Please welcome to the main stage, Rivka Shulman-Rosenbaum. Good morning, everyone, and welcome to our Diabetes Plenary Lecture. I have the honor of introducing our speaker, but I am Dr. Rivka Shulman-Rosenbaum, the Vice Chair of the Diabetes Disease State Network for ACE. This morning, we have the pleasure of hearing from Dr. Archana Sadu. She'll be speaking about diabetes technologies in 2022. Dr. Archana Sadu is the Director of the System Diabetes Program for Houston Methodist and Director of Transplant Endocrinology at Houston Methodist Hospital. She holds faculty appointments as Assistant Professor of Medicine at Weill Cornell Medical College and Texas A&M Health Science Center. Dr. Sadu obtained her medical degree at the David Geffen School of Medicine at UCLA with honors and completed her internship and residency in internal medicine, followed by fellowship in endocrinology, diabetes, and hypertension at UCLA. Dr. Sadu is one of the faculty at Houston Methodist in the Division of Endocrinology, Diabetes, and Metabolism. She has unique expertise in endocrine care of patients undergoing organ transplantation and in particular, pancreas transplantation. Dr. Sadu performs clinical research, has numerous publications, and is an invited lecturer at national and international scientific meetings in her area of expertise. She has been extremely active in ACE leadership as the section editor for the ASAP Board Review in Diabetes, and she is the Vice Chair for the planning of this annual meeting that we are all truly enjoying. It is my pleasure to introduce to the stage, Dr. Archana Sadu. Good morning, everyone. Thank you for that wonderful introduction. And also thank you to ACE and the program committee for giving me this opportunity to talk to you all today. And what a wonderful place to have our first in-person meeting after three years, San Diego. I'm not sure which scene is nicer, the picturesque San Diego outside or seeing all of our colleagues and meeting new ones. I think it's been a great combination of both, and I hope you're having a great conference so far. So I was asked to speak about diabetes technologies in 2022, and it was a little overwhelming because we have had a lot going on. So I'm going to try to be very brief and pertinent to technology that will apply to your patients and your practice as soon as you head back. But we could talk about advancements in diabetes technology all day. This is my conflicts of interest. So I just want to put it in perspective. We now have over 100 years of experience with insulin therapy, but this scenario is still all too common for our patients. Every time they take a dose of insulin, they're wondering, is this the right dose this time? Because if they take too much, they go crashing down, something that every patient on insulin fears and has deadly consequences. And if they take too little, well, there's symptoms of that as well, and another potential life-threatening complication of DKA. Now if they're lucky and they have those untoward consequences and get medical care, they'll be okay. But if they weren't so lucky, some of our patients still suffer from this, dead in bed syndrome. A hundred years later, we still have not been able to eliminate this fear. So my objectives today are to do an overview of types of diabetes technology. And again, we could spend all day here doing that, but I'm going to be pretty pertinent to what you can take home and take back to your office, I should say, for your clinical practice. And then I'm going to briefly highlight recent advancements in diabetes technology. There have been many, but ones that we can look out for in this coming year. And then finally touch on barriers and future directions in this area. Now I like to put everything in clinical perspective. Even when I'm sitting in the audience listening to a lecture, I try to apply it to some patient that I've seen recently or had seen. So I'd like to start by introducing you to two of my patients. This is Imelda. Imelda is a 49-year-old Latina woman living with type 1 diabetes since the age of eight. Unfortunately, Imelda did not have great access to care. So she developed all the diabetes complications, retinopathy with significant vision loss, nephropathy leading to end-stage renal failure, and neuropathy. And in 2002, she underwent a kidney transplant. Very common scenario for patients at this stage. She also had a pancreas transplant after the kidney in 2008. Unfortunately, she was one of the few complications that had immediate graft thrombosis. So it left her back on insulin therapy. And the kidneys also don't last forever, but she had 11 years of dialysis-free life until the kidney failed in 2013, and she was back on dialysis. And this is when I met her. Her A1C was 11.5, and she had frequent hypoglycemic episodes, and that was her major concern. She did not want to go crashing down and was high risk for doing so. A little bit about her background. She's an artist, and she lives with her mother, and she shares her artwork with me. And she described this piece to me, saying that the nails in the left eye represented her retinal detachment. And the rose petals in the right eye represent the floaters she experiences. Now because of all her medical history, she's on disability, and she has Medicare for her insurance coverage. Now contrast my other patient, Dustin. Dustin's a 31-year-old male who's living with type 1 diabetes since age 4. Now he had a different history of management of his diabetes. He started insulin pump therapy very early, since age 9, and he's been on all the reiterations since then. He has absolutely no evidence of any complications of diabetes. And Dustin is a math guy. He has a bachelor's in information systems and decision making, and a master's in analytics. And he works as a senior data scientist for one of our prominent oil and gas companies in Houston. So those of you who have managed type 1 diabetes patients with this kind of a brain, you know what I'm getting at. He wants his glycemic control perfect. And he's frustrated, despite all this technology, that he cannot predict the glucose with the dose of insulin he's taking every single time. So in that backdrop, I want to now take a look back. We really have come a long way. And I like this picture, pardon if it's a little bit blurry. But if you go back to 400 to 500 BC, we were really using ants to test to find if there was glucose in the urine of patients with diabetes. If there was glucose in there, the ants would be attracted to the patient's urine. And then we moved on from there to where we had people tasting the urine of patients with diabetes. And then it wasn't until 2,000 years after that that we had the first chemical reaction that can be used in urine of patients with diabetes to kind of test if there's a degree of glucose and to what degree of glucose that is. We still use this now, and we call it the urine dipstick. Now fast forward to blood glucose testing. So the first blood glucose glucometer was large and was really limited to the physician's office. It wasn't until we were able to make those glucometers smaller, more portable, that we offered them to patients at home. And that was a game changer. Now patients could actually go home and test their glucoses and react to it in real time. And the next big game changer, of course, was CGM. And when we first had CGM, really it was for professional use in 2004. But now we've come, 20 years later almost, to patients being able to check their glucoses minute by minute if needed. And not only that, we're now using the CGM technology to integrate into insulin dosing with automated insulin delivery systems. So what is diabetes technology? We could sit here all day and talk about many, many aspects of diabetes technology. But for the next hour, I'm going to focus on these sections. So we have technology for patients. Very commonly we use insulin pumps and CGM as tools for our patients. But we also now have smart insulin pens, and I'll talk a little bit about that. And what about lifestyle tracking devices for physical activity, for physiologic parameters? We can even include that, as well as nutritional guidance, into what we can call diabetes technology for our patients. And then automated insulin dosing calculators, a very important part for our patients to help manage day to day. And then for us, healthcare professionals and healthcare systems, we can look even at more global technology like telemedicine. This has really revolutionized how we care for our patients, especially during the pandemic. And remote patient monitoring. This is coming into the scene as something that we will be able to offer to patients and get input, as well as recommendations made in a very real-time fashion for the patient instead of the once every three or six month intervals that we are often seeing them. And where we're just embarking, in my opinion, is artificial intelligence for clinical decision making. This is going to also revolutionize making the right treatment recommendation for our patients. And we're just, again, starting in this area. So this just visually gives you a visual of all these diabetes technologies. We have glucose meters, smart pens. We have apps and advisors for our patients. And then all the IT resource and support that goes with that. Consultation tools for us when we're doing these telemedicine visits. And then all the way through artificial pancreas. So I want to take a moment to talk about Moore's Law. I'm wondering how many in the audience know about Moore's Law? Oh, wonderful. Because I didn't until very recently. And if this conference was happening in Silicon Valley, we would all know what Moore's Law is. So Moore's Law states that the number of transistors incorporated in a chip will approximately double every two years. And that's a quote for Gordon Moore, who is one of the Intel co-founders. So just to give you an example of Moore's Law, or the most remarkable example, has been the development of the cell phone. From the initial version, all the way to the smartphone that we use now. Here's another example of Moore's Law in action. Does anyone recognize this ad? Maybe our younger audience does not know what Radio Shack is, but when I was young, this was a common Sunday paper ad. So in this Radio Shack ad, you'll see, this one was from 1991. There's 15 different electronic devices that they're advertising. And in today's dollars, the retail cost of all 15 of those would be $6,449. But now, 13 of these devices can be incorporated into one, with a cost of only $554 on average for that device. And not only that, that device will fit in our pocket. And that is what we all know commonly as the smartphone. So Moore's Law is happening in other technology industry. Is it happening in diabetes technology? Well, this timeline shows, from the beginning, the discovery of insulin in 1920s to current day. And you can see the very first insulin pump was developed in, the prototype was in 1963 by Dr. Arnold Kaddish. And it was supposed to be the first portable insulin pump. Doesn't look so portable, giant backpack, and it required IVs for delivery of insulin. Now we've come a long way since that first prototype. We're all the way now to 2022 and offering tubeless insulin pumps. So it took a lot longer than a cell phone, but we are making progress along the way. But are we prime time? Nick Jonas says we are. This is a Super Bowl ad from this year. The first time ever diabetes technology ad was in that prime time scene. Can't get more prime time than Super Bowl, right? And some might even ask, we may be in a diabetes technology revolution at the moment. Well, I'll give you a few examples of why that statement may be correct. So just in the first two months of this year, the FDA has approved the first bolus dosing from a smartphone app, the first tubeless automated insulin delivery system, the first 180 day implantable CGM, and gave a CGM the first time a breakthrough designation for inpatient use. So we've made a lot of progress in just a few months of this year. And coming up very soon, and hopefully by the end of the year for sure, will be another advanced hybrid closed loop system. I'll talk to you about that. The first hybrid closed loop system that is going to minimize patient interactions and even two smaller and even more convenient CGM devices. So I want to just start with the CGM evolution here. This has been a critical step in diabetes technology. This graph shows all of the different types of CGM from their inception back in the 1990s. So this graph is representing the MARD. So MARD is the mean absolute relative difference. It is the difference between what the interstitial glucose measured by the CGM compared to a gold standard blood glucose. So it's how we decide whether a CGM device is accurate enough. And of course, the smaller the MARD, the more accurate it is. So the first versions of CGM, MARDs were around 20, 25%. But you can see that with each new development, the MARDs of all CGM now are less than 10%, and the most recent one being as low as 8.5%. So we're getting much, much more accurate with this technology. And that's important because CGM is now getting, because of its accuracy, can integrate into insulin delivery. And this represents the progression of automated insulin delivery systems over time. Initially we had sensor augmented pump therapy, where if the glucose hit a threshold, the pump would stop. Then the next step was a predicted low, and the pump would stop. And then we went on to overnight closed loop. And when that worked well, fully hybrid closed loop. And now we're on to the era of advanced hybrid closed loops, where we're not only adjusting basal insulin in the background, also giving correction boluses. We'll need some patient interaction in terms of mealtime, but we are progressing towards a fully automated pancreas. Now the key brain that connects the CGM and the insulin pump is a machine learning algorithm. And we needed to develop this area as well in order to make that artificial pancreas a reality. So there's three different types of machine learning algorithms available in our pump systems today. This is the first one, the proportional, integral, and derivative, or known as PID. So PID has been around for a while, since the 1940s. And it's actually frequently used in automotive industries, such as cruise control and self-driving cars. Now it's a very simple algorithm. I'm not going to go through the whole of it, but you can see the picture there. You have the patient's glucose, goes into this brain, and PID has this three terms, very linear in their computation, and it spits out an insulin dose for the pump to deliver. Its advantages are that it is simple, and it does have a low computational burden. But because of its simplicity, it really just relies on what the measured glucose is and what the desired glucose is, and all the other variables that our patients face that would affect that computation cannot be accounted for in PID. The next one is the model predictive control, or MPC algorithm. This is a little bit more complicated. You'll see in that controller portion, there's a little bit more circuits going on. Now I'm no expert in any of this. I learned it just to get an idea of where these systems are coming from when they're recommending doses. But you know just to give you an idea of how it works. The last one I mentioned was a linear model. This one's a little bit more sophisticated. It can actually do prediction of glucose and that's what's helping our patients stay out of hypoglycemia. So it has a little bit more complexity and is multivariable which is good. But because of its complexity it has a high computational burden as well. And it relies on complex mathematical models and glucose and insulin dynamics are quite complex. So the third one is fuzzy logic. No it's not a typo it is fuzzy logic. So this this one is described as most likely to mimic the expertise of the physician. I'm not sure what that says about us physicians if we have fuzzy logic but it is accounting for all the multiple variables that our brains are doing when we're making recommendations to patients and then translating them to a rule. So with these rules then the controller will fuzzify and defuzzify again not my area of expertise but be able to integrate many different rules together to give an output. It is much more higher cost maintenance and requires a lot more expertise. So back to Moore's Law. Are we catching up in diabetes technology. Well this timeline shows a number of clinical trials using automated insulin delivery systems from 2010 to 2021. And you can see that we really are mushrooming in this area. And because of these trials we've gone from sensor augmented pump therapy to closed loop to hybrid closed loop and we're now embarking on dual hormone closed loop. But there are a group of patients out there and we should know about this that don't feel that we are advancing fast enough or well enough. So I wanted to make you aware of this open source do it yourself hybrid closed loop system that's available on the Internet. And that group goes by the hashtag. We are not waiting. So this is a very smart group of individuals. I have a patient who has used such a system but they have figured out how to do it beyond what we have been able to offer them through FDA approval and commercialization. And it's just I just highlight this not that we should recommend it to anyone but to show that their patients are frustrated and they may be seeking alternatives to what our FDA recommends and there is no safety around this. So it's scary scary that they have to feel like they do this and they get complete you know instructions on the Internet on how to create these systems with the products they already have. So with that I'm going to move on to current state and where we at today and what to expect hopefully in the next 12 but maybe definitely by 24 months. So we're all familiar with the Dexcom CGM the G6 version. So and the pipeline soon is the G7. So it has a new platform completely. It's going to be 60 percent smaller than the existing G6. It's smart is going to be even less than the existing G6 which runs around 9 percent. This one's going to be in the low eight. So look more accurate easier to apply with a transmitter incorporated in the sensor shorter warm up time and interoperability. So this is this is another key development in diabetes technology where different parts of a system can be put together. So you don't have to go with one manufacturer for the whole thing. It is actually available in Europe and pending here at FDA approval. The other thing the other system now is Libre 2. It's really taken off especially in this pandemic time. We're all familiar with the Libre 2. This one came out in 2020 and it's unique and it's the longest where 14 days measuring interstitial glucose every minute has the alarms has a really good Mart as well. Nine point two percent. It's been authorized by the FDA to be part of integrated CGM and that's coming up in the near future. And so is the Libre 3 which is going to be even smaller as well the size of two stacked pennies. And then another great development this year is the ever since E3 being the first six month implantable CGM prior version was three months and it has a small sensor about the size of a long grain of rice and a transmitter that the sensor is actually inserted with a small office procedure under the skin and then the transmitter overlies that and then transmits the data to a mobile app. So just to conclude the CGM part it really has changed how we manage diabetes. It's changed vernacular entirely whereas previously we were focused on the A1C. We're now starting to talk about or should be talking about TIR T I R or time and range. And that's what this data this prolific data we get from all day and all night for patients. We can look at time and range time below range and time above range and have those discussions. But is time and range really going to improve our patients outcomes. It's great to have a good glucose value but what does that mean for their comorbidities and complications. Well these are just two studies the one on the left showing you that higher time and range less microalbumin area and the one on the right showing that the time the amount of time and range is directly proportional to all cause mortality and cardiovascular mortality. All right so I'm going to move on to now hybrid closed loop systems and just describe the more prominent ones particularly the ones available or soon to be available here in the US. Again a very familiar hybrid closed loop the tandem T slim with control like you. The new thing for tandem is in February of this year the FDA cleared the system to be allowed remote bolusing through a mobile app. So no. So again making it more convenient for patients less devices for patients and that's expected to launch soon. Also in the pipeline is a smaller version called the T sport. Now it was important to get that mobile app a bill applicability for boluses because now that you can make the actual pump smaller so there is no screen there that you have to use. So for our very athletic and active patients this is going to be an option as well. Now this year we had I mentioned the first tubeless A.I.D. to be to be approved and this is the Omnipod 5 approved in January. It has adaptive basal rates that go with each pod. So it remembers what the pods before did and uses it in the next pod when the patient changes it. It has more flexibility in targets and we know Imelda needed a higher target because hypoglycemia was her major risk factor and her life expectancy was not the same as Dustin who wanted it as low as he can go for as long as he can be. So this offers a little bit more there too. And it also does automated correction boluses and has smartphone capability. Now interoperability is more promising with this system. The first version is with the Dexcom CGM but in coming soon will be also with the Libre. And Medtronic this has been the longstanding manufacturer in this space. We've been using the 670 and 770G in the last couple of years but coming up soon is a 780G. This version is already available in Europe and is pending FDA approval here. So new things for this a new sensor for them a Guardian 4. It's going to now offer not just the basal but bolus corrections that the others offer two different targets. Previous systems only offered one and it's going down to age 2 and above in terms of FDA approval if granted and still has all the Bluetooth sharing remote monitoring as well. Now this is exciting. New kid on the block hopefully if approved is the beta bionics islet and it comes with a nice story. So the developer and founder of beta bionics is a Ph.D. in biomedical engineering. His wife is a pediatrician and she unfortunately diagnosed her their son at 11 months with type 1. So since that time they've been working on creating an artificial pancreas for patients with type 1. So this is going to be very different in terms of the other systems. The only thing needed to enter for initialization is a weight no carb ratios no sensitivity factors needed and no the patient does not have to put in and count carbs for it to do a meal bolus. So minimizing patient interaction making it more automated and it's going to be interoperable as well with multiple different sensors. The first pivotal trial is with insulin only and that is actually completed recruitment and the data has been locked. So it's moving along but it also has the potential of by hormonal and they've just started the recruitment for pivotal trials with DASI glucagon and insulin. So the next couple of slides I've attempted to kind of consolidate the information of the existing systems are soon to come for you and all of the differences in their features because like Imelda and Dustin not all patients will need the same thing. And now we actually have options. So just just to let you know for instance the glucose sensor is different for different systems. Some require no calibrations. Others may require one or two calibrations a day. The algorithm the brain behind them that I mentioned can vary as well. And with that brain the features will will vary age indications whether you want to go down to pediatric ages how long it takes to warm up and what we as health care providers have to enter into the system can be different and the targets what targets do the patients need and want and some other features as well here how it regulates the basal insulin how it adapts can you also control it from a smartphone and the primary one I want to focus on this slide is the pharmacy benefits. So all of these systems are costly and usually have a lot of paperwork to go through and are usually DME except for the new Omnipod that's actually has a pharmacy benefit so they can go pick it up there. So do these systems really do. Do they benefit our patients in terms of how they work for glycemic control. I'll just summarize this by saying yes these are all the pivotal trials to some of these major systems and you can see that they all have some degree of A1C reduction around half a percent. They all improve time and range and they all improve hypoglycemia or time below range. So to get to the ultimate end of the artificial pancreas though we have the sensors we have the pumps but this diagram shows we're still lacking a lot more. We're doing way better with hypoglycemia prediction way better with improving hyperglycemia but we're not accounting for other physiological variables of the patient physical activity even temperature blood pressure stress level and sleep. So still a lot more to go in terms of automating all of that. I mentioned smart insulin pens. So the patient that's appropriate for AID is pretty unique. They have to be educated they have to have resources and they have to really want to be able to manage their glycemic control even minute to minute. Not all of our patients meet that. So we now have these smart insulin pens that can help them as well. So we're integrating CGM into a smart pen which means that the pen has insulin and is tracking insulin delivery for us and the patient and they have apps that can help the patient dose for bolusing as well. This system is only applicable for rapid acting insulins but we have other systems other smart pen caps we call them that go on to the patient's usual insulin pen and we can use them for both basal as well as rapid acting and then we integrate those with CGM as well and get the full AGP report to make our dosing recommendations. So you can see on the right it's packaged all into one and with these devices we're now able to offer remote physiologic remote monitoring for our patients. They upload at home we download in the office and we can complete the loop and make real time recommendations and changes. And in the hospital space we've actually taken a leap here. As you all remember all too well the pandemic made all of us have to think how we deliver care differently and the hospital setting was was one of those especially in the ICU where COVID patients were admitted frequently in the very beginning and because of the cytokine storm of the infection the use of steroids glycemic control became a prominent problem and typically we would do manage that in the hospital and ICU with IV insulin but that increased the risk of exposure to the nurse who had to do hourly blood glucose monitoring. So we started at Houston Methodist a pilot of using CGM in these ICU COVID patients. We tested out two different CGMs actually and it was really just a feasibility study with the intent of reducing that exposure for the nurse but still being able to safely deliver insulin therapy for our patients and the pilot was unexpectedly successful. Thus far we have not had FDA clear CGM in the hospital due to concerns of accuracy. But when we looked back in our group of patients we found that the MARDS were much better than we expected and remember these are very sick patients. We allowed at least one presser use. They were on dialysis. They were on ventilators but we still showed MARDS of 11 and 13 percent. That's really good. And then another way to assess accuracy for CGM is Clark air grids. If you're familiar with that if you can stay in the A and B zones for a majority or most of the time that's pretty good and we found that both CGMs functioned well there. And we weren't the only ones. This is a list of all the CGM studies done during COVID for COVID in the hospital setting. So the FDA had allowed us to use CGM when COVID started in April 2020 and all of these studies began. They're very small but all of them are showing really good mark of 9 to 14 percent and really demonstrate the potential for glucose telemetry as we like to call it in the hospital setting. And following that just this March one of the manufacturers Dexcom was able to get a breakthrough device designation which means that we're going to speed up this process for the FDA and hopefully have inpatient CGM very soon. So I'm going to move on to barriers and there are many. I just wanted to show you this graph here from the type 1 diabetes exchange comparing how we're doing with managing patients and their A1Cs from two different type periods 2010 to 2012 and then 2016 to 2018. And you can see if we even use the universal A1C target of 7 percent we're not doing that well despite all of the developments I've shown you. So how are we doing with CGM use in this patient population. Well it's gotten better definitely from 2011 to 2018. It's a steady increase and that increase seems to be in more of our younger patients and kids and adolescents and then A1C by technology use. This is a great graph. So just to take a moment to orient you. The black bars are just injection therapy and blood glucose monitoring so no technology. The horizontal stripes there on an insulin pump with blood glucose monitoring. The white bars are there on injection therapy but with CGM and then the diagonal bars we're combining CGM as well as insulin pump therapy. And you can see here that those without technology had any technology had the highest and those with both types of technology had the lowest A1C. So any technology seems to be better than none in this patient population. Other barriers of course we can't ignore is socioeconomic status. I mentioned Imelda didn't have resources for for care whereas Dustin started on pump technology as a kid. So no surprise to us A1C is directly proportional to household income and so is the use of technology. So CGM users also were higher if they had private insurance and higher incomes and using that CGM resulted in lower A1Cs for those patients. Just to list a few other things that came to my mind it's definitely not comprehensive but in the development track we have safety accuracy I'll talk a little bit about that the FDA approval process is long and laborious hopefully we can make advancements there interoperability we're really doing well there cost effectiveness we really need a lot more in that area and what is the eligibility criteria for patients to use this technology and in clinical decision making our patients need more education on this our clinicians need more education on this which is why you're all here and then we need more advancements in AI to help us make those treatment recommendations safety so it all comes back to safety as it should but we really have some questions what are we considering the gold standard comparison for safety all of the CGM marred studies I showed you they use something called the yellow springs instrument which is a very old instrument if you've all worked with actually they're not even available anymore I don't think they're manufactured anymore so should we be using other types of comparisons venous capillary arterial blood glucose for our CGM studies and then for our pumps how low should our hypoglycemia be rate be and what should we compare that to no technology or the normal patient should we be saying we should have no hypoglycemia or hyperglycemia so just questions that we need to to answer before we can continue to move forward now barriers in health care health care professionals this is a study done by ace actually and dr. lowing and I presented these results at the cardio metabolic conference last fall and I wanted to share them with you it was a study really of a survey as an email survey sent out to 200 endocrinologists and about 200 primary care and also about 500 patients both with type 1 and type 2 just asking questions about their experience with digital health tools in general and what they experience what their bear what their thoughts and perceptions were about barriers and it came out as we would expect that cost coverage and patient adherence were the top three challenges and that was across the board for endocrinologists as well as primary care physicians and then another barrier that came out of this survey was the work that was required to implement that technology whether it be on the patient side to learn it or the provider side to incorporate it into the daily workflow of their office downloading reviewing the data etc and I mentioned cost-effectiveness I did do as much of a literature search as I could try to find these studies but there were very few and far between here's an example of one using one system the Medtronic 670 hybrid closed loop versus previous versions that were not closed loop and this is a Swedish study so may not be universally applicable applicable as their health care systems very different but they looked at a number of different things including life expectancy quality of life etc and they do something called an incremental cost-effectiveness ratio so in in that patient population it's a small study they found that their ICER was 164 thousand Swedish kronor and their willingness to pay for any intervention was 300,000 so in this study they suggest that at that willingness to pay threshold hybrid closed loop is probably cost-effective or in the last few minutes I want to touch on the future and what I think is coming up here everything I've showed you is is vast but in type 1 diabetes all the studies are really focused with technology in type 1 we're now finding technology is just as good if not even better for type 2 so this is a new space a much bigger market even force that for the implementation of technology this is one study from Kaiser Permanente that looked at patients either with type 1 or type 2 who were started on CGM versus not and they found lower a1c lower hypoglycemia and lower utilization of health care services and this other studies the mobile study looking at type 2 patients treated with basal insulin and similarly patients with with CGM had higher time and range and lower time above range as well so and they did not find a difference in hypoglycemia but we're getting better glycemic control in our type twos so we hear a lot of buzzwords now in health care everywhere artificial intelligence machine learning big data maybe this will be the Holy Grail for us progressing certainly a lot of work happening in the big data space and the intention of all of this is to have early detection of disease early prediction of complications and prevention of complications and also to offer a patient centered management approach this is one example of this machine learning and in the inpatient space to predict hypoglycemia and patients admitted to the hospital this was done at Johns Hopkins and they looked at 43 different variables that already exist in their electronic health record and found these top three variables to be the most predictive of a hypoglycemic episode and tried to validate them with external validation so we're doing this we have so much information now especially with these universal EHRs that we can start you applying artificial intelligence to manage patients again interoperability it's great to have that flexibility of different types of technology but they all work together for all of our patients needs but also we need to be a little bit more expansive in our digital tools we need to include our fitness trackers our nutrition apps and then see how we can improve insulin dosing for patients data sharing platforms here's another space that I believe we can really do better with everyone every manufacturer has their own platform and many of you there's pictures of all of these wires crisscrossing to one computer and offices office practices that download all these we really need to streamline this and make a universal platform that all of these different types of technology can can be incorporated into and finally a word on this new breed has anyone heard of them physician ears no yeah it's brand new and in fact a Methodist we collaborated with Texas A&M the College of Medicine and College of Engineering and created the N med program so I mentioned a PhD in biomedical engineering and a wife physician coming together for the islet hopefully we're going to spur much more of that collaboration so we had the first class in 2019 and I'll have the first graduating class in 2023 and there's some other programs across the country as this as well so quickly going back to our patients so Imelda finally got her to the point where she was willing and able to use an insulin pump and we chose the pump that would work for her and that she could practically get and you can see her blood glucose values here no hypoglycemia but we can still do a lot better with hyper we did bring her a1c down from 11 to 9.5 but she was very definitely afraid of bolusing and that was the challenge so finally we got her a sensor and allowed her to bolus more so when she first came and I got this download I almost fell out of my chair and she had the biggest smile on her face I'm doing it well now so got her a1c down to 7.2 now I know some of you are thinking why would you choose that sensor when she had the other pump Medicare at the time they did not approve the sensor that would go integrate with the insulin pump for automated insulin delivery that has since changed very recently so we'll get her into the hybrid closed-loop space soon Dustin though he's always been in the hybrid closed-loop still not happy and every visit we analyze and overanalyze the glucose and why it didn't go there but his a1c is great not good enough he wants to be at six he tells me every time I want to be at six so we switched him to a different system because he was not happy with that algorithm in the other system his algorithm was better and we switched him to a different system and I'm sorry I don't have a pointer but if you notice the middle of the graph there where he becomes hyperglycemic he overrode the system he didn't trust the system as well but shot up into the hypoglycemic range and then he went hypoglycemic because he overrode and bolused above and beyond and then he got hypoglycemic and what did the system do and he corrected it he quickly took carbs but the system is now like thinking oh my god he's getting too high and kept bolusing so he and the system were fighting so we're still not there right because the system doesn't know what he did how many carbs he took and is is working on prior you know prior patterns but we do get in there and there's days like this where he's completely in target and that's what he wants so are we keeping up with Moore's Law we all have drawers like this in our house I don't think our patients have drawers like this with diabetes technology yet but we certainly are easing the burden and with that thank you for your attention and look forward to your questions thank you dr. Sadu for such a fascinating and comprehensive lecture that is also very clinically relevant I'll ask anyone with questions to please come to the microphone yes we said Lila we've got Sumter South Carolina X excellent review my question is that they the CGM in ICU do you encounter any problem when the patient is getting presser agent whether your CGM would have a falsely low reading so surprisingly enough we actually had allowed one presser on patients and many of those patients were on pressers and we did not encounter any problems that that was really satisfying to see that the CGMs were still working despite alterations in perfusion for the patient's skin now two pressers or the dose of the presser all of this still needs to be worked out but we did allow one presser next question I believe the the only way to get around it we we have to the sensor catheter in intravenous or whatever in so that has been done we still and I recall studies 15 years ago trying to do intravenous glucose sensor and catheters sensors intravenously intra arterially we still don't have the products yeah thank you next question Dr. Hershkovich UMass Medical Center congratulations on this talk however I have two big concerns one is you you are presenting this from a big academic center but the reality let me tell you what the reality is we have 20 minutes per patient out of which five we lose when the patient gets in and the vitals are done so we're left with 15 minutes it's kind of hard to do all this analysis and convey the real points to the patient in 15 minutes so one suggestion that I will make coming from an academic center and having worked in an academic center is that you academic centers have to come together mass gen Methodist West Coast East Coast to find a way that pay that providers in the real practice can convey the messages and use this technology which by the way is advancing but not to the advantage of the people on the ground so this message is and patients can be really treated because in 15 minutes you have to convey all these things so you have to find an algorithm for that the second issue that I want to bring up and I hate to break the ice here but I will break it is that all these patients on state insurances do not pay for CGM analysis so what are we physicians supposed to do in the practices work for nothing because they pay nothing and then on top of that we have to work for nothing so and I will give examples BMC one of the biggest biggest providers in the Boston area for people who are indigent so you're fighting on the PA to get the CGM and then they don't pay for the CGM analysis so if you have 2,000 patients in the practice like that you're working for 2,000 patients for free thank you thank you and I don't think anybody here would deny anything you said but I we are making some progress as I mentioned in one of my patients I couldn't integrate her CGM with her pump because Medicare would not approve it that way but it's changing it's changing at a slow pace I agree and just to comment on what we're doing in academic centers versus versus you know non-academic or small practices you know things like the AGP report are exactly what you're talking about before we used to have reams of paper of the glucose data and the pump data and you had to spend 20 minutes going through that that print out but now we have something the AGP report you can look at and immediately in seconds look at their target range look at their below hypoglycemia range and say we need to do X so we're getting there and just you know that's why I kept also emphasizing artificial intelligence we're really not utilizing in medicine the advancements in AI that other technologies other industries have used to help make that decision in a split second for us and just to add on having medical assistants or other support staff in the office that can get all of that AGP report ready for you so that when you walk in the room you can right away just start looking at it and not have to figure out those platforms as well those platforms will get it all done for you about it before the patient even comes so all you're doing is clicking a button to download their their personal information so all right next question David Feinstein from Dallas I used all the technology that you discussed I'm sure many of the people in room do also my patient some of my patients who do the best and this was not recommended by myself are an extremely low carbohydrate diets I have a tribe of such a patients they run hemoglobin a1c in the fours or the low fives they have shockingly good blood sugars and they're not using some of this technology using home monitoring but it's outstanding how well they might do this is something we're going back to the 1920s before we had almost insulin what experience have you had with patients on extremely low carbohydrate diets like 10 to 15 grams per meal or have you had any experience you didn't comment on it is it bad or is it good so yes I do have experience with extremely low carb diets one clear population is patients after ruin why gastric bypass who have an assiduous doses we use CGM with them because they have to be in that extremely low carbohydrate diet and it works it works remarkably well now in patients with diabetes I don't recommend it unless they are on these advanced closed hybrid hybrid closed-loop systems so just last month and the last two months I've had two type one patients undergo gastric sleeves where they're you know changing diets overnight right first they prep with a liquid diet then post-op they're doing very small meals and these hybrid closed-loop systems one of them stayed on it the whole time no issues it actually adjusted to the glycemic patterns now when he started eating more issues but when he was in that very low restricted diet phase post-op I was deathly afraid that it wouldn't adapt quickly enough or well enough and I kept checking on him and he's like no I'm doing fine one last very quick question maybe yeah just a couple of comments I'm a doctor Baliga I have an office in Columbus Georgia and over like Alabama I was just looking you know listening to your talk like one of the best talks ever I bet you didn't sleep for the past two months there is you know two months you didn't sleep I know that but there are a couple of things I sort of empathize with the gentleman who talked about the difficulties down in the trenches but while waiting for all the big institutions to come together they can all come to my office and I'll show them how to do all these technologies I adopt everything but about 15 to 20 percent of my patients regardless of what I do they do not get control there's always a difficulty so the one of the things that I adopt adopted recently is the Bigfoot unity device I'm glad that I started using it in the last like six months I've got more than 120 patients and all of them do so well so my recommendation people when in trouble just go for the Bigfoot unity device and use all the technologies see I was the first in the country to do ever since in a small place like Columbus so we can do it in a small places you know we don't have to of course we look up to you for the inspiration but down in the trenches we can do a lot of stuff thank you thank you very optimistic comment thank you and unfortunately we're out of time thank you everyone and thank you dr. said do for a fantastic lecture

diabetes technologies

advancements

CGM devices

continuous glucose monitoring

hybrid closed-loop systems

smart insulin pens

glycemic control

hypoglycemia

barriers in healthcare

patient education