Welcome to today's webinar, and thank you for being here. I'm Dr. Noel Barengo, and I'm the chair of the American Diabetes Association Public Health and Epidemiology Interest Group Leadership Team. And welcome. I'm Elisabetta Patorno, and I'm the immediate past chair of the Public Health and Epidemiology Interest Group Leadership Team. So Noel and I are pleased to be moderators for this important webinar today that will discuss diabetes risk scores. Welcome, everyone. The International Diabetes Federation has estimated the number of adults with diabetes is expected to rise from 425 million in 2017 to about 630 million by 2045. Type 2 diabetes mellitus does not cause specific symptoms for many years at onset, which explains why between 25 and 50% of the cases of type 2 diabetes remain undiagnosed at any time in the community. Plasma glucose, either fasting or to other levels after 75 gram glucose load, and HbA1c levels are recommended methods for type 2 diabetes diagnosis in the general population. However, these are invasive, expensive, and time-consuming procedures, and hence are not suitable for mass screening. It has been shown that the most cost-efficient methods for type 2 diabetes mellitus screening in a general population is the use of a non-invasive tool for risk stratification as the first step, followed by a blood test for glycemia. The screening for impaired glucose regulations should be targeted to individuals at high risk of type 2 diabetes mellitus. This webinar will provide a brief overview of the current experiences in developing and implementing diabetes risk scores in different parts of the world. So experience from Europe, South America, and India will be presented by our expert panel today. We are fortunate to welcome three experts to provide insight on this topic today. So thank you for coming today, Dr. Jakob Duomi-Lehto, Dr. Pablo Aschner, and Dr. Nandita Arun. Here's a glance of today's agenda. So after a few minutes of announcements, we'll first welcome Dr. Duomi-Lehto to discuss the development and population-wide implementation of diabetes risk scores and the experience in Europe. Then Dr. Aschner will discuss the development and implementation of risk scores in South America. And lastly, we'll hear from Dr. Arun about the Indian diabetes risk scores. Today's webinar will be recorded and you will receive an email with the link to the recording. We encourage you to share your experiences and bring questions by the diabetes screen tools to this event. So the speaker will take questions as a panel at the end of the event, but don't wait until the end of the session to send in your questions. Instead, please use the Q&A box in your control panel, type in your question there, and make sure to use for all your questions the Q&A box. You can also use the chat box to share experience with other members, participants of this webinar. And we'll be using the chat box to send you also links during this announcement segment. I also wanted to thank, take a moment to thank all the members of the leadership team for their work throughout the year to provide opportunities to the interest group members. So each of those you see here is a valuable member of interest group of the leadership and has plays an important role in providing the webinars and all activity we develop. So recently, our group hosted a webinar in improving access to diabetes care in underserved populations. And you received the link, see the link to the recording in that chat. We're also working to having two more webinars, one webinar, a joint webinar with American Public Health Association with the physical activity group and America's Network for Chronic Disease Surveillances in the first week of April. So look out for these events as well. Keep up to date with events for this interest group as well on the Diabetes Pro member forum. This is an ADA member exclusive forum where all members of the interest group can connect and share their experiences. Next, I'd like to highlight other ADA webinars that are scheduled for next month. So to register these, please go to the link in the chat and on the screen and we're great if you can attend. Wonderful. So I think we can now start our webinar. So it is my pleasure to introduce the first speaker, Dr., sorry, Tilueto, and sorry for that. Here we go. So Dr. Jaco Tilueto is professor emeritus of public health at the University of Helsinki, Finland. He has graduated in medicine and sociology and has a PhD in public health. His research interests include the epidemiology and prevention of non-communicable diseases such as diabetes, cardiovascular disease, cancer, and dementia. Dr. Tilueto has coordinated numerous scientific research projects. He has contributed to many landmark studies and has been involved in many epidemiological studies. His Finnish diabetes prevention study demonstrated a 58% reduction in the incidence of type 2 diabetes with a lifestyle intervention. This finding was later confirmed by similar trials from many countries. He developed a simple non-laboratory type 2 diabetes risk score, FINRISK, the Finnish diabetes risk score that has been validated and applied widely globally. Dr. Tilueto has received many prestigious scientific awards including the ADA Kelly West Award and the Harold Rifkin Award. He has published over 1,900 scientific peer-reviewed publications and is one of the most cited author worldwide in four fields, epidemiology, public health, diabetes, and cardiovascular disease with 230,000 citations and an H-index of 229. So it is without further delay that I turn it over to Dr. Tilueto today. Welcome. So thank you very much for the introduction and thanks for inviting me to join this webinar today. It really is one of my most favorite topics, how to identify people at the high risk of diabetes. So I will start with a brief introduction. Start my presentation, where did it go? Bye. I'm sorry about the delay with the slides. So I'm going to discuss with you about the development of a risk score for type 2 diabetes. First of all, it is a screening test and it's not intended to be a diagnostic test for type 2 diabetes, but it should be easier to perform than a diagnostic test. So the positive result from the screening test requires confirmation with the diagnostic information. The main questions in type 2 diabetes identification is how to identify the people who are already affected with diabetes or hyperglycemia. Or secondly, very important is to identify people who may have type 2 diabetes in the future. The defined risk, as I pronounce it, the aim was to develop a tool that was simple, inexpensive, high people at a high risk of type 2 diabetes, and that it should be able to be applied in a general population with the lay people even, and no blood drawing, no special measurements or trained persons would be required. So we developed this based on a 10-year follow-up prospective study where we had a baseline explanation in 1987 and then we validated it in both cross-sectional and other independent prospective studies. And the scores for the individual items from a multivariable model. So we first had the survey in 1987. We had cases of type 2 diabetes. We had another cohort where we had individuals for five-year follow-up and we identified cases of diabetes in five years time. So in the analysis of a multivariable model, we could see that several parameters, we actually had about 20 parameters in the beginning, but we then finally squeezed out the most important ones, which were body mass index and waist circumference, both independent, and age, glucose elevation, in a history, hypertension, and then lifestyle variables, diet, exercise. And then with the odds ratio, we could identify the significance and the parameter estimate was rounded up to the number that developed the risk score. And we could see that the higher risk score people had, the higher was the risk of diabetes. Actually, at the highest points, almost everybody became diabetic in 10 years time. And at the low score, hardly anybody under seven points became diabetic. So it was then evaluated in the so-called ROC curve, Receiver Operating Characteristics curve, which gives the area under the curve, which is a useful parameter for a screening test. The area under the curve, if it's over 0.8 or 85%, 80%, it's considered good. And very rarely we have a screening test, which is excellent, that provides 0.9 or higher area under the curve value. And the other thing here is that the best screening specificity and sensitivity is given in the corner, which is closest to the upper left corner here. And in our case, it was 10. So that would have been the best, the screening test in our case. And then we developed a simple assessment form of where people could easily calculate their own score. The only measurements where needed a body mass index and weight circumference. And in the Finnish risk score, we had on the other side, we had a table for calculating body mass index. And of course, by computer, it's very easy. So these are the only measurements would have been needed. And then we had another survey later on, and we tried to see how well this risk score is actually performing for identification of already existing non-diagnosed type 2 diabetes. And the risk score was working very, very well to identify the undiagnosed type 2 diabetes based on oral glucose tolerance test. Also, when we combine type 2 diabetes with the pre-diabetes abnormal glucose tolerance, the Finnish find risk was working pretty well, but already at the lowest level, about 20% of people had the abnormal glucose tolerance due to other reasons. Then the question is, what is the cutoff value we should use in a clinical practice? The best would have been 11, and then the sensitivity would have been about 70%, but it would mean that about 10% of people, however, I would say that only less than 40% would have been needing the glucose tolerance testing. Then if you go to the higher value points, then about 50% of people, 50% of sensitivity, and about 25% of people would require further assessment. In many places, 15 points has been used because about 35% of sensitivity, which is not too bad, but about 15% of people would require further assessment, which would be feasible in primary care in many countries. Now there are several other risk scores developed, actually many, and this is one of the European Danish risk score, and you can see here, the parameters are basically the same as in the find risk, except there is a gender that the men have a higher risk, which is true in the middle-aged men have a higher risk than the middle-aged women in Europe. We did not include the gender because after 65, women have a higher risk. So it would have been complicated to actually have it in the model. Another one is the UK EPIC-NOVOC risk score, which is based on the clinical information from a medical records. And that includes also the laboratory parameters, which is different than the others. But again, here, no laboratory test is needed because it's taken from the medical records. And the area under the curve was pretty good. Of course, not as good as the find risk, but still okay. Now, today, the find risk, when I looked at the PubMed, there were more than 200 publications where the find risk was mentioned in the title of a paper. And this has been developed in different countries, maybe slightly modified if needed, but really in many, many countries, it has been done and translated to about 50 different languages. Now, we applied also defined risk in the Finnish Diabetes Prevention Study, where we could show that the intervention worked very well. And we tried to stratify people by defined risk and in the control group, we could show that the higher find risk value at baseline, the higher risk of type 2 diabetes during the follow-up, whereas in the intervention group, the trend was flat, completely flat. So we could abolish this increment of type 2 diabetes with a high risk, with a lifestyle intervention, which is a good news. Also, in the others in there were randomized to a usual care and intensive intervention when they had been identified with a high value of find risk. And there was about 50% risk reduction with the intensive intervention in these people who have been identified with a high find risk value. Now, there are many other populations where defined risk has been applied and modified. And this is one of the completely different population from Oman and the differences are usually that the score points become different than originally. And so the relative value of a certain factor is different in different populations. And also in some populations like in Australia, ethnicity need to be included. In the UK, they were doing the risk score with the colors like traffic lights, green, yellow, and then red. I did my own risk score with this. And they also had a gender, male, female ethnicity because in the UK, many ethnicities exist and so on. So basically, otherwise, they had a waist and BMI and high blood pressure. So altogether, I got the 17 points, which means that quite high. But out of 17, 13 points came from my age, which is 70 or older. I'm not quite sure whether it's really that I should have a sort of moderate risk because of my age, because I don't have a risk. Anyhow, in Saudi Arabia, we developed, NOEL was included in the project, the risk score where we actually diagnosed diabetes based on the one hour post-challenge glucose tolerance, plus a fasting or hemoglobin A1c. And we could see that we could improve the find risk to some extent. We applied to find risk in the US population where the population was in Eric's study was including black and white people. And as you can see, the risk score was predicting diabetes basically in the same way in the black and white people. And then another very different population in Indonesia, also the rock curve, area on the curve, 0.73, not too bad. And the, whether it was the original or modified find risk was approximately the same. And a very interesting study was coming from Algeria where they had refugees from a Saharan area. And they found that the find risk is working very well with these individuals. The TEC2 is a multinational collaborative project where more than 18,000 people were curved, whether it was the original find risk score value or extended model. Extended model, a little bit better, but not really much. And so this sort of global study also produce information that this score is working well. We applied the risk score in the European D-Plan project where we had 25 centers from 17 European countries. And when we looked at the production of find risk for the oral glucose tolerance test at cross-sectional analysis. So we could see that there was a 0.74 in the area under the curve. So approximately the same as in many other studies. Some years ago, we published data from one D-Plan project cohort, which is a huge hunt study in Norway which is almost 50,000 participants and 10 year follow-up. And we could show that the increment with the find risk was as it was in the original find risk. But the only thing was that in the highest risk, only about 25% in 10 years developed diabetes while in our original was more than 50%. So a question was why less than in our original study, less people develop diabetes. So we looked at our data from another screening program where we had about 9,000 individuals and tried to see what actually is the risk with the highest category. And it was like in a Norwegian population about 0.3. So maybe there is something that has happened over the years that more contemporary populations may have a little bit lower risk than what we observed in the original find risk population developed based on 1987 data. And another issue is now the metabolic syndrome is predicted very well with the find risk. Coronary heart disease, stroke, all cause mortality, all of them, we had a collaboration with the Albert Einstein Hospital in Sao Paolo where we looked at the various things. For instance, here, C, reactive protein, very strong correlation with the find risk, Framingham CVD risk score. You can see that those people who had a high find risk, 82% had a high risk in the Framingham score, whereas in the low risk only about 28% had a high. And then litter, fatness was extremely strongly correlated with the find risk. So that when we looked at the area under the curve, it was actually more than 80% for men and 0.9 for women, which would have been excellent prediction. So for many other outcomes, we have a possibility to look at the find risk. Also with the quality of life, this was our Finnish cross-sectional study where we correlated find risk with the quality of life and whatever way to measure quality of life, there was a very strong correlation. The higher the find risk, the lower quality of life. Now, how this risk score is being used. The first use was actually the National Diabetes Prevention Programme in Finland, where we had a population strategy and a high risk strategy where we actually used the find risk. And then if people had over 15, we had a cost tolerance test and then interventions. Now, the knowledge we had already was that it's not only over 15 that have a high risk, but it's also up to 17, slightly elevated, moderately elevated from 12 to 14, so that these people need some information and we provided them information, but not the glucose tolerance testing. Only the glucose testing was done over 15. The others received the lifestyle intervention advice. So in Finland, when we started to apply the find risk, there was a huge increase in the number of individuals who went to the website of the Finnish Diabetes Association and looked at their own risk. And then we couldn't have a campaign sign, mass media, immediately a lot of people went to do the risk test. So it's a combination of making a public health announcement to go there and do your own testing. And that really leads to a huge amount of individuals who are actually doing it. So these are the things we included and they all were based on the empirical risk assessment, and then we could inform people what to do. In the International Diabetes Federation a couple of years ago, they introduced prevention of diabetes as a year of diabetes, and they included a defined risk in the risk assessment. And it is available there in many languages. And therefore people can use it online at the IDF website. We have tried to see if a new model could be developed. Model could be developed. And we took more recent surveys that have been carried out in Finland with a large number of individuals. And we evaluated if the original risk score is better than the new one, and the new one produce exactly the same result as our original one. So in summary, diabetes risk scores are very, very useful to detect undiagnosed diabetes or dysglycemia, including the prediabetes. They have been validated in many different population ethnic groups and slightly motivated. They can very well predict the future development of diabetes. But some population-specific issues need to be considered. The main parameters are the same. Human are the same, but the cut points may be slightly different because of various factors. I don't think that a single universal diabetes risk score is possible. It's not possible, but it is possible to implement scores in all populations, but we need some sort of validation. And we have demonstrated, and a diabetes risk score is not only predicting diabetes, but also identifying people at the risk of other non-communicable diseases and health problems. And good news is that we have evidence that people at the high risk identified by a risk score can benefit from healthy lifestyle advice. Thank you. Thank you very much, Dr. Tuomi-Lehto. And next, I would like to introduce Dr. Pablo Ashner. Dr. Pablo Ashner is Associate Professor of Endocrinology at the Havariana University School of Medicine. At present, he's also Director of Research at San Ignacio University Hospital and the Scientific Director of the Colombian Diabetes Association, Bogota. He graduated in medicine surgery from the Havariana University, Bogota in 1974, and then took postgraduate training in internal medicine and endocrinology at the Military Hospital, Bogota and at Cambridge University in the UK, qualifying as an endocrinologist in 1982. He has also a master's degree in clinical epidemiology. His main research interests include diagnosis, control and treatment of diabetes, epidemiology of diabetes and its complications, as well as the metabolic syndrome in Latin America and the role of diabetes associations in primary healthcare. He has authored or co-authored over 90 abstracts, articles and book chapters in the field of diabetes research. He's also a member of numerous societies, including the Latin American Diabetes Association and the Pan American Endocrine Society. He served as a member of the WHO Expert Advisory Panel on Chronic Degenerative Diseases and the WHO Ad Hoc Diabetes Reporting Group. He's also the past president of the Colombian Endocrine Society, the Latin American Diabetes Association and Latin American Diabetes Epidemiology Group. And currently he's the president of the Latin American Group for the Study of Metabolic Syndrome and serves as chair of the IDF Task Force on Guidelines. Dr. Ashner, welcome. Thank you so much for accepting our invitation. Thank you. Thank you, Noel. I'm just looking at, trying to upload this. Okay. The problem is somehow I am not finding my presentation. I'm sorry, but... Don't worry, take your time. Well, I shouldn't. When I tried to share it before it was, I was unable, it was closed and now that it's now that it's available When I had the same problem, I went back and re-opened my presentation. Oh, okay. Okay, I'll try to do that. Thank you. Okay. Well, it seems to be stuck. But I'm... Yeah, it's, it's stuck. It's not responding. Can you maybe close your presentation PowerPoint and... That's what I'm trying to do. And it seems to be... Wait a minute. I'm going to... Okay, now I could close it and now I'm going to open it again Okay. That's it. And now I can share it. Right. Okay. Sorry for this delay. Technology. Okay. So thank you again for the invitation. And I'm going to show our experience with the risk or particularly the fin risk in or find risk in South America and specifically in Colombia. And I want to start by showing that the impure glucose tolerance is quite frequent in the Americas, both North, Central and South America. It's around 10 to 12% the prevalence in most countries. There are hotspots like Mexico, which is considered to be part of North America, also in Nicaragua and in Guyana. But otherwise it's more or less similar around the Americas. As I said, if we compared North American and region with South American and Central American region, we can see that we have less, a little bit less prevalence of IGT, probably because North America includes Mexico. And also same with prevalence of diabetes. But I want, what I want to show is that in our region, there are more people with IGT than, than diabetes. That means that there are more people that risk of developing diabetes than actual people with diabetes. And the conversion rate from IGT to diabetes worldwide has been more or less around 13% at one year and 60% at 20 years with a wide range between different countries and regions. There was a study performed in Colombia, which was led by our chairman, Noel, and also with Jaco was involved. And it can be calculated from that study that the annual incidence in our region is around 4.6 to 6.5%. So it's, it's an important issue, this high risk of people with IGT and the high conversion rating, at least in some regions. And the other problem is that in our region, around one third of the population with diabetes is unknown, undiagnosed. It's about a half in, sorry, a quarter in North America, but in our region, it's about one third. Again, an important problem. So we do need to use risk scores to identify both people with IGT and people with unknown diabetes, which, who by more or less definition are asymptomatic. As Jaco already mentioned, the test should be the score, risk score should be non-invasive and should be validated to identify people with high risk in our case, both for unknown diabetes or IGT, because they would be the people candidates for blood testing. So that's the main purpose from my point of view of the risk scores, at least to identify people who have the problem, not, not, there's the other option to predict people who will develop the problem, but that's, that's not our case right now. So fin risk is the most used risk score in Latin America and has been endorsed by IDF. As Jaco already showed, this is more or less the risk score where we have different characteristics, which don't need blood tests, except you need to measure height, weight and waist circumference. There's an issue which should be changed or perhaps or validated in, in Latin America. And that's the issue about waist circumference because it varies from one region to the other. So what we did is to adapt the waist circumference for the values for our Latin American people, which came from a study we did in five countries in Latin America. And our, our waist circumference, our cutoff was 94 for men and 94 women. So we changed the waist circumference score, giving the highest score to those who were above that cutoff, one, one single cutoff. And with that change, we also changed some questions like, this was actually changed in other areas as well in Spain, for example, and other places, the physical activity, instead of asking if you do less than an amount, a monthly, weekly amount, we actually ask, do you do at least 30 minutes per day of physical activity? Obviously we, we didn't include Paris because that's mostly from the, from the Finnish score and other minor changes. But the change in the waist circumference didn't actually make any difference in men, exactly the same area under the curve, curve, whether you use the original fin risk of, or what we call the LA fin risk or fin risk LA, which has this change in the waist circumference. But in women, it does make a difference. The fin risk with the waist circumference adapted for Latin America definitely has a higher area under the curve. And these changes, these differences are significant. This is for identifying people who might have diabetes. And when we included IGT and IFG, that means identify people who may have diabetes or any other glucose abnormality, exactly the same. In men, there's no change, 94 centimeters of waist circumference is exactly the same as the cutoff for Europe. But in women, if there is a change, the 90 centimeters gives a higher area under the curve. And the difference is significant. There has been attempts to simplify this fin risk. And actually our chairperson, our chairman, Noel, did a study here in Colombia with more than 2000 people trying to simplify the risk score by finally only having four variables, sorry, five variables, and using our own cutoff, the 94 and 90. And with this simplification, sorry, it's four. The point is that BMI was not included as I'm showing it here, but it was not included. And with those four, he actually showed that it did work. But the fact that BMI was not included might be a barrier because actually identifying people with obesity is an important, it's an important modified risk factor. So it should be in the risk score. But in fact, with a simplified risk score, which was called cold risk, actually the area under the curve is exactly the same as with the fin risk and the Latin American fin risk. So it can be used as the other one, the fin risk LA, except for what I mentioned, that it doesn't include BMI and it actually should. So I think that the fin risk LA is easy to perform. And it's so from our point of view, it's validated for our region. So we looked at the best cutoff, the best score to identify people who could have diabetes or IFG or impaired glucose tolerance. And in fact, the score of above 12 was the best score with a high sensitivity and an acceptable specificity. And that was what we suggested as the cutoff to use the fin risk in our clinical practice. And in fact, it was included as a recommendation in our clinical practice guideline, which was commissioned by our government, by the health ministry. We have universal coverage in Colombia and for the health ministry, it was important to have an evidence-based guideline, which would serve as the guide for treating diabetes in our country, type 2 diabetes, trying to be as more evidence-based as possible, but also when possible, cost-benefit proof. And actually, as part of this work we did, we did a cost-effectiveness study of the screening model and screening and diagnosis model, comparing using the fin risk before doing a blood glucose, a fasting blood glucose, which would then go to an OGTT. So we compare that with just starting with a fasting blood glucose. And this was for testing people in general population. And the fin risk came out to be cost-effective and therefore it is recommended as a screening method for type 2 diabetes in Colombia with a cutoff of 12 points to proceed to a diagnostic test that would mean an invasive test, a blood test. And the recommendation was strong and the evidence was considered moderate. So right now, fin risk is included in our national practice guidelines for diabetes. Now, there have been many studies proving the benefit of prevent diabetes prevention, primary prevention in people with impaired glucose tolerance with different strategies, including many drugs and a bariatric surgery. But I want to stress the fact that the diet and physical activity are quite effective in preventing diabetes as has been shown, including the DPS study led by YACU. So we wanted to use the fin risk to identify people with IGT and put them in a primary prevention program in two cities in our country, Barranquilla, which is in the Caribbean coast, and Bogota, the capital, which is in the Andean region. So we wanted to screen people with a fin risk LA and then incorporate those who had IGT in a prevention program. We knew by the previous study I already mentioned that if we screened people with the general population, around 35%, one third of that population would have a high score. And among them, around 17% would have impaired glucose tolerance. So we designed this study. I'm not going to explain it completely because I'm just interested in showing you what happened with the screening and how we identified the IGT people and then what we did with them. And in fact, we screened around almost 4,000 people because we wanted around 150 people with IGT. And we found that actually those who had higher than 12 points in the risk score were not a third, but a half. And then when we tested them with the OGTT, around almost 14% had IGT, which is a bit lower than was found in the previous study I mentioned. Almost 6% had type 2 diabetes and they were referred to the health service, but those with IGT were included in a structured program, which I will explain now. But how can we actually lower the people with a high score and have a higher yield of people with impaired glucose tolerance for these kinds of studies or these kinds of strategies? Well, we can increase the score, as already Jaco mentioned, and if we use more than 14, then we increase the specificity to 81%, which is much better to give a better yield of IGT. And the sensitivity is still not bad, 64%. There are some questions which need to be adapted, transculturized, as one could say, which might make the test more precise. And one of the issues is the issue of lack of physical activity. Originally in the FINDRISK, the question was the lack of physical activity, which included individuals who didn't do any strenuous physical activity during work or during their leisure time, spare time. But the question has now turned into the daily physical activity of more than 30 minutes, and that for many people, and we found that in our study, this PREDICOL study, many people interpret it as actually extra physical activity or exercise, this kind of exercise. And so they all answered they didn't. They didn't do more than 30 minutes per day, which would be equal to more than 3.5 hours per week. 80% of the people in our population we studied actually said they didn't do any kind of any physical activity because that was the image they had. But actually, the people we studied live in the low and medium income area of the city, where they have to walk more than 30 minutes every day to the market, to work, etc. And the streets are quite steep. This is an image of the area, although we cannot see in detail, but these people do walk a lot during the day. So I think this question has to be reorganized to really identify what we mean by physical activity. Anyway, those people who had IGT were included in a diabetes prevention program based on the National Diabetes Prevention Program developed by the CDC in the U.S. and adapted for Hispanics. But that program includes around 26 sessions, and participants must complete at least 22. So we asked Betsy Rodriguez, who developed the adaptation for Hispanics, if we could adapt it even further for our study by reducing it to just 10 group sessions related to nutrition and physical activity. And actually, we did that intervention during one year, and these sessions were led by coaches from the community, which we trained with an investigational research group. And We believe that this is a better idea, more feasible, because we cannot use diabetes educators for this. They have a lot of work with people with diabetes, we trained coaches from the community. We managed from the 153 subjects who were identified with IGT, we managed to get around 57 of them to complete the program. These are the characteristics. There was a change, a significant change in the median to our blood glucose post-glucose oral load of 75 grams. We did see a weight change in one year. The mean was 2.4 kilograms, but among those who lost weight, which were the majority, 78%, their mean weight loss was almost 4 kilos. We were aiming at 5 kilos, but we believe that almost 4 kilos was quite a good result. Regarding the glycemic status at one year, 60% of those people who went through the program reversed to normal glucose tolerance, and only 1% progressed to diabetes. This is a quasi-experimental study, so we cannot actually we do need to study further these results, but we can conclude that the fin risk is an adequate method to screen people with unknown diabetes in the general population. That's why we included in the guideline, as long as the identified cases are properly diagnosed and incorporated in the health system, so that they will get the proper health care. This we can more or less guarantee because we have this universal coverage, but it's not the same in all the countries. Also, fin risk, this adaptation, the fin risk delay, is an adequate method to screen people with impaired glucose tolerance who have a high risk of developing diabetes, as long as the identified cases are incorporated into a diabetes prevention program. We have shown that this diabetes prevention program we have adapted from the CDC Hispanic program has demonstrated its feasibility, at least in our country, and its effectiveness, although it's a quasi-experimental trial, to reduce weight and reverse to normal glycemia in people with IGT. Thank you very much and looking forward to the questions and answers. Thank you so much, Dr. Aschner. We can now welcome our last presenters, Dr. Nandita Arun. So Dr. Arun is Director and Consultant Diabetologist at the Dr. A. Ramachandra Diabetes Hospitals. She did her undergraduate studies in Sri Ramachandra Medical College in India, during which she received the Gold Medal in Preventive and Social Medicine and post-graduation a medical degree in General Medicine. Dr. Arun underwent training for research at the University of Cambridge in the UK, training for diabetes at Christian Medical College Bellore, and received a diploma in diabetes from Cardiff University. She has also conferred fellowship by the Royal College of Physicians, Glasgow, and a fellowship by the International Medical Science Academy. She underwent training for healthcare management at the Indian Institute of Management. Dr. Arun is a member of several diabetes and medical societies. She has been an organizing member for many international conferences. She has several international publications and presentations. She's a co-investigator in many drug trials and she's on the advisory board of various committees. Dr. Arun received the Young Research Achiever Award at JPF in 2019. She was honored with a COVID Warrior Award in December 2020, and she was acknowledged as the Diabetes Care Personality of the Year. Dr. Arun received the Outstanding Young Investigator Award at the TN Kidney Foundation Tanker Awards Ceremony, and the RSSDI Young Investigator Award from the Research Society for the Study of Diabetes India 2021. So it is now my pleasure to turn it over to Dr. Arun. Welcome. Many thanks for accepting, having accepted our invitation. Thank you. Thank you so much. Very happy to be here. So after those two wonderful lectures that we've heard from Professor Yako and Professor Pablo, in the next few minutes, 15 to 20 minutes, I will be talking about the similar programs that we have conducted and from the Indian perspective. So I bring greetings from Chennai, which is the southern part of India. And before I move on to the risk core per se, I am first going to present to you data about which is the prevalence of type 2 diabetes in our part of the world. As many of you might already be aware, India is one of the countries with the largest, one of the largest countries with the largest number of people living with type 2 diabetes, and with a growing prevalence of type 2 diabetes. And this is a slide just that shows us just that, the prevalence of type 2 diabetes. On the left, you see an India map with the prevalence statewide in 1990. And about 15 years later, the same statewide prevalence, as you can see, the colors have actually changed and transitioned from the light blue and the blues that you see, which is a lower prevalence of less than 4 or 5%, which is all transitioned to an orange or a red or a yellow, which is a higher prevalence. And in southern part of India, which is where we are from, Tamil Nadu and Kerala and Karnataka, is where you have very high prevalence of type 2 diabetes as well. Now, we also earlier did believe that type 2 diabetes was more affected due to urbanization, and it is more in the urban areas in India. However, this is data that you have in front of you that I'm showing you here. We have actually been reporting the prevalence of type 2 diabetes and again in southern India from Chennai and Tamil Nadu, in both in the rural and in the urban areas. And what we have actually found, what you see here in front of you, is that there is in fact a significant exploding increase in explosion of prevalence of type 2 diabetes in India, in southern India. But what I want to bring to your notice or what I want you to take note of over here is that the significant increase that you see is not just in the yellow bars, which you're seeing here, which is not just the urban areas, but also in the rural areas. And this is very striking because as I mentioned, it used to be believed that type 2 diabetes is more prevalent in the urban areas and it's more a disease of the affluent, but it is not so. Also, the villages are getting equally affected, and a mere prevalence that was just a single digit around 2% in the 1980s is now in 2016, the last that we had reported it, almost 14%. So, we have exploding numbers in our part of the world. What is even more alarming, and Professor Pablo also showed us his prevalence of pre-diabetes in South America, but here you have the prevalence, this is the change in the prevalence of pre-diabetes or the trend in the prevalence of pre-diabetes in India. So, we have a comparison between, again, the urban and the rural areas. You have the city, town, and the village, and you have the total prevalence projected in front of you, comparing it between the city, town, and the village, right, over a period of a decade, right, from 2006, and then going back to the same areas in 2016. And this shows us that there is a significant, sorry, a significant increase in the prevalence of pre-diabetes, again, not just in the city, but also in the town, and also in the rural areas, also in the villages. And this is something which is very, very alarming, because just like Professor Aschner was addressing, people with pre-diabetes are a ticking time bomb, and these are people who are waiting to convert to type 2 diabetes. And I presented to you the previous few slides, a growing prevalence of type 2 diabetes in our part of the world. So, when you also have an alarming increase in the number of people with pre-diabetes, this is just to show you, this is the huge burden that we have, that we need to deal with. And here, you also have data to show you that there is a rise in prevalence, not just in the male, but also in the female counterparts. Now, when we tease out the same data that we looked at the previous slide, this is the total diabetes prevalence in the city, town, and the village, which is 22% in the city, 20% in town, and around 14% in the villages. When we tease out the newly detected cases from this prevalence that you have here in front of you, in the city, the newly detected patients or the newly detected subjects contributed to around 8.4%, in the town, 8.3%, and the newly detected prevalence was around 6.7% in the villages, which then, what that shows us is that the undiagnosed cases or the undetected cases was actually almost 50%, about 38% of cases in the village, in the city, 41% in the town, and 50% of cases in the villages remained undiagnosed or undetected and unaware that they have type 2 diabetes. So, the best solution to this would be, as was addressed in the previous talks as well, mass screening and screening to detect patients, early detection of patients with type 2 diabetes so that they can be treated to prevent the risk of complications as well, which is what brings us to the Indian diabetes risk score. The risk score is nothing but a very simple screening tool and a non-invasive tool, more importantly, which helps us to identify patients with higher risk of type 2 diabetes. And how was this tool arrived is what I'm going to talk to you about in the next couple of slides. So, we had data with us from a national survey that had been conducted and we had data for about 10,000 subjects from various cities with a good representation across India, southern India, Chennai, and Bangalore, northern India, which was New Delhi, central India, which was Calcutta, eastern, and almost every city in India, which was covered, and we had data from 10,000 subjects. These 10,000 subjects that we had data, these are people that we had screening data from. We took that data and what was actually done was a multiple logistics regression analysis with diabetes as a dependent variable. Now, this helps us identify the risk factors which significantly contribute to an increased risk of type 2 diabetes. So, the factors that actually turned out are what you have here in front of you and the significant p-values is what you have on the right-hand side. So, an age as with an increasing age, an age less than 30 is not significant, 30 to 44, 45 to 59, and more than 59 are the significant values that you had. As you can see, the p-values were significant. The male gender was not significant because the p-value was only 0.7. A positive family history of type 2 diabetes was another significant variable. A BMI of either more than 25 or less than 25 also turned out to be a significant variable contributing to a higher risk of type 2 diabetes. A higher waist circumference among males more than 85 centimeters and females more than 80 centimeters and a sedentary lifestyle or sedentary or light physical activity. So, these were the significant variables when we derived it through the MLR or the multiple logistic regression analysis. Once we found these significant variables, the beta coefficient, I think this was also explained in the previous sessions, were then multiplied by 10 to derive the risk score. And so, you have age where you have a beta coefficient of 1.05 for 30 to 45, and when multiplied by 10, you get the value of 10 and so on for each of the variables. This then brought us to derive the diabetes, Indian diabetes risk score, and here you have in front of you the complete score. The variables, as you saw in the previous table, age 30 to 44 years, 45 to 59 years, or more than 59 years, and each of these has a risk score of 10, 18, and 19, respectively. A positive family history of type 2 diabetes would have a contributing risk score of 7. A body mass index of more than 25 against 7. A higher waist circumference, 5, and a sedentary physical activity would be 4. The maximum score when you add up should be around 42, and any person with a risk score of more than or equal to 21 has a higher risk or a higher probability of developing type 2 diabetes or being undetected with type 2 diabetes. So, once the risk score, and this is what the score that you have in front of you, was derived, the next step was to validate this in the Indian population. So, then what we did was went ahead and tried to validate this. We derived the data, as you saw from the previous, to the MLR with diabetes as a dependent variable, and all that was done in the first 5,000 subjects from the previous 10,000 that I showed you earlier. So, the risk score was then validated in the remaining or the second half of 5,000 in the national data that we already had. The score was also found to be valid in another previously conducted Chennai example that we had, a survey that we had conducted earlier. Thirdly, the score was not applicable and not validated in patients in the white population in the UK. So, this was what helped us to validate the score and actually find its applicability. We then went ahead, and again, this was discussed in the previous session, but this is the ROC curve with our Indian scoring, to look at the sensitivity and the specificity and the area under the curve. This was very well explained by Professor Yako that the area under the curve, anything more than 0.6 or 0.7 is said to actually represent a higher sensitivity or a specificity. And from the ROC curve, we were able to infer that the sensitivity and the specificity of our score was actually quite good, with a sensitivity in the first cohort of around 76%, 72%, and 73% in the first, second, and the third cohort, and a specificity of around 60% in all the three cohorts. So, the use of the risk score has many advantages. The most important advantage is that it helps us to identify those people with a higher risk of undetected diabetes. It also avoids unnecessary testing, especially when we want to conduct mass screening. Unnecessary testing and a waste of resources can be avoided. It is cost-saving. It is a non-invasive, simple tool, which can be done by any layperson, and even a paramedic staff or anybody can use this in their clinics. And therefore, being such a simple non-invasive tool, which is easy to use, this can actually be used in the national scale all over the country in any clinic for screening patients. Thank you. Thank you very much to all the speakers of today's webinars. Excellent talks. We have now about 15 minutes for questions and answers. I saw there were some questions, I think, in the chat, but also in the Q&A, and I'm going to select the first ones we have in the Q&A. There are some already answers, but let's have the open question first. The first question is, any comparison made between the FINDRISK and the UK PDS risk engine? And any of the speakers, if they may know, maybe Professor Tuomilehto, it's a European thing, do you know anything about that? I don't know, as we have seen here now, the diabetic risk scores usually have risk factors non-modifiable and modifiable. And these are relations to some extent. But not much, the risk factors are the same and different populations at different time points or different years may have different interpretations. So therefore, when we are comparing results from one study or one population to another population, we have to consider what are these populations, what is their history, socioeconomic status, and so on. I would like to make one example. In 1980, there was a national diabetes survey in China. Prevalence of diabetes was 0.8%, 0.8%. Less than 1%. And later on, we have seen that China has a huge prevalence of diabetes nowadays. But in a short period of time, we can see the huge differences. The other way we have to interpret the data these people who were living, who were born at a certain period of time in China, they were not living long enough to develop diabetes. Now they do. And this is the long everything is some factor that has not been taken into account in a proper way. When we are comparing populations. So it's very difficult to compare populations without having, I would say, complete understanding what the populations are, what the history of the population, what the current status of a population, what is the survival of a people who are living nowadays and so on. So, yeah, it's a very complex issue, but at the end, and we have a data from today, and that's what we need to use. Okay. Thank you very much, Jaco. And if you, Dr. Asher, please. Yeah. I would just like to add that as Jaco showed, the finders could also predict cardiovascular disease without actually doing any invasive tests. The UK PDS does include invasive tests, lipids and HbA1c. And on the other hand, it's applicable to people who already have diabetes. While the find risk is earlier, it predicts cardiovascular disease, even in people who don't have diabetes. So I would say that the aim of this course, and therefore the use of this course, is quite different. I think they might correlate in predicting cardiovascular disease, but their aim, and as I said, and applicability are quite different. Thank you very much to both of you. There's two questions to Dr. Nandita. The first one is, how do you derive the cutoff point of 21 to classify people at high risk? And the other one was like, did you investigate where the cutoffs for scoring might be different for North Indian versus South Indian populations? Right. So the first question is a cutoff value of more than 21 is what showed us a higher sensitivity of an increased risk of type 2 diabetes when it was applied to the population. To the population that was being studied. And the incidence of type 2 diabetes actually shoots up, it doubles once the score crosses 21. So that is the cutoff value of how that was derived. The second question was, what was it again? Was it validated in the northern parts of India? Yeah, between North and South India. Right. It was not tested in different populations within India, but it was derived as I projected in my slide. It was a pan-India population that was included in the study to derive the score itself. So there was a good representation of subjects all through right from Southern to Northern India. We had subjects equally distributed right from Chennai to Hyderabad to Maharashtra and even up to New Delhi. So there was a good representation all through. Thank you very much. There's another interesting question from Michael Bergman that he asks is, is it appropriate to refer high risk individuals to lifestyle interventions or prevention programs without an oral glucose tolerance test? I would like to answer yes, definitely. This is a real primary prevention and that how we should do it and not go for a glucose. We should have the risk and based on the risk to do whatever needs to be done. And as I showed in my presentation, then with a simplified risk, we could easily identify many other health problems, not only diabetes. And to do interventions, that would be fine. No need to do a laboratory test. But people like numbers. Dr. Arsene, what do you think? Well, there's the issue of whether to aim the intervention at the higher risk and the benefit few people or aim it at a lower risk and benefit a little bit less on more people. And there's no real final solution to that. It depends on the resources. In our case, this program we developed, although it was reduced to 10 sessions and relatively, I would say, not very expensive. We used coaches from the community, et cetera. The adherence to the program was not as good as we expected. And this was among highly motivated people because we demonstrated that they actually had abnormal glucose tolerance. We did include an intervention on the community and even on medium, I mean, on people who had more than 12 in the risk score, but didn't have IGT in the OGTT. We haven't analyzed that, but it might give some insight into this issue of whether they also would benefit. So we're going to analyze that and see whether the OGTT wouldn't be necessarily at the next test to do. Okay, thank you. There's one more question related to the models of developing the risk score from Melina Selbrand. And she asked like, why isn't smoking included in the risk score? Well, I can see this is important question. Smoking doubles the risk of type 2 diabetes. On the other hand, when we're smoking in our fine risk, a new model, it did not improve the prediction significantly. The one thing is when we have a multivariable model to predict a risk of disease, when you have five or six variables, whatever their initial, or significance is another one is very, very limited. So if you have five variables and you have five more, you don't really much improve the model. And that's partially statistics, partially it's multicollinearity and some other reasons. So to build the model where you have too many variables, it's not useful. And for practice also, it is useful to have a minimum number of variables in the model to predict a risk. Yeah. I would add, I don't know if you agree Jaco, that maybe there's also interaction, for example, between smoking and BMI. So that in a model, the strongest variable will sort of push out the weakest variable if there is interaction, but I'm not sure about. Hmm. Okay. Dr. Arun, in diabetes score, did you check for smoking as well or just your conclusion? Yes, of course, all the variables. So when the MLR is, I think what both Dr. Jaco and Ashna said, I agree with them. It's only about what factors turn out to be significant when you run the MLR. So if the stronger factors and the most significant factors which turn out is what we derived into the score or included into the score. So since smoking did not turn out to be a significant contributing variable, that's the only explanation for, that's the only reason why it was not included into the score. But like Professor Ashna was saying, maybe smoking is strongly associated with a higher BMI or hypertension or many of the other factors, but for type two diabetes, it did not directly turn out to be a significant variable when the MLR was run with diabetes as a dependent variable. Okay. Thank you very much. As we, there are no more questions at the moment. We also have only two minutes left to the webinar. So I would like to ask each of you, please, if you can give me maybe in 20, 30 seconds, your main conclusion or main recommendation in regard risk scores to our participants of this webinar, please. And let's start with Dr. Arun, ladies first. So your summary of. Well, first of all, thank you for having me here, Dr. Noel and Dr. Patoma. And I'm sorry, I hope I got your name right. It was an absolute pleasure. And well, I think mass screening is one of the best ways to go about this. And I think there's two ways to, we are dealing with a huge burden of type 2 diabetes, both globally and in our part of the world. I presented to you data that we have huge growing numbers in India as well. And the best way to handle this burden would be early detection and screening, number one, and of course, primary prevention. So if we could identify people at higher risk of type 2 diabetes and intervene earlier to prevent the conversion to type 2 diabetes, then that would be the best way to handle this burden. And here we presented and we've discussed a simple tool that makes life a lot easier, that empowers not just doctors or physicians or clinicians, but even anyone, like I said, a paramedical staff or anybody sitting in a clinic can actually just run through this code and it helps to identify people. It helps in saving resources, saving time, and it is non-invasive. So I think this should be utilized at a national level in each country. And of course, like I think Dr. Ashner was also saying, each country should have a specific risk code that will help them identify patients as well. Okay, thank you very much. Let's continue going to the time zones. It will be Dr. Dhuamilek, next. So prevention of a disease, non-communicable disease, requires both a high-risk approach, population approach. If not doing both at the same time, the efforts will be really, really minimal. I mean, the benefits would be minimal. And this is very often. We may consider the huge efforts for a high-risk approach, like identifying high-risk people for diabetes, but at the same time, it's not supporting preventive measures. So we have to coordinate both of them. But on the other hand, as a medical profession, we have a reason, an important reason to help people who are at risk or to have a significant disease like diabetes, which will... So, yeah, it is important to identify people, to help them, but at the same time, I think we who know what to do, we should try to influence our societies, whatever way we can to improve the situation, to make the modifiable more easy. Today, unfortunately, industry is working much more efficiently than the healthcare system. Okay. Good, thank you very much, Dr. Tummelecht. And finally, Dr. Aslu, your final remarks. Well, I agree with what has been said. I just want to say two things. The first is that I believe this is the 10th year after the first publication of the FINRISK in diabetes care, is it not, Jaco? I think it's 10th anniversary. Yeah, yeah. So congratulations. Thank you, thank you. And I think that the FINRISK actually is a good tool, at least in our part of the world, but I would just insist that screening by itself does not improve health. You have to be prepared to what to do with those who are screened positive, both if you are looking for undiagnosed diabetes, or if you are looking for IGT, usually both. And you have to have a strategy of what to do with those people. Otherwise, screening by itself would be useless. Thank you. Thank you very much. So finally, I would like to thank each of the speakers for being available today for our interest group webinar. Please, everybody look out for our next events of our public health epidemiology interest group. And for all members of our group, if you're not members yet, please sign up for members and send us emails if there's anything else you'd like us to do for you. Thank you very much. And also thank you for everybody at ADA who helped organize this webinar. And good night to Dr. Haroun and Dr. Tomi Lechte. Thank you very much. Okay. Thank you. Thank you.

webinar

diabetes risk scores

type 2 diabetes

International Diabetes Federation

diagnosis

non-invasive risk scores

diabetes screening

development

implementation

validation

benefits

early detection

complications