One of the clinician leaders who had been scheduled to present on Wednesday, March 11, at HIMSS20, the annual conference of the Chicago-based Healthcare Information and Management Systems Society, which had been set to be held at the Orange County Convention Center in Orlando, was Osama Hamdy, M.D., senior endocrinologist and director of the Obesity Clinical Program and the Inpatient Diabetes Program at the Joslin Diabetes Center in Boston. Dr. Hamdy had been scheduled to present on the topic, “Reinventing Diabetes Care with Telehealth.”
As the session’s description notes, “With one in three Americans projected to have diabetes by 2050 and a shortage of endocrinologists to treat them, it’s clear that new approaches to diabetes care and management are needed. Case studies demonstrate the power of telehealth to reinvent care delivery for diabetes with real-time access to health experts at any hour, live monitoring of symptoms, instant notifications of high need for intervention, and remote coaching and consultation. At Avera Health, telehealth monitoring, education, and intervention for women with gestational diabetes has saved $500 to $1,000 per patient, decreasing C-section rates 20 percent, eliminating hospital stays longer than two days for one-in-four women, and increasing vaginal deliveries without complications by nearly 27 percent. This session will explore strategies for using telehealth to more effectively manage diabetes in men, women, and children, with lessons learned from Avera Health and Joslin Diabetes Center, part of Harvard Medical School.”
Dr. Hamdy, who has been a practicing endocrinologist for 35 years and who has practiced at the Joslin Diabetes Center for 22 of those years, spoke recently with Healthcare Innovation Editor-in-Chief Mark Hagland about that presentation, days before HIMSS20 was canceled. Below are excerpts from that interview.
Could you share with us some of the background behind your upcoming presentation?
Certainly. Currently, we are facing two problems in diabetes. One is the rising cost of diabetes care and of insulin. The second problem is the capacity to see larger numbers of patients, as the numbers keep going up. The cost is around $327 billion dollars spent every year. That number is going up by 40 percent every five years’ that’s a huge amount of money. And 43 percent of it is spent on hospital-based care, to treat complications from diabetes.
The second problem is the number of patients involved. In the U.S., around 50 million people have diabetes, and the number of endocrinologists available is decreasing. Technically, we have around 6,300 endocrinologists in the United States, but only 1,000 are available to see patients. 10 percent are retiring every year, and another 4,300 are working in pharma companies or NIH or are academic researchers, so the availability of physicians to see patients is very small. So there is one endocrinologist for every 30,000 patients; that ratio simply doesn’t work. And even the most recent statistics of the AMA [American Medical Association] find that there are around 6,200 hospitals. So,, mathematically speaking, you have 1.3 endocrinologists per hospital.
Also, the classic diabetes practice is not doing a good job right now. It’s not because of training or clinical practices. On average across the entire population of diabetic patients in the U.S., A1C went from 7.8 to 8.4 in the past five years among patients with type 1 diabetes being treated by endocrinologists. The problem is that system doesn’t work well. We only see patients every three to six months, and don’t have enough capacity to see them more often.
And per the EHR [electronic health record], for established patients, we spend an average of 35 minutes per patient, but only 8 minutes are spent in direct contact with the patient. What we’ve found is that if the endocrinologist can see the patient every one to two weeks and then see them for follow-up visits every three to six months—if that person isn’t using insulin, it takes four months to get them to a1c target, but if a patient can be seen every three to six months, it takes 24 months to get them to target. The problem is not the training, the volume, but the model itself isn’t working. A lot of time is wasted on EHR documentation.
So what we’ve tried to do is to imagine what the clinic of the future will look like in the technological era, and how different types of intervention can work. So we’ve developed the Joslyn Home Model. It was developed between Joslyn Diabetes Center and AmericanWell; that partnership started in October 2018.
And the model is trying to solve many of the problems we have right now. For example, in this model, we have five new improvements—five pillars of the model.
The first pillar: we need the visit to be efficient and short. We need to keep clinic visits to a maximum of 15 minutes, with all of it being direct contact. The second pillar is around the frequency of visits. Visits should be every one to two weeks instead of every three to six months. And can be with anyone on the team: nurse practitioner, diabetes educator, dietician, behavior therapist, exercise physiologist. So anyone on the team could be seeing the patient weekly or biweekly.
The third pillar involves developing a very concise, focused interaction, with no wasted time. We send patients a questionnaire before every interview, asking them what they actually need from the visit. The fourth pillar: make scheduling and cancellation easier for both patient and physician, for flexibility. So we’re trying to cross over time and space, including via telehealth. And the fifth pillar involves making billing very, very simple. It can be a bundled payment or cash payment.
And this model works for both type 1 and type 2 patients, correct?
Yes, for everyone. In the pilot, we tested type 2 patients and those with very high a1c levels. We send a box home with every patient—a glucometer connected wirelessly to the cloud, send them a scale for weight, a blood pressure cuff, so they do all this at home, and all the data is communicated to us. And we wanted to see if this model works, and what time is consumed, and practicality and patient acceptance.
What is the system of interventions involved in the interventional program?
We review data on blood glucose on a daily basis. We can see the patient’s blood glucose; we also know if they don’t test, because the timeslot is empty. And we can see who’s at risk.
The typical model [involving sometimes-daily phone contact with patients on the part of nurse case managers] is not suitable for chronic care models like diabetes. You need continuous interaction and modification based on changes to blood glucose. For example, one of our patients was 87 years old, in and out of the hospital with pneumonia. His a1c was very high, more than 9, which is very dangerous. But it was too difficult for him to physically come here, but we were able to communicate with him, and corrected his insulin regimen, through his home phone.
The system—when we select a time, it sends a text to the patient and provider with a time. He spoke with an endocrinologist or diabetes educator or anyone on the team. In other words, we’re involved in constant case management for the patients at highest risk.
So what were the results of the pilot?
We completed the pilot after six months; in the pilot, we managed the care of 17 inpatients with very a1c; the average a1c was 9.5. Average age was 59, not young. 70 percent are on insulin. So they are complicated patients. And we try to do the model in the most difficult cases. For example, one patient has a double amputations and 20 other medical conditions—cardiovascular disease, lung disease, kidney disease, etc. And we have a patient who lives in Arizona, we’ve never physically seen him, his a1c was 14. So we’re testing the model on extreme patients. We found we were able to reduce a1c from 9.5 to 8.3, so 1.2 percent in six months.
A recent study shows with patients with 9.0 or more, that any decrease of their hemoglobin a1c, will save you 24 percent of their patient costs in first year and 17 percent in the second year. And if you translate that into money, that’s $2,500 and $1,690 in the first and second years. So you can see big savings. If you can reduce a patient’s hemoglobin a1c to lower than 7, you’re saving 72 percent of costs, which translates to $1,500 every year.
So you can imagine the impact. For example, one of our patients was a female patient who had undergone a double amputation. Her hemoglobin a1c had been at 9.4, and we reduced it to 7.1 in three months, and by the end of the pilot, she was down to 6.1. She could only come into the clinic in a wheelchair, and that made things difficult. But, with this program, she got the benefit of frequent visits. And the reality—how much time will physicians take in telehealth visits. 70 percent of the visits can be completed in less than 20 minutes. And the average visit for providers, for MDs and NPs, is 11 minutes. And educator, dietician, etc., between 15 and 18 minutes. Behavior therapists, 22 minutes. This is telemedicine. You see them on the screen and speaking with them. And they can do it from their screen or smartphone.
What have been the biggest lessons that you and your colleagues have learned overall, so far?
There are a lot of lessons. The most important is that we actually have to use technology to its utmost, to change the chronic care model, so that people don’t have to take a half day off from their job, drive a long way, park, and then be actually seen by a doctor physically only for 8 minutes. That model is obsolete. With the current technology available, things can be made much easier. For example, we have the list of all our patients. I can immediately see among those 17 people in the pilot, who didn’t test, and I can send them a text. In the future, this could be computer-generated and even AI-facilitated. But the most interesting aspect is, if I need to refer my patients to anyone, I can refer them to a dietician in California or Arizona, and can make an appointment for the patient, from anywhere. Imagine a military base.
And the physician can be anywhere physically, too, and the physician has more time for himself or herself. I cannot tell you: the technology will take over and will make our lives much easier. This is the most scalable model of chronic care possible. If this model is scalable, everyone will do better. And at Joslyn, we are very sophisticated and very much invested in education. And we could train primary care physicians on our model. So we solve the problem of physician scarcity as well. What’s more, patients love the model, they love being able to reach anyone at any time. And they won’t cancel appointments. Cancellations were no more than 25 percent, which is a very low rate. And it was mostly cancellations from providers, not from patients.