the number of vendors joining or dropping out of the WIC program by vendor type. g Large vendors included chain supermarkets, independent grocery stores, supercenters (e.g., Meijer, Wal-Mart Supercenter), and other mass merchandisers with limited food products (e.g., Kmart, Target, and Wal-Mart). WIC, U.S. Department of Agriculture’s Special Supplemental Nutrition Program for Women, Infants, and Children

Large vendors

143 (89.9)

Small vendors

26 (50.0)


9.7 (0.1)

0 (0.0)


0.0 (0.0)

Pharmacies Latino fresh FV (n⫽14) Large vendors

143 (89.9)


61.0 (0.2)

Small vendors

28 (53.8)


36.3 (0.4)

0 (0.0)


0.0 (0.0)

151 (95.0)


83.6 (0.1)

44 (84.6)


59.0 (0.2)

113 (95.8)


32.4 (0.1)

Pharmacies Canned vegetables (n⫽11) Large vendors Small vendors Pharmacies

Canned vegetables, low-sodium (n⫽10) Large vendors

145 (91.2)


53.6 (0.3)

Small vendors

14 (26.9)


7.8 (0.1)


11 (9.3)


2.3 (0.1)

Large vendors

140 (88.1)


71.4 (0.2)

Small vendors

11 (21.2)


8.9 (0.2)

8 (6.8)


2.1 (0.1)

Canned fruit (n⫽5)

By vendor type, selection of commonly consumed fresh FV and Latino fresh FV improved pre- to postpolicy (p⬍0.05). Among small vendors, selection of commonly consumed fresh FV declined pre-policy but stabilized post-policy (p⬍0.01), whereas selection of lowsodium canned vegetables was similar pre-policy and increased post-policy (p⬍0.05). Selection of canned vegetables improved at pharmacies post-policy (p⬍0.05). No interactions were found between neighborhood characteristics and pre- to post-policy change in FV availability or selection (not shown).

Pharmacies Frozen vegetables (n⫽10) Large vendors

144 (90.6)


82.3 (0.2)

Small vendors

28 (53.8)


29.1 (0.3)

7 (5.9)


1.0 (0.1)


62.4 (0.3)

Pharmacies Frozen fruits (n⫽8) Large vendors

135 (84.9)

Small vendors

11 (21.2)


9.7 (0.2)

2 (1.7)


0.0 (0.0)

Pharmacies a

Discussion Results show small improvements in FV availability and selection at WIC vendors pre- to post-policy change. Findings suggest that (1) large vendors that previously did not offer fresh FV (e.g., mass merchandise stores) added fresh FV; (2) WIC vendors expanded stocking of culturally specifıc FV to attract such customers; and October 2012

All assessed fruits had no added sugar; all assessed vegetables had no added fat. b n⫽number of vendors that carried at least one variety. c Expressed as percentages FV, fruits and vegetables

(3) small vendors and pharmacies met WIC stocking requirements by stocking canned/frozen forms of vegetables. A study limitation is that unobserved market


Table 3. Pre- to post-policy changes in FV availability and selection overall and by vendor type,a-c OR (95% CI) Fresh Commonly consumed FV

African-American FV



Latino FV


Low-sodium vegetables

2.53** (1.31, 5.35)

1.72 (0.84, 3.98)


2.69* (1.17, 6.22)

1.84 (0.91, 3.72)

1.97* (1.05, 3.70)

2.15* (1.06, 4.37)

2.27* (1.31, 5.48)

1.69 (0.94, 5.54)

1.62 (0.81, 3.25)

0.93 (0.25, 3.48)

1.01 (0.41, 2.48)

1.43 (0.91, 2.25)

2.10 (0.86, 5.12)

1.83 (0.65, 5.17)

1.18 (0.47, 2.94)

2.11 (0.95, 4.69)

2.80 (1.13, 6.93)*

1.93 (0.68, 5.53)

1.38 (1.02, 1.88)*

1.25 (0.92, 1.69)


0.71 (0.12, 4.18)

1.06 (0.04, 25.53)

1.34 (0.34, 5.24)

2.24 (0.19, 25.74)

1.67** (1.14, 2.47)

1.14 (1.01, 1.42)

1.17 (1.02, 1.33)

1.22** (1.07, 1.40)

1.13 (0.98, 1.30)

0.96 (0.77, 1.20)

1.09 (0.82, 1.46)

0.92 (0.69, 1.21)


1.67* (1.03, 2.69)

1.13 (1.01, 1.43)

1.22* (1.06, 1.36)

0.84 (0.68, 1.04)

1.05 (0.91, 1.20)

0.88 (0.71, 1.09)

1.02 (0.74, 1.40)

0.93 (0.69, 1.25)


1.71* (1.06, 2.76)

1.17 (0.78, 2.19)

1.05 (0.73, 1.58)

1.32 (0.95, 1.85)

2.01* (1.03, 3.84)

1.05 (0.53, 2.07)

1.34 (0.79, 2.29)

0.80 (0.33, 1.93)


1.04 (0.93, 1.20)

1.09 (0.95, 1.21)

1.58** (1.31, 1.91)

1.17 (0.18, 7.45)

1.35 (0.06, 30.18)

0.81 (0.32, 2.08)




AVAILABILITY 2.14** (1.31, 3.50)

Change by vendor type Large

3.56* (1.22, 10.34)


1.07 (0.51, 2.24)



2.64 (1.09, 6.38)*

5.95** (1.74, 20.29)

SELECTION Overall change Change by vendor type



Note: Bold indicates significance. a Estimates adjusted for seasonality and neighborhood characteristics: population density, median household income, and racial/ethnic composition (percentage non-Hispanic white, non-Hispanic African-American, Hispanic, and Other residents). b Robust SEs clustered on vendor, but not neighborhood, were used. There was an average of 1.4 vendors per neighborhood. Binary logit results were consistent when using two-way clustering (vendor and neighborhood). A two-way clustering option is currently unavailable in statistical software for fractional logit regression. Nonetheless, fractional logit results clustered on neighborhood were consistent with those presented here. c Based on follow-up tests conducted on regression coefficients: (2010 –2009) – (2009 –2008). An OR ⬎1.0 for this contrast indicates that the post-policy change from 2009 to 2010 was greater than the pre-policy change from 2008 to 2009. d Could not be estimated because of lack of variability in the outcome by year *p⬍0.05; **p⬍0.01 FV, fruits and vegetables

Zenk et al / Am J Prev Med 2012;43(4):423– 428

Overall change

Zenk et al / Am J Prev Med 2012;43(4):423– 428

changes or store initiatives between 2009 and 2010 are alternative explanations to the attributed WIC policy effect, although the authors are not aware of any such changes/initiatives. Nonetheless, a few studies—with various designs and in various states—found even larger positive changes in FV availability and selection post-WIC policy. Without controlling for pre-policy trends, a study of WIC vendors found signifıcant post-policy increases in fresh fruit availability in all three states examined: fresh vegetables (New Hampshire); canned and frozen fruit (Wisconsin); and frozen fruit (Pennsylvania).18 In a study of WIC and non-WIC retailers in fıve Connecticut towns, FV selection at WIC vendors increased by 23.8% in lower- but not higher-income neighborhoods, but it was similar preand post-policy among non-WIC retailers.19,20 Lack of consensus regarding differential change by neighborhood income suggests the importance of future studies with suffıciently large samples to stratify by income. A third study of WIC- and non-WIC retailers in two lowincome, minority Philadelphia neighborhoods found greater improvements post-policy in healthy food availability at WIC vendors, but did not report results for FV specifıcally.21 Based on research that U.S. rural, low-income, and minority neighborhoods have poor access to FV5–11 and that neighborhood food availability is associated with diet and weight outcomes,22–26 it is important to identify policy approaches that increase FV access. Larger subsidies to create more demand and increased retail-stocking requirements (which currently differ across the U.S. and should be tested) may yield larger improvements in FV availability and selection.27 Also, complementary policies, particularly for small vendors, may be needed for equipment (e.g., refrigeration) and linkage with suppliers.18 Further, given that the current study did not fınd larger increases in rural, low-income, or minority neighborhoods, targeted approaches may be needed to reduce disparities in FV availability and achieve environmental equity. The authors thank the Illinois Department of Human Services, Bureau of Family Nutrition, including Penny Roth, for their contributions to this study. The authors also thank Sumithra Murthy, Susan Hobson, Ganga Vijayasiri, PhD, Paige Ricca, and Chang Park, PhD for excellent assistance with data collection or analysis. The authors appreciate support from the U.S. Department of Agriculture Economic Research Service RIDGE program, Robert Wood Johnson Foundation Health Eating Research Program (No. 65852), and Robert Wood Johnson Foundation (No. 64702). October 2012


No fınancial disclosures were reported by the authors of this paper.

References 1. Braveman PA, Cubbin C, Egerter S, Williams DR, Pamuk E. Socioeconomic disparities in health in the U.S.: what the patterns tell us. Am J Public Health 2010;100(S1):S186 –S196. 2. IOM of the National Academies. WIC food packages: time for a change. Washington DC: National Academies Press, 2006. 3. U.S. Department of Agriculture Food and Nutrition Service. Special Supplemental Nutrition Program for Women, Infants, and Children (WIC): revisions in the WIC food packages; interim rule. Fed Regist 2008;72(234):68966 –9032. 4. U.S. Department of Agriculture Food and Nutrition Service. Special supplemental nutrition program for women, infants, and children (WIC): revisions in the WIC food packages; delay of implementation date. Fed Regist 2008;73(52):14153. 5. Sharkey JR, Horel S, Dean WR. Neighborhood deprivation, vehicle ownership, and potential spatial access to a variety of fruits and vegetables in a large rural area in Texas. Int J Health Geogr 2010;9(1):26. 6. Liese AD, Weis KE, Pluto D, Smith E, Lawson A. Food store types, availability, and cost of foods in a rural environment. J Am Diet Assoc 2007;107:1916 –23. 7. Zenk SN, Schulz AJ, Israel BA, James SA, Bao S, Wilson ML. Fruit and vegetable access differs by community racial composition and socioeconomic position in Detroit, Michigan. Ethnic Dis 2006;16(1): 275– 80. 8. Baker EA, Schootman M, Barnidge E, Kelly C. The role of race and poverty in access to foods that enable individuals to adhere to dietary guidelines. Prev Chronic Dis 2006;3(3):A76. 9. Larson NI, Story MT, Nelson MC. Neighborhood environments: disparities in access to healthy foods in the U.S. Am J Prev Med 2009;36(1):74 – 81. 10. Morland K, Filomena S. Disparities in the availability of fruits and vegetables between racially segregated urban neighbourhoods. Public Health Nutr 2007;10(12):1481–9. 11. Block D, Kouba J. A comparison of the availability and affordability of a market basket in two communities in the Chicago area. Public Health Nutr 2007;9(7):837– 45. 12. Shadish WR, Cook TD, Campbell DT. Experimental and quasiexperimental designs for generalized causal inference. Boston MA: Houghton Mifflin, 2002. 13. Zenk SN, Schulz AJ, Izumi B, Sand SL, Lockett M, Odoms-Young AM. Development, evolution, and implementation of the food environment audit for diverse neighborhoods. In: Israel BA, Eng E, Schulz AJ, Parker E, eds. Methods for conducting community-based participatory research for health. 2nd ed. San Francisco CA: Jossey-Bass, In press. 14. Zenk SN, Grigsby-Toussaint DS, Curry SJ, Berbaum M, Schneider L. Short-term temporal stability in observed retail food characteristics. J Nutr Educ Behav 2010;42:26 –32. 15. Grigsby-Touissant D, Zenk SN, Odoms-Young A, Ruggerio L. Availability of commonly consumed and culturally specifıc fruits and vegetables in African-American and Latino neighborhoods. J Am Diet Assoc 2010;110(5):746 –52. 16. American Community Survey Offıce. The 2005–2009 ACS 5-year summary fıle technical documentation. 2011. 17. Papke LE, Wooldridge JM. Econometric methods for fractional response variables with an application to 401 (k) plan participation rates. J Appl Econ 1993;11:619 –32. 18. Gleason S, Morgan R, Bell L, Pooler J. Impact of the revised WIC food package on small WIC vendors: insight from a four-state evaluation. 2011.


Zenk et al / Am J Prev Med 2012;43(4):423– 428

19. Andreyeva T, Luedicke J, Middleton AE, Long MW, Schwartz MB. Changes in access to healthy foods after implementation of the WIC food package revisions. 2011. 20. Andreyeva T, Luedicke J, Middleton AE, Long MW, Schwartz MB. Positive influence of the revised Special Supplemental Nutrition Program for Women, Infants, and Children food packages on access to healthy foods. J Acad Nutr Diet 2012;112(6):850 – 8. 21. Hillier A, McLaughlin J, Cannuscio CC, Chilton M, Krasny S, Karpyn A. The impact of WIC food package changes on access to healthful food in 2 low-income urban neighborhoods. J Nutr Educ Behav 2012;44(3):210 – 6. 22. Bodor JN, Rose D, Farley TA, Swalm C, Scott SK. Neighbourhood fruit and vegetable availability and consumption: the role of small food stores in an urban environment. Public Health Nutr 2008;11(4):413–20. 23. Powell LM, Auld MC, Chaloupka FJ, O’Malley PM, Johnston LD. Associations between access to food stores and adolescent body mass index. Am J Prev Med 2007;33(4S):S301–S307. 24. Franco M, Diez-Roux AV, Nettleton JA, et al. Availability of healthy foods and dietary patterns: the Multi-Ethnic Study of Atherosclerosis. Am J Clin Nutr 2009;89(3):897–904.

25. Moore LV, Diez Roux AV, Nettleton JA, Jacobs DR Jr. Associations of the local food environment with diet quality—a comparison of assessments based on surveys and geographic information systems: the multi-ethnic study of atherosclerosis. Am J Epidemiol 2008;167(8): 917–24. 26. Zenk SN, Lachance LL, Schulz AJ, Mentz G, Kannan S, Ridella W. Neighborhood retail food environment and fruit and vegetable intake in a multiethnic urban population. Am J Health Promot 2009; 23(4):255– 64. 27. Cole N, Jacobson J, Nichols-Barrer I, Fox MK. WIC food packages policy options study. 2011.

Appendix Supplementary data Supplementary data associated with this article can be found, in the online version, at

Did you know? When you become a member of the ACPM ( or APTR (, you receive a subscription to AJPM as a member benefit.