Hospitals and health systems across the U.S. are all feeling the impact of the growing problem of chronic disease, and diabetes is one of the most pervasive and costly chronic diseases. An estimated 30.3 million people in the U.S. have diabetes, or one in 10 adults, and at a cost to healthcare of $245 billion per year, according to the Centers for Disease Control and Prevention (CDC).
The CDC estimates that, if current trends continue, the number of U.S. adults with diabetes could rise to one in five, or even one in three, by 2050.
At the same time, food insecurity also is a widespread problem across many communities, and many healthcare leaders note that studies have shown a correlation between low food security and poor diabetes self-management. In central Pennsylvania, food insecurity is particularly serious in many of the communities that the Danville, Pa.-based Geisinger Health System serves, according to health system executives. While 12.7 percent of the U.S. population and 18 percent of children are food-insecure, in many of the counties Geisinger serves, those numbers are even worse: 14 percent of the overall population and 23 percent of children. One in eight of these food-insecure people has diabetes, according to Geisinger executive leaders in an article published in Harvard Business Review.
With the aim of addressing food insecurity, as a significant social factor impacting health, and to improve patients’ diabetes management, Geisinger launched an IT- and data analytics-driven Fresh Food Farmacy initiative to provide fresh, healthy food to diabetes patients, at no cost to the patients. The health system initially launched the program in July 2016 as a pilot project at Geisinger Shamokin Area Community Hospital in Coal Township, in Northumberland County, which has the second-highest rate of long-term diabetes complications in central Pennsylvania.
“Geisinger is very focused on approaching medical conditions from a population health perspective—we not only want to take care of each person that’s in front of us and provide them with the best state-of-the art care that we can, but we also look at our patients in their communities and see how we can improve the overall health of those that we serve,” says Andrea Feinberg, M.D., medical director of Health and Wellness at Geisinger Health, and the clinical champion of the Fresh Food Farmacy.
Andrea Feinberg, M.D.
While Geisinger healthcare leaders are using an old approach, essentially “food as medicine,” to tackle medical conditions, the Fresh Food Farmacy initiative is an informatics-driven project that relies heavily on data integration, analytics and mobile technology to do everything from tracking clinical outcomes to managing the food supply chain.
Jonathan Slotkin, M.D., director of spinal surgery in the Geisinger Health System Neurosciences Institute, also serves as medical director of Geisinger in Motion, a department focusing on digital engagement and patient- and provider-facing mobile device technologies, within the division of informatics at the health system. Slotkin helps lead the informatics work that underlies the Fresh Food Farmacy project.
“For me, it’s fascinating from an informatics standpoint because it brings together traditional clinical care, which we’ve all gotten pretty good at, but it also brings in the harder issues around data and the transactional level of social determinants of health, costing, supply chain and distribution, things that most medical systems are not yet optimized for. I think as we all endeavor to manage and help our populations with social determinants of health, we are going to be faced with these challenges more and more,” he says.
An IT-Driven Effort to Provide Healthy Food
The Fresh Food Farmacy pilot began two years ago with just six patients, all diagnosed with Type 2 diabetes. To start the program, Feinberg says the project team first queried the health system’s electronic health record (EHR) to identify adult patients in selected zip codes who had a diagnosis of type 2 diabetes and hemoglobin A1C levels over 8.0, indicating that their disease was not controlled. Well-controlled diabetic patients typically have A1C levels under 7.0, Feinberg says. The program also targeted patients who received primary care through Geisinger physicians to enable the team to study and track clinical outcomes.
The project team also leveraged the MyGeisinger patient portal tool, which links to the health system’s Epic EHR, to screen patients for food insecurity. Care managers also called patients to screen them for food insecurity. “Our hypothesis was, and still is, that if you have unmet social needs and if you are faced with food insecurity and you cannot afford to pay for healthy, nutritious food, then your diabetes cannot be well controlled,” Feinberg says.
Slotkin notes, “This is an interesting informatics question because it is a large population health problem and we are approaching it that way, but at the same time, it very much affects individuals and individual families. We have a nice span of human touch and human interaction combined with the use of our EHR and our patient portal to leverage the enrollment to this program.”
Early on in the project, program leaders recognized specific data governance issues that needed to be addressed to make the program more effective, Slotkin says. “In a lot of health systems, medical problems being added to the EHR problem list is something that usually only happens under physician direction. We felt that food insecurity was both pervasive enough and devastating enough that we needed to liberate it, so non-physician providers, like dieticians or others, could have the authority to add food insecurity to the problem list. Food insecurity does have a recognized diagnostic code. We took that through our senior leadership governance committee, and Andrea was successful in getting approval for non-physicians to add that to the EHR problem list.”
Jonathan Slotkin, M.D.
The program has now been expanded to 115 patients, with the goal of reaching 250 patients for the Fresh Food Farmacy in the next six months. Feinberg notes that the food provided to patients is enough to feed the patients as well as the family members in their households. “Once we get up to 250 patients, we’ll be feeding somewhere between 750 and 1,000 people per week,” she says. Geisinger has several community partners, including the Central Pennsylvania Food Bank, that provide food, mostly fresh fruits, vegetables, lean proteins and whole grains, for the program, at a reduced cost to the health system. Patients enrolled in the program are provided fresh, healthy food at no cost.
Patients who fit the screening criteria and show interest in the program are referred to an enrollment class, where they meet with their care team and receive a “prescription” for healthy, diabetes-appropriate food. A multidisciplinary team comprising a program coordinator, nurse, primary care physician, registered dietitian, pharmacist, health coach and community health assistant help to develop a nutritional counseling plan for each patient. Patients receive enough food weekly, along with recipes, to prepare healthy and nutritious meals twice a day for five days, and patients also receive more than 20 hours of diabetes education. “Our program is different than a regular food bank or food pantry in that while we believe that closing the meal gap and ending hunger is very important, we also think that taking care of diabetes is not only about the food, it’s about understanding your diseases. We expect our patients to attend our diabetes self-management class to help them to understand this very complicated condition,” Feinberg says.
Feinberg and other project leaders have seen significant improvements in clinical outcomes for patients enrolled in the Food Farmacy program, to date. With 12 months of healthy food and lifestyle changes, project leaders have seen patients’ A1C levels drop more than two points, from an average of 9.6 before the program to 7.5. “A two-point drop, or 20 percent, is very, very significant. That’s why we feel confident that we’re on the right path,” Feinberg says. “We’ve seen their A1C levels come under control, and we see very significant improvements in blood pressure, as well as total cholesterol improved.”
What’s more, utilizing payer-side claims data, project leaders drilled down into the care costs for patients in the program and found that costs dropped by two-thirds, on average, across the program.
Lessons Learned Around Behavioral Change and Analytics
There have been a number of lessons learned along the way, such as recognizing that it may take several months for patients to maintain behavioral changes. “Another lesson we’ve learned is that people who have these unmet social needs, they have a lot of unmet social needs. So, if you don’t have money to buy healthy food, it’s very possible that you also don’t have reliable housing or reliable transportation. That makes it hard for us to recruit patients and to get people engaged. Once they are in the program, they love it and we have a very low attrition rate,” Feinberg says.
Feinberg and Slotkin both agree that robust data analytics plays a critical role in the success of the Fresh Food Farmacy project. “The data analytics is huge; we have an incredible dashboard that we use and it tracks what’s going on with the patients and the program. Without that, we would not be able to support the work that we’re doing,” Feinberg says. “We all understand that if you don’t have the money to buy healthy food, then your diabetes can’t be under good control, but I need to be able to prove it, by showing the decrease in A1C levels, the drops in cholesterol, the drops in blood pressure and weight, and then follow that up by showing that it is saving money for the health system. Without that, then it’s just a nice program.”
She adds, “To other healthcare providers that are interested in this type of program, we recommend not only finding good community partners, but also make sure that you have great data analytics behind it to be successful.”
What’s more, IT leaders on the project are taking steps to improve data capture processes and are currently beta-testing a mobile app based on Apple’s CareKit platform. The app enables patient engagement, access to educational content, communication, and real-time data exchange with a provider-facing clinical dashboard on iPads. For instance, patients can wirelessly sync their Bluetooth-enabled glucometer or blood pressure cuff with the app and that data is transmitted to the provider-facing dashboards.
“What we’re seeking to do here is to connect our patients with the provider team as a method of digital engagement, but also as a method of actual clinical care and lifestyle tracking, both for the patient themselves and in communication with the care team,” Slotkin says. “What we really like about this app that it is HIPAA secure, with cloud backend syncing with multiple providers on dashboards who can drill down to the individual patient level and also look at large populations of patients at the same time, to look at things such as lab data, glucose levels and blood pressure, as well as physical activity.”
Slotkin adds, “This project has really brought out some of the absolute best in the analytics infrastructure at Geisinger. We have a range of data questions here, which include the delivery of care at the individual level and clinical outcomes, of course. Then there are the Farmacy-related questions of medication adherence or the decreased need for medication, as we’ve observed in a number of patients. And then also we have cost of care changes.”
Geisinger work closely with its health insurance company, Geisinger Health Plan, to pull in claims data and integrate it with clinical data. “A lot of the claims side data, so the reduction in per member, per month charges comes from the luxury at Geisinger of being able to get that payer-side data,” he says.
In addition, executive project leaders are looking at technology solutions to address supply chain issues related to food distribution. “Right now, we’re doing that on a homegrown system and we want to improve on that. Our future roadmap right now, we’re asking some great questions around potentially even the need for customer relationship management (CRM) solutions that can handle distribution and supply chain,” he says.
What’s more, project leaders want to incorporate more social determinants of health data from community partners. “We want to be able to communicate with a patient’s non-Geisinger social worker in the community, and a lot of that data stream is not well-meshed with current versions of electronic health records, at least in their pure form. So, whether we do this through an expanded use of conventional EHRs, or we build our own, or we work with a CRM partner, these are things that we’re actively contemplating as we speak,” he says.
“We found historically that these are often very disengaged patients; they have been disappointed, they don’t know how to take care of themselves and they don’t have resources to buy healthy food,” Feinberg says. “We’re giving them the education and the tools they need, we’re removing the obstacles, and with that, we follow the improved health and improved fiscal outcomes as well.”