Seth Berkowitz, assistant professor of medicine at the University of North Carolina at Chapel Hill, and Lori Tishler, vice president of medical affairs at the Commonwealth Care Alliance, co-authored a study analyzing whether meal delivery programs could reduce healthcare costs among Medicare and Medicaid beneficiaries.

What was the point of the study?
Berkowitz: We wanted to find out if home delivery of meals would reduce the use of a handful of healthcare services and medical spending. We looked at three different groups: one received medically tailored meals customized to participants’ healthcare needs (diabetes, renal conditions, etc.); the second received healthy, non-tailored food; and the third received no food. Five days of lunches and dinners were delivered each week.

Who took part in the study?
Berkowitz: Nutritionally vulnerable patients of the Massachusetts Commonwealth Care Alliance, a nonprofit community-based health plan. The organization works with adults who are dually eligible for Medicare and Medicaid.

What were the results of the study?
Berkowitz: People taking part in the medically tailored meals program had fewer emergency room visits and inpatient hospital admissions and used less emergency transportation than those in the control group. People in the non-tailored program also had fewer ER visits and less emergency transportation. They didn’t, however, have fewer inpatient admissions.

When you subtract the program costs from the estimated healthcare savings, participants in the tailored meal program saved $220 per month. The non-tailored program reaped savings of $10 per participant per month.

You chose food delivery. Why that particular solution?
Berkowitz: The association between good nutrition and good health is clear. When people are really sick, they may not be able to go out and shop or prepare food for themselves, so food delivery can help solve that problem.

Tishler: In Massachusetts, between one in six and one in eight people are food insecure. That doesn’t necessarily mean they don’t have food, but they aren’t exactly sure where their next meal is coming from. And that will include people in the commercial market, there is no question on that. Particularly for large employers, their low-income workers may have real challenges. Look at people with children. I would be willing to bet a number of their workers’ children are getting subsidized meals at school or in the summer. Given that, it’s really easy to think about this area.

What do commercial payers need to know about “social determinants” of health?
Berkowitz: Adverse social determinants of health, that is, social circumstances that increase the chance of getting sick or that make it more difficult to manage illness, are very common in the United States even among individuals with commercial insurance.

Tishler: About 80% of things we think about that affect people’s health and illness we can’t fix with a visit to the doctor or an antibiotic. It’s much more related to the neighborhood they are living in, what kind of food they are eating, how much education they have had, what language they speak and if they are socially isolated.

There is also data in other studies showing that hospital readmissions markedly decreased when people got meals delivered to them after a hospitalization. People who are much more likely for readmission and increased cost are exactly the people who aren’t going to have a community that is able to bring them food.

It is so compelling to see what seems like a pretty simple solution and see if it can move the metric on things that are causing us to spend so much money on healthcare.

Most of the work in this space has been done with Medicare/Medicaid. Why do you think that’s the case?
Berkowitz: Medicare and Medicaid are set up specifically to care for more vulnerable members of our society, so the populations they care for often have a high prevalence of adverse social determinants.

Social determinants don’t just affect low-income individuals, correct?
Berkowitz: Social determinants are important for everyone—even individuals historically considered to be middle class are finding it harder to afford housing or nutritious food. People often expect the prevalence of food insecurity or housing instability to be lower in middle class samples than it is. It’s hard to know how common it is unless you ask.

What can commercial payers learn from what Medicare/Medicaid are doing in this space?
Berkowitz: I think it’s really important to realize that a lot of what goes into staying healthy, or into managing chronic illness, occurs outside the healthcare system. Working with other sectors, like social services, makes it much easier to help people stay healthy.

What did you learn from this study’s outcomes?
Berkowitz: I was really encouraged to see the association with lower emergency department visits and hospitalizations. These are outcomes that I think patients really care about—not having to go into the hospital.

Tishler: It surprised me that there was such a significant difference between the medically tailored meals and the non-medically tailored ones. I think that surprised all of us. We would like to say that food is food, but this article suggests that there may really be advantages to certain kinds of food. They both made a difference, but the medically tailored meals actually saved more money in monthly medical costs.

I think their average cost was $843 versus about $1,400 for their comparison group. We would all like to see more research that confirms this or begins to answer that not only that happened but why it happened.

Many of our members live on a fairly low income and in relative food deserts [areas with little or no access to healthy foods because of lack of farmers markets and grocery stores]. One of our medically tailored meals was broccoli and red peppers and salmon. And it would be harder for them to put that kind of meal together on a good day and even more if they’ve been ill or in the hospital.

What kind of barriers do you see for employers/brokers to do work in this space?
Berkowitz: I think finding the right partner agency is really important. People need to be dedicated to making nutritious and tasty culturally appropriate food.

Tishler: Barriers are really about seeing food (and other social determinants) as an important thing to improve health and reduce costs. Delivery of food might seem out of scope for a healthcare organization or health plan, but according to our data and others’, it might be a cost-effective way of preventing more expensive care—at least in certain settings.

How could they overcome challenges?
Tishler: I think there are different ways to do this. There are home-delivered meals that make sense for a population like ours. But in another setting, like after someone is hospitalized for myocardial infarction, you could send people heart-healthy meals so they know what to eat or teach them how to make healthy food. For the commercial market where people have a lot of resources, there would be a lot of ways to think about this.

It will require health plans and companies to think more broadly about what health is and what’s worth paying for. It would be a great benefit but would also benefit people’s health. Especially if the company made its choices wisely.

What other opportunities are available in the social determinant space? Are there areas other than food to target?
Tishler: Stable housing can make a huge difference in someone’s use of the healthcare system, costs and quality of life. And we also see that people live longer with stable housing. I’m not saying that every health insurance company or employer should get into the housing business, but when they think about what they want to work on in employee assistance programs, I guarantee there are homeless people in their workforce.

I was seeing patients at a homeless shelter for women, and it’s amazing. You can’t always tell who is homeless and who the staff are. Some of those women were putting themselves together and going to work. So that’s something to think about for companies.

Improving access to exercise is one thing, and that can be done relatively easily. They can subsidize gym memberships or make sure there are places to walk at work or start a walking program for the company at lunchtime.

They can help people find community resources, which are there in all different kinds of communities. And some people like education about what healthy food is. You don’t have to buy it for them, but you can do some demos at work and people can see a great way to use a zucchini.

Anything that helps decrease chronic stress—whether around social stressors or creating workplaces that are welcoming—all of those really do contribute to health. Helping people understand they are valued helps a lot.

What can employers and brokers expect regarding results when targeting social determinants?
Berkowitz: I think the most important thing is to focus on improving health. It may not be easy, as the consequences of adverse social determinants accumulate over the lifetime, but organizations that just get into this trying to make a quick ROI are unlikely to prioritize the right interventions.

What kind of measurement should be used when seeking population health improvements?
Tishler: While metrics having to do with improving cholesterol or blood pressure are impressive, the cost ramifications of those are much longer-term. Generally, I would look at measures having to do with emergency room visits, hospital costs and readmission. It’s much less expensive to feed people than to pay for three nights in the ICU.

Some payers are beginning to use social assessments along with traditional health risk assessments. Are these a good option?
Berkowitz: Incorporating social risk will likely be important, but there are a lot of unanswered questions—how to collect the data, how often, what questions to ask and what to do with it. Getting too far out ahead of the evidence could wind up leading to disappointing results.

Tishler: There are many different measures that people can use to look at social determinants of health. They can be tailored to specific populations or more general, but I definitely encourage organizations to find out more about their populations. I think there will be some obvious things they find, but they may also be surprised at the level of need.