Four types of food vouchers, same dietary outcomes

Voucher restriction and frequency did not appear to affect fruit and vegetable intake.


Nutrition assistance in the U.S. can operate in several different ways, and voucher programs in particular are surrounded by ongoing policy debates. Do people eat better if their benefits are restricted to the purchase of certain healthy foods, such as fruits and vegetables, or do unrestricted benefits work just as well? Does it help long-term planning if vouchers are provided monthly, or are weekly vouchers better because they encourage more frequent shopping and a healthier diet?

A randomized controlled trial published in the April Health Affairs aimed to find answers to some of these questions. Researchers in San Francisco assigned 359 low-income adults to receive one of four types of food vouchers. Two groups received $20 vouchers weekly or monthly that could be used to purchase any type of food, while the remaining groups also received vouchers weekly or monthly but could only use them to buy fruits and vegetables. Vouchers were delivered monthly by mail for six months and were redeemable at 19 conveniently located vendors, including three grocery stores, 15 corner stores, and one farmers' market.

A study that compared restricted and unrestricted food vouchers found similar use among groups Image by Noel Hendrickson
A study that compared restricted and unrestricted food vouchers found similar use among groups. Image by Noel Hendrickson

The study's prespecified primary outcome was change in fruit and vegetable intake between month 0 and month 6, while secondary outcomes included changes in scores on the Healthy Eating Index and the Alternate Healthy Eating Index, voucher use rate, change in food insecurity, and ease of voucher use. At six months, voucher use rates were high and participants reported decreased food insecurity. However, use was similar among those with restricted and unrestricted vouchers, and fruit and vegetable intake did not change significantly in any of the four voucher groups.

Lead study author Sanjay Basu, MD, PhD, an ACP Member who is director of research and analytics at Collective Health and faculty at the Center for Primary Care at Harvard Medical School in Boston, recently spoke to ACP Internist about his study's findings.

Q: What prompted you to do this study?

A: It seemed like a number of health care organizations are trying to address social determinants of health and integrate some of these food support programs as one strategy to do so. There's an ongoing debate about how targeted or untargeted to make these kinds of food vouchers. There's one set of debates around targeted versus restriction, and the second ongoing debate results from behavioral economics, where it's been suggested that people who get benefits on a smaller interval might be able to make purchases of more perishable foods or have less long-term planning cycles, rather than getting a large amount of benefit once a month and then spending it mostly on bulk items as opposed to fresh fruits and vegetables. We wanted to test both of those theories, since to my knowledge they hadn't really been tested very much in a randomized controlled trial setting.

Q: Did your study find an effect on intake with any type of voucher?

A: The theory didn't hold in either direction. It seemed like overall just having more funding understandably improves food security and improved people's overall purchasing of fruits and vegetables, but there wasn't any real difference between the four groups.

Q: Was that what you expected to find, or were you surprised?

A: I was pretty surprised. We partly wanted to do the test because we had genuine equipoise between the various theories. But prior to the trial, there was widely held belief within the field that was really supportive of the idea of the restriction, the idea of more frequent interventions, under the premise that prior behavioral economics literature supported it. So we were surprised to find that this was not the case for our trial. It may be because the amount overall was inadequate. We also didn't test a biweekly voucher instead of just weekly or monthly, so there may be more subtle approaches to continue modifying. But overall, there wasn't really much of an effect.

Q: Participants did use the vouchers, and return rates were good. If people with restricted vouchers were using them to buy the fruits and vegetables, why do you think the consumption didn't necessarily increase?

A: Among our population, and I think in general this is true for many populations that are low income, there's just an incredible amount of negotiation that has to take place to make ends meet. That means that people are often finding every imaginable source of funding and kind of making it work by trading off different expenditures all the time, at least among my patients. They become real experts at finding a dollar here in order to make a dollar there go further. That's likely what happened is that this is yet another program amongst a mix of so many other complicated siloed programs. Perhaps we were looking too narrowly at just supporting food vouchers and … restrictions instead of looking the opposite direction at broader supports that are more flexible for people to use across domains as widely varying as rent, food, prescription drug expenditures, and so on.

Q: Participants reported that the restricted vouchers were easier to use than the unrestricted vouchers. Why do you think that was the case?

A: That was interesting. It may have been related to behavioral theory about cognitive load. It's less cognitive load not having to choose because it's yet another thing to do during the day, one less item to make decisions about. One of our participants likened it to a friend of theirs who was a Somalian refugee who arrived in the United States, went to the grocery store, and sat down in the breakfast cereal aisle because he was so overwhelmed. He didn't know what to pick. It's that kind of cognitive load that we wanted to see if we could address, but it didn't seem to manifest in a real meaningful change in diet, even if it was a noticeable change for people psychologically.

Q: What are other considerations surrounding the concept of restriction, for example, its effect on recipients' autonomy?

A: It's an ongoing discussion and debate that I think needs to be had. A lot of it is also around framing. The framing of the restriction to be on SNAP [the federal Supplemental Nutritional Assistance Program] has very much been around the word “restriction,” of whether it will discourage participation, what to restrict, and so on.

What's interesting to me is a few years earlier the analogous debate took place in the WIC program, the Women, Infants, and Children program. It wasn't framed as a restriction. It was framed as targeting taxpayer funding for nutrition dollars to nutritious foods. The WIC program is limited. You can't buy everything with it, and there's a certain group of items that are purchasable. That's actually been fairly effective from a dietary standpoint in terms of people's improved outcomes medically. In fact, some have attributed some of the decrease in childhood obesity to the WIC program.

The SNAP debate is very different. It is, in effect, a subsidy of many food manufacturers, and so there's a lot more money at play. The debate has been very centered on the idea of restriction rather around the question of whether nutrition dollars should go towards nutritious foods or not.

Q: What do you think are the main implications of your findings?

A: The main implication is to perhaps move on towards a broader set of tests around how vouchers might work, particularly looking at a broader set of social indicators that we might be able to affect with what's restricted. [We might look into providing] support systems, ones that can help stem people's challenges. Around here, rent is often a big concern. In addition to food, paying for prescription drug copayments is a major concern. [We could see] if we can offer the ability to be flexible around those items.

There's a large and understandable worry that people will misspend funding, and yet I'm surprised that some of the international studies around income supports have not supported the idea that there's large misuse. I think we need to do that test nationally and see if it really manifests the same way.

Q: Do you see any potential policy changes on the horizon because of these and other results?

A: It seems a lot of policy changes are likely to take place at the city, county, and state level. We do have a number of localities who asked us after this study to help them implement local programs. We see further interest in expanding that as municipalities on the more local level are interested in moving more quickly, regardless of federal stalemates that may be taking place.