Interview Summary

So, let’s get into it. Before we dig into the study itself, can you tell our listeners more about Produce Prescription Programs?

So, the programs that we evaluated in our study gave participants electronic cards. So, either a debit card, a gift card, or a loyalty card that they could use at retail grocery stores and some farmer market partners as well to get free fruits and vegetables of their choice. The food could be either fresh or frozen fruits and vegetables. These programs acknowledge that nutrition education alone likely is not sufficient to increase healthier eating for many low-income households, for which just the cost of healthier foods is really the primary barrier in improving their diet. These programs seek to overcome that barrier, and it’s really exciting to see that clinics across the country are turning to these. I think physicians, from our conversations with the clinics we partnered with in this study, are really excited to have a tool they can use and provide to patients to offer higher-quality care and help improve their nutrition when they’re talking with their patients about managing diet-related illnesses like diabetes, hypertension, or obesity.

Thank you, Kurt. Could you tell me a little bit about why you and the team decided to do this particular study?

Food insecurity has been consistently associated with poor health outcomes, higher healthcare costs, and lower dietary quality. Many food insecure households tend to have higher rates of diet-related chronic illnesses. We worked with a wonderful organization called Wholesome Wave. They have operated Produce Prescription Programs across the US, and they have been collecting data on these programs for about five years now. Wholesome Wave reached out to our research team at Tufts University’s Friedman School. They wanted to see if these programs are working at improving health outcomes. The key thing, I think, to contextualize where this study stands, is prior to this work, there had been a handful of studies that had shown, not surprisingly, that Produce Prescription Programs can improve dietary quality and improve food insecurity. However, very few had gone that extra step to see if Produce Prescription Programs were associated with improvements in really important clinical outcomes. Things like hemoglobin A1C, which is an important measure of average blood sugar levels in the past three months. This is critical for managing diabetes, and outcomes like blood pressure and obesity and overweight. Previous studies had found mixed results on these outcomes, and most had been very small, maybe about 50 participants. We built on this by doing the largest analysis to date. Our study had nearly 4,000 participants from 22 programs across the US, from 12 states. These ranged from cities like Los Angeles to Minneapolis, to Houston, to Miami, Idaho – so all over the US.

So, it was the fact that previous studies have had relatively small samples, and some of these other studies did not take all of the sort of important measures of health into consideration. So, you were able to build on that past work in a unique way in this study?

Exactly. I think the key thing is that Wholesome Wave had excellent relationships with their partner clinics. This meant that the clinics were willing to share medical record data with us. This is always just very challenging and many other studies weren’t able to go that extra step. We were blessed with access to a lot of medical record data and we were able to do analyses that looked at important clinical biomarkers. I will say though, our study is a step in the progression. I’ll be the first to admit, we did not have a comparison group in this analysis. So, the results that we found also could have occurred due to other reasons. Such as, for example, perhaps when someone is referred to a Produce Prescription Program, their physician might also make other referrals, or perhaps change their medications because this patient has been identified as high need. We certainly built on previous literature by having a much larger sample size and pooling data across the US, and to me, our findings really provide us with a strong rationale to continue research into this area. But also, to confirm our findings with randomized trials similar to what you would do, for example, for drug research.

That is helpful to hear. One of the things that’s really important about what I understand of this study is the fact that you worked with Wholesome Wave, and that allowed you access to a lot of different programs across the US as you described. Could you give us a little bit of a sense of what some of those programs were, and how did they provide the support that you were able to study in this project?

Most programs provided the benefits on electronic card. It’s similar to in WIC or SNAP, where participants have an EBT card, they can use it at retail settings. It’s administratively very simple. From interviews among SNAP participants and other research, this tends to be lower stigma – when you’re using a card at a checkout. These cards gave about $50 a month for six months on average for the adult programs, and $110 a month for the pediatric programs. Some of the children’s programs were also a little longer. Some of the child-focused programs that we include in the analysis were up to 10 months.

So I understand from this description that the Produce Prescription Programs also look different. There were some programs for children, some for adults. How did you manage that? There’s a lot of other things that could be going on. How do you sort of do that in this work?

Participants were referred to these programs because they were either food insecure or were recruited from a clinic that served a low income community and were very likely to be food insecure. Individuals also had a risk factor for poor cardiometabolic health. So, this means that they either had diabetes, high blood pressure, or were overweight or obese at baseline. That was really the common thread across all of these programs. We did all the analyses for adults and children separately. So, we report changes in fruit and vegetable intake, and changes in self-reported health status separately. Food insecurity – we did assess at the household level, but then for other outcomes, for example, hemoglobin A1C, we restricted that to the adult population that had diabetes when they enrolled into the program. For blood pressure, we restricted those analyses to adults with hypertension at baseline. For Body Mass Index, we restricted analyses to adults who were overweight or obese at baseline, and then did those same analyses separately for children for age and sex, only looking at children with overweight or obesity at the start of the study.

This is really important then. Thank you for that clarification. You know, you’ve talked about some of these critical measures such A1C for diabetes. You’ve talked about obesity measured in BMI. Can you tell us a little bit more about the importance of these findings and what they mean in real terms for participants’ health?

Absolutely. So, we found that participation was associated with improvements in dietary quality and food insecurity. For example, among adults, they were reporting that they were eating, on average, about 0.8 more cups per day of fruits and vegetables by the program end. And food insecurity rates were cut by about a third. So, the program seemed to be working as intended, which was excellent to see. But then looking at the clinical biomarkers, for example, hemoglobin A1C among those with diabetes dropped by 0.3 percentage points. And among those with uncontrolled diabetes, those having chronically high blood sugars that are very difficult to manage, dropped by about 0.6 percentage points. So, to put that in context, that’s about half the effects of commonly prescribed medications to manage high blood sugar levels. So, for just a simple change in diet, that is I think fairly impressive and very encouraging to see. And the effects on the reductions on blood pressure were also about half as large as we would see with commonly prescribed medications. I think it’s really important to highlight that, one, we don’t know if these changes will be sustained long-term when the program ends. There might be some participants where this program caught them in a moment of crisis perhaps, in which this helped stabilize them, and maybe they would be able to maintain these new improvements in dietary intake long-term. But I suspect many participants might not be able to maintain this healthier eating because the cost of healthier foods was the main barrier to healthier eating at the onset of the program. And so, this is an area that we’re really interested in looking at in future research. But I will say, if one were to maintain these improvements that we would see in hemoglobin A1C, blood pressure, and BMI among adults, they really were clinically meaningful. And if sustained long-term, it could reduce risks of things like heart attack or stroke years down the line. Any reduction in these biomarkers can really have a meaningful impact on patient quality of life. Things even like averting diabetes complications with damage to the circulatory system, to nerves, to the retinas in the eyes and having vision loss. So, sustained long term, I think these really are meaningful impacts on health and well-being. The last thing I’ll say, is for children, we did not see a change in BMIs for age and sex. I’m not too surprised, given it’s a relatively short program on average six months. But also, child BMI is a notoriously challenging metric to move. But I do want to highlight that among the households with children, we did see an improvement in fruit and vegetable intake, and reduction in food insecurity, and self-reported rates of higher health status. And I think if we’re thinking about childhood development, to me, that is still an important success. We know that having enough access to food in the household and having higher dietary quality is really, really important for childhood development and well-being. There’s certainly a nutrition causal pathway here. But it’s important not to forget that there’s so much stress and anxiety when someone is experiencing food insecurity, about not knowing necessarily where your next meal is going to come from. Just worrying constantly about managing household budget, about trade-offs, say between buying healthier food, paying for medications, paying for other needs. So, I suspect these programs are improving health outcomes both through a nutritional pathway, but also through like a mental health pathway, perhaps reducing anxiety for some households as well, which can also have an impact on things like blood pressure.

So, given these results, it says that there’s some important implications of these Produce Prescription Programs for the health and well-being of the participants in this study. I mean the fact that just changing fruit and vegetable consumption through a program like this had an effect similar to half the effect of some medications is really a powerful finding. I have got to ask, what are the policy implications of this work?

There is very exciting momentum across the US federal programs, promoting produce prescriptions and other forms of, for better or worse, what is known as food as medicine. And these programs all provide free healthy food to patients in partnership with the healthcare system or through a physician’s office. So, Produce Prescriptions are the focus of this study, but the other Federal programs also include things like medically tailored meals, which are healthy, home-delivered meals, often to patients with even more advanced chronic disease who also might have activity limitations or disabilities that makes it really challenging for them to shop and cook independently. Last fall there was a historic White House Conference on Hunger, Nutrition & Health, in which the Biden administration in the summer prior engaged at a national level, major stakeholders in the anti-hunger space, large health systems, researchers, and government agencies to bring together a policy agenda to address hunger in the US, and the really high rates of diet-related chronic diseases. And it’s important to note that, I think it is the first recommendation in the section of the final national strategy from this White House conference under the healthcare sector highlights Produce Prescriptions as a policy priority and expanding them in Medicare and Medicaid. So currently Medicaid, which is the federal health insurance program for individuals with low incomes or who have a disability, Medicaid is managed in partnership between each state and the federal government. And this means that states have some flexibility in how they manage their Medicaid program. They can apply for what is known as Section 1115 waivers. The federal government can approve these waivers if the state makes a good case that if they propose an innovative pilot, an innovative change, that they can make the case is likely to improve health outcomes and remain budget neutral, then the federal government can approve them to pilot this idea. So, currently in Massachusetts, we are several years into an 1115 pilot that is actually paying for produce prescriptions and medically tailored meals through the state Medicaid program. And currently about 10,000 people in Massachusetts have received some sort of nutrition support through the state Medicaid program. About 10 other states now have either similar waivers approved or pending approval to allow other Medicaid programs to do something similar. So, this is a really exciting area where expanding access to these programs is happening. And then in Medicare, which is the health insurance program for older adults, in Medicare Part C, which is the Medicare private health plans, those health plans can choose to cover Produce Prescriptions as a benefit. They’re not required to, but they have the flexibility to offer that service if they would like. And then finally, I’ll just say that there’s also new pilots that were announced in Indian Health Services and the Veterans Health Administration. All these examples show that across the federal government there are exciting pilots and expansions occurring to cover Produce Prescriptions, and other foods and medicine programs. However, they remain unavailable to the vast majority of Americans who might benefit. And so, they’re not a core component of any of those programs at the moment. And rather, these are pilots that are being tested, but very exciting movement, nonetheless.

Thank you for that really comprehensive set of examples of how policy is implicated in this work and potentially the need to expand this work. It makes me think of USDA’s Gus Schumacher Nutrition Incentive Program. That would be another way that folks could access some of the benefits of federal dollars to support produce-type prescription programs. How do you hope to build off this study in future research?

At UMass Medical School, we’re the official independent evaluator for the state Section 1115 waiver, which means we are essentially responsible for evaluating if things like the Produce Prescription Program in the Massachusetts Medicaid program is improving health outcomes. So, that is what we are working on right now. In our partnership with the state, we actually have access to all of the claims and encounters data within Medicaid. And we also are working with several health systems that are also sharing medical record data with us so we can evaluate the impacts of food as medicine programs on hemoglobin A1C, blood pressure, and BMI, so similar outcomes to this study. We actually have a large study funded by the NIH in partnership with former colleagues at Tufts University that is doing a deep dive on the Medicaid medically tailored meal program. And we’re hoping to do something similar for the Produce Prescription Program in Massachusetts Medicaid. And the nice thing about these studies, they aren’t randomized trials. Since this is a kind of a policy rollout, anyone who’s eligible for these programs can receive the benefits. But we will be improving upon this study that we just published by leveraging two really strong comparison groups, and using some statistical techniques to make sure we’re identifying patients who are as similar as possible to those who are receiving services, but ultimately didn’t enroll. So, we’re excited to have these results a couple years down the line and see if these programs are working in the context of a large state Medicaid pilot.