Hello everybody. My name is Earl Hirsch. I'm coming to you from the University of Washington in Seattle, specifically, I'm at the diabetes institute. Um right next to the space needle across the street from amazon and also across the street from the climate pledge arena where our new NHL hockey team is about to start their season. Um It's great to talk to you about this very important topic to me. I've become interested in the economics of diabetes care for several decades and I think that this topic of C. G. M. And the managed care setting, you will find very interesting. These are my dualities. And let's get started. Do our home monitoring tools really improve outcomes and costs. Well back in the really, really early days, It's hard to know. This is the glue urine glucose testing that was used back in the 30s, 40s and 50s even there was some use of this chemistry set in the 1960s, believe it or not. But then in the 40s and 50s we moved to urine glucose testing with the clinic test. Eventually, in the 1980s we moved to finger stick glucose testing, which was quite painful for many people. And then eventually we moved to the whole topic of uh meter glucose testing, which actually Was quite controversial even into the early 1990s because the cost effectiveness data on finger stick testing really did not meet the rigour that we have today to see if something is both effective safe but I would add to that cost effect. So here we are in this 100th anniversary since the discovery of insulin. And do we have data to suggest that our older tools had an impact on acute complications and costs. And the reality is we don't really have the data especially when we go back and look at the urine glucose testing. We can argue we sort of have this data for the diabetes control and complications trial. But that was a type one study and we really weren't thinking about the cost and the cost analysis the way we are thinking about it now. So what about our newer tools because now um in the American Diabetes Association standards of care especially in the world of type one diabetes and in type two patients who take insulin. Now our standard of care is continuous glucose monitoring or C. G. M. Here in the United States right now we are fortunate to have four different systems available. But I think the real question the first real question about C. G. M. Is will it reduce visits to the emergency department? And Do we have any data on this? And as it turns out and you will see all of these data are relatively new. This came from 2020. Looking at the Association of Cumulative multi morbidity glucose control and medication use with hypoglycemia related E. D. Visits and hospitalization among adults with diabetes. And this was a report looking at over 200,000 adults with diabetes, you can see that they were more elderly. I have to be very careful when I say elderly. Since I'm getting close to that very quickly. Their overall glucose control was actually pretty good especially for this age group. That's sort of where we want them. We don't want them to low. The majority of these people as you can expect had type two diabetes. And 69% of them were either on insulin with or without a soul final Yuria. And this is looking at this survey between The beginning of 2014 to the end of 2016. What was found, Well looking at the crude rates of hypoglycemia related Emergency Department visits and hospitalizations. And if we look at events per 1000 patient years, what you see as you see it's very steady until the end of the commercial age group. Commercial insurance age group. But then once we start Medicare age at the age of 65 we see this increase Of these crude rates of hypoglycemia and for the over 75 year olds it gets higher yet. And perhaps not surprising when looking at annual household income. The highest rates are with the lowest income and the lowest rates are with the highest income. I think we see that pretty much with everything We then we'll get rates stratified by total comorbidities. Not surprising to me, the more comorbidities one has the higher the risk of hypoglycemia. Into the E. D. And into the hospital. And this one I find particularly interesting because we see this over and over again we see this U. Shaped curve Where as the Hemoglobin A. one c. increases. We see these higher rates of hypoglycemia rates. It then has a nadir Between 5.7 and 6.4. It then goes up a little bit When these a one CS get less than six and 5.6 and think there are many reasons for this glucose control is too tight. Maybe the A. One CS are reading too low because of advanced ckD or chronic kidney disease. But we see this quite a bit where it's actually these higher levels Um of a. one c. That have the higher rates. And I can only speculate why we see this At least in the world of Type one diabetes. And I think this is probably too true in the type two world is that somebody does an occasional glucose. They see it's in the 3000's they give a dose of insulin. They don't recheck and they end up in the E. D. Or even worse. That on my view is a total speculation. This is also not surprising somebody who doesn't make any insulin. Type one diabetes a 34% increased risk in hypoglycemia related E. D. Visits or hospitalizations. And I think any emergency room physician will tell you this is what they see. Especially the older type ones who um who mostly have minimal hypoglycemia awareness. This is a complicated slide the annual cost of health care resource use from hypoglycemia. And we're looking at resource use with no hypoglycemia in the blue, non severe hypoglycemia in the orange and in the gray we're looking at severe hypoglycemia. And what I want you to pay attention to is that when we look at the total cost of severe hypoglycemia, It's over $15,000. And this is in $2013. But this is obviously an extremely expensive situation. We talked about hypoglycemia. What about keto acidosis and hyperglycemia emergencies here in the United States and abroad. Well first in the U. S. Now these data came out in 2020 And we only have data to 2015. So this does not include The COVID-19 epidemic. But note this increase of e visits, an increase in inpatient ketoacidosis From 2009 to 2015. It will be very interesting to see what happens once we get through 2020 and 2021. With these curves we know they haven't gone down. But this is what we had published as of last year. If we now look at hospitalization in the U. S. Population by age group. And I really want to focus on hyperglycemia hyperglycemia hospitalizations. What you see in this youngest age group as you see this bump up In hyper Glycemic Hospitalizations. But the numbers also are going up in the 45-64 year olds. They're not going up for hyperglycemia in the elderly. Um age group are the oldest age group. And actually between 2009 and 2015 hyperglycemia crisis increased by 81%. In the United States 81%. Whereas we would think things would be going the other way with all of the developments and diabetes. We are actually going the wrong direction. And in fact if we look at length of stay with ketoacidosis now we're just looking at keto acidosis. The length of stay and the blue went down where as the cost actually increased in the orange. In fact the cost is $26,500 per episode. Yet the length of stay actually went down. We're getting them out sooner. But the number of people have obviously Um increased. And this is a big deal because we saw how it increased up to 2015. Um This is the problem. So I took the liberty of doing some math And I only had the data to make the conversions to 2020. But let me show you what I did in 2014 there were Almost 189 admissions for keto acidosis And $26,000. $526,556 per admission. So that was we spent $5 billion $5 billion dollars on the k. And there is this thing called tom's medical inflation calculator and I went and I did the math that actually increased the amount we spend on deka by almost a billion dollars. I could round that up to $6 billion. I'm not an accountant. I'm not a CPA, I'm not a mathematician But we'll call it $6 billion. Okay So the number of cases of zika in the U. S. Cost Well in excess of $6 billion dollars In 2020. Remember this is 2014 data And we know the data went up to 2015. I actually think it went sky high during covid. Um And with more you glycemic DK from the SdlP two inhibitors. So this is a very conservative number were well in excess of that now so can see GM reduce the cost of decay. That to me is really the question. This was a UK study not from the US. And they looked at almost 2500 people with type one diabetes from 30 hospitals in the UK. The mean age was 34 more women than men Diabetes duration of 14 years. There were non obese and the hospital admission for hyperglycemia and deka with using a C. G. M. A freestyle. We break it went down from 5% To 1% in six months. That's an 80% reduction of ketoacidosis. Let me repeat 80 by using the freestyle libre A C. G. M. In the UK. Okay well that's one study. Let's go to another UK analysis again using the freestyle in the break. This was a prospective observational study with 900 people using the liberate all type ones A little older than the last group, a little longer duration than the last group. But the DK was reduced from 10 to 2 episodes in the six months following the initiation of the library. And this is where the science to me gets interesting. It was exactly an 80% reduction. We had two studies showing the exact same thing. This is real world observations using the liberate continuous glucose monitor. And I find this probably at this point not coincidental. Okay, so let's take it to the U. S. You remember what The cost of Deka is in the us. It's well in excess of $6 billion. So what happened if we would reduce the cost in the us by 80%? Well, I did the mass. This was the number, it was actually close to six billion. And you multiply that by .2. So the number is 1.2 billion. Therefore the cost savings. The 80 Conuction in costs is 4.78 billion dollars. Now, I can't tell you what a billion dollars is. That's not how I Those are the numbers I used to balance my checkbook but suffice it to say this is a heck of a lot of money by investing in the freestyle. We break in those two trials. Okay. Similar reductions in acute diabetes complications in France. This is a true real world evidence study. It's called the relief study. So we're going to a different country um, in europe, looking at 74,000 people with either Type one or Type two diabetes, initiating the freestyle de bray in the french national claims database. So this was a database claim survey. Well, look at type one diabetes first and it's interesting if we look at deka, The reduction was 56% over 56 hypoglycemia in this study didn't do anything. But if we look at the overall reduction Of acute events, there was a 49% reduction. And in type two diabetes, we also saw this reduction, not, not as much of reduction as far as the actual numbers, but percentage wise, it actually wasn't far off. There was a reduction in hypoglycemia with the tattoos, which is very interesting to me and we saw this overall reduction Of acute events of 39% in this type two population. This is from France. What about reductions of hospitalizations? 12 months after starting the library. Well, this study looked at, what were they doing with finger stick testing before not testing at all. Testing 123 times a day, 4 to 5 times a day or more than five times per day. And what's interesting is you saw essentially the same amount of reduction in hospitalizations. 12 months after starting liberate in red with people not testing at all As you saw in people testing over five times per day. It was 54% compared to 51.2%. And and we've learned this in this country also that the frequency of testing is probably not a good way to look at outcomes when looking at C. G. M. Data such as this. Okay, so that was a 54% reduction. Let's translate that To the US cost. Remember in the UK they found an 80% reduction. Well we're using a different number. But what we're still seeing is a $3.23 billion dollar reduction in costs. I don't know of any health care economists or for that matter any politician that would not like to see that more data can see Gm improve outcomes and cost more real world evidence in the U. S. Now this was a study that Richburg install and I were involved with where we defined acute diabetes events as a combination of inpatient and emergency outpatient events. This would include hyperglycemia, hypoglycemia, deka hypoglycemic coma, hyper osmolarity and secondary. We looked at all cause inpatient hospitalization. Not just related to the diabetes events but everybody going into the hospital. We looked at over 2400 people. They all have type two diabetes. They were all receiving multiple injections, a basil insulin with the Crandall insulin before and six months after starting the freestyle re break. This was just published earlier this year. The age was 54 years. And what did we find was quite interesting when we will get a cute diabetes events. And remember this is all the glycemic events you could think of. This was the line before acquisition. The red is the line after acquisition. The hazard ratio is .39. In other words there was a risk reduction of 61% very significant wearing in this case the liberate compared to not those were the acute diabetes events. And now if we look at all cause hospitalizations, we also see a similar type of result. The hazard ratio was not quite as quite as high but in my view it was still very impressive with a 32% risk reduction. And you know when I first saw these data for the first time after the statisticians um went through I just like now I I got goose bumps um simply because we don't usually see in medicine and science. We don't see data like this now. Yes, this is retrospective but its real world evidence and this is what has happened with the use of continuous glucose monitoring. We concluded that these findings provide support for the use of Siegmund Type two diabetes patients treated with short or rapid acting insulin therapy to improve clinical outcomes and potentially reduce costs. Okay now, as you know, there are actually four C. G. M. S on the market but there are really two types of C. G. M. That the majority of patients in the United States where now we looked at two different databases. One IBM markets can research database. And that was for reviewing the decks calm and the freestyle right? We had over 3500 type ones total. You can see um 1100. Almost 1200 with dex com more than that on freestyle liberate. And we had Almost 4000 Type 2s. You can see also there were morley braised than decks calms. But all of these patients were using multiple injections. They were all adults for both the type one and type twos. The library and the decks com users were propensity score matching and I have become very um knowledgeable about propensity score matching matching to normalize the group. So you can do comparisons looking at demographics, provider visits comorbidities, insulin pump use and baseline visits. And here's sort of the bottom line of what we found after the propensity score matching. We set, we found that when we looked at all cause hospitalization and acute diabetes events. First in type one, we didn't see any difference between the liberate and the decks calm. At least in this analysis. They were the same in this real world data For all costs, hospitalizations for type one and also for acute diabetes events. There there were no differences after the propensity score matching. If we look at um acute hyperglycemia events and acute hypoglycemic events. When we separate them. Again, no difference between the library and blue and the decks calm in pink and no difference with the hypoglycemic events. That may be surprising to many people. It was surprising to me but this is what the data showed. And we also when we move now to type two diabetes, when we look at all costs hospitalizations we see wired wider confidence intervals but there were no differences with lee break compared to dex calm with all cause hospitalizations or acute diabetes events. The peas are non significant for all of these and then looking at acute hyper glycemic events and acute hypoglycemic events in our Type two population. Again, just like we saw with the type ones, there were no differences at all. So with these real world experience conclusions we can conclude that patients with either type one or type two diabetes experience similar reductions of acute diabetes events and all cause hospitalizations when using either brand of continuous glucose monitoring. My last point is penetration of C. G. M will only increase and I think it's become clear especially as we see the research ongoing. The guidelines showing how effective this is for patient safety as this becomes easier both for the patient and the provider. We know that the C. G. M. S are going to get smaller whether it's liberating or whether it is dex calm. These liberals are going to get smaller and easier for the patient to use and in fact what I foresee seen in the next few years is the sensors on the back of a watch, name your brand and they can actually measure through a radio frequency measure the glucose without any penetration of the skin. And that is what I hope we see and I think we will see within the next five years. So I would like to conclude that acute diabetes emergencies are increasing in the us both deka overall hypoglycemic emergencies and hypoglycemic emergencies. The data continues to accumulate from around the world that C. G. M. Reduces the incidence of these life threatening events. More formal cost effectiveness. Studies assessing the impact of this technology in different countries are now needed. I think we can all agree on that and I would like to thank you very much. This is the main campus of our medical school. This is the main university of Washington campus. This is a husky stadium, one of the best places to watch college football game, especially when our team is winning, which unfortunately is not this year. Thank you all for your attention
Related Presenters
