‘Diabetes Control A Good Indicator For COVID Outcomes’
Mumbai: Many people lose their lives to COVID-19 because they also suffer from pre-existing conditions or comorbidities. And a key comorbidity is diabetes. A database in Scotland tracks more than 300,000 diabetes patients in the country, including their current health condition, and helps maintain a continuous communication channel with the patients.
We speak with Deborah Wake, co-founder of MyWay Digital Health, who started this as part of an academic project which later turned into a venture. The database puts together the data and then uses machine learning and artificial intelligence to understand the trends and to respond to them. Wake is a physician and a clinical reader in medical informatics and diabetes care at the University of Edinburgh.
Tell us about your database, and how it links to the challenge of COVID-19?
Over the last 15 years or so, in Scotland, we've taken a national approach to managing diabetes. We've linked up all the data from all the healthcare systems, including the primary care, general practice systems, the hospital systems. We've also started bringing in data that people with diabetes themselves produce: Things like data from glucose meters--they might be measuring sugar levels at home--activity monitors, weights and blood pressures and other things that they can do in their own houses. We use that data to help understand our population better. It allows us to respond in a very personal way to them and their needs.
We use it [the database] with clinicians to help us understand the problems patients are having. The computer simulation of the data can also help us personalise our approaches to these patients. It can pick up the high risk patients--for example, the ones who overdo certain testing, the ones who are likely to have a reaction to certain drugs, and the ones who are going to respond in a specific way to certain treatments. That way, we can plan and support them personally.
We also use our computer systems to communicate directly with patients--not just with the clinicians--by helping them understand their condition better, giving them access to their data, and explaining it in a really understandable way so that they know what their targets are for the key parameters, and also giving them personal links to things like self-management advice, education courses and direct feedback on the results, which will hopefully improve their healthcare.
How did the database come together? How do you keep this database running actively? For instance, right now, do you know how many patients across the country are following the advice of their doctors?
Yeah, we do. The database is pretty unique--99.9% of people with diabetes in the country are registered on it. And pretty much the day they are diagnosed with diabetes, the systems will pick it up, and push the information onto our registries. We collect information at least on a daily basis from all the key systems. So it is completely up to date in terms of the data--to the extent that if I see a patient in the hospital clinic, and they've been to the general practice the day before, had some tests done, all that information is available to me. And a lot of the detail will come in almost real time. So it is a real, live functioning system that is continually updating and through the additional artificial intelligence, machine learning analytics that we put into the system, it can constantly be flagging--again in real time--patients that are high risk, patients that should get started on certain treatments, patients that need additional support.
That's really important at the moment with COVID, because we're trying to reduce the number of patients we see face to face in the clinic, because we know there are significant risks with clinical contact. So it helps us stratify which patients we really need to be seeing face to face, [and for whom] we can safely defer treatment. And it helps us rationalise the limited resource we have and keep people safe during COVID. We're also using a lot of the remote communication functionality with patients so that if we're not bringing them in, we can communicate with them through the platform, get them to upload data, and keep that conversation going. We've now been able to develop it into other areas. So similar systems are now running across large regions in England. And we're starting to pilot those in international territories as well.
What is your sense of the number of patients with diabetes who have died from, or have progressed to severe COVID-19 in the last few months?
We know that diabetes is a progressive condition--particularly type II diabetes. So the longer you have it, the more likely you are to develop complications like heart attacks, strokes, amputations, kidney failure, blindness, and so on. And these complications are very costly. About 80% of the cost of diabetes is spent treating these complications, and for most healthcare services, between 10% and sometimes even up to 20% of their entire health budget is spent treating diabetes complications--most of which are preventable, most of which we know the cause of, and if we implement the right treatment at the right time and get these patients to better manage their lifestyle, most of these could be prevented.
In terms of COVID, more recently, we found that people with diabetes are at much higher risk of having poor outcomes with COVID, either being severely infected or dying from COVID. And in fact, in the UK, about 25%-30% of COVID deaths were in people with diabetes. In other environments, as much as 40% of people who died from COVID had diabetes. And we know the main risk factors are age--if you're older, then your outcomes are poorer. But also, it's linked to things like obesity--the more overweight you are, also the higher your blood sugar levels. And in general, the control of your diabetes is a very good indicator for how you're likely to respond and what your outcomes are likely to be to conditions like COVID. So the more we can do to help people keep their weight managed well, keep their sugar levels managed well, and reduce other risk factors like blood pressure, cholesterol, and self-manage their condition better, we can really have a huge impact on their outcomes for conditions like COVID--and more generally, the risk of developing other diabetes complications.
You mentioned that 80% of costs of managing diabetes come from other conditions. How have you seen this linkage between the two evolve? Are people able to understand what happens when they have diabetes, and how things can go wrong?
One of the key things we're trying to do is use the data to really help inform patients and give them a much better understanding of diabetes. When people are diagnosed, they may have limited symptoms, [and] these complications occur a few years down the line and often people either won't take notice of that or just may not have a good awareness of the risks that they're putting their body under, by not maintaining good control of their diabetes.
Within our system, we are trying to make it clear to patients what their targets should be for things like blood sugar, blood pressure, [and] cholesterol. If they're not achieving those targets, we push them more information to help them understand why they need to be achieving those, and what the long-term risks are if they don't. The other thing that we're doing in a more granular way is developing a risk-prediction engine for patients, which allows them to look at their current parameters and it will predict, on a very personalised basis, their risk of all these different complications. We don't want to scare people. But we want to be realistic about what risks there are.
But more importantly, what we then do is we give them a separate engine that allows them to change those risks, so they can see what happens if they give up smoking, or lose 10 kg, or start taking their cholesterol tablets again. They can really begin to understand the impact they can make by lifestyle changes to their downstream risk. And that can be really quite eye-opening for patients to be able to kind of actually see that, visualise it, understand what they can do to modulate that risk, and it can be a real motivator. And we feed that into the kind of goal setting and behavior change that we then work with patients to try to achieve.
How many patients do you have right now as part of this database?
In Scotland, we've got just over 300,000 people with diabetes in Scotland, and they're all contained within our diabetes database. A certain percentage of them will have chosen to also use our data-driven system in terms of accessing the health record and use more of the deep functionality. But we've got data in Scotland going back about 20 years on about half a million people with diabetes. And that's obviously been a great resource for data mining and artificial intelligence, and starting to understand the trends within the data for people with diabetes.
And whenever we go into new environments, obviously, it takes a bit of time for us to make those connections into the IT systems to create that level of data integration. But that's part of what we do when we go to a region. We understand that the data is so important that the first task we do is understand the key IT systems in the area, put in new systems for the coalitions to start collecting their data in a more transparent way, and set up a system where the patients understand the functionality and can start to upload their data. And we're now doing that and other regions like England, the US, the Middle East, and we're starting some pilot work in the north of India as well.
As you look ahead, do you see this database becoming expanding beyond diabetes?
Absolutely. We're already doing some work in the at-risk group. These are people who don't yet have diabetes, but who are at risk of it. And they're almost as big a group as the people with diabetes. They need a slightly different approach. But actually, a lot of the fundamental principles are the same.
But we're also starting to look at other disease verticals like cardiovascular disease, hypertension on its own, kidney disease, and so on. [For] so many of these conditions, the same principles apply. It's just the messaging and the educational support that's slightly different. So we've already started working in these other disease verticals, and we're keen to kind of expand the work that we've done into these other areas.
People with diabetes often don't just come with one condition; often, they have other comorbidities. So it's really important for us to kind of take that whole personal approach and try and make sure that we're giving them appropriate advice--even if they've got, let's say lung disease or some other condition that is linked, but that requires some additional support.
For those who do not have access to such a database, what is your advice in terms of what they should or should not be doing--particularly if they're either borderline diabetic or diabetic, at the time of COVID-19?
One of the first things we do often when we go into new territories where we can set up a database straight away, is just put in some really good educational advice for people with diabetes. We make sure that there's good public access to good public health education. And we're actually going to be running some free online courses in India, for example, which we will advertise. And that'll be open to anyone with diabetes to come join us for a couple of days and learn more about what they can actually do to improve their outcomes.
But a lot of it is quite straightforward advice. A lot of people with diabetes are carrying too much weight. So it's important to get your weight down to what would be seen as the normal Body Mass Index. And maintaining that kind of weight is really important.
Activity and diet are key--so trying to stick within government guidelines for activity, making sure you're staying active and maintaining a good diet, trying to avoid high carbohydrate foods, sugary foods or starchy foods as much as you can, keep your calorie intake down.
It's also about turning up for your diabetes appointments, taking your diabetes medication, looking after your feet because a lot can go wrong with your feet if you've got diabetes, and that can lead to significant problems--so checking your feet every day, washing them, looking after them, making sure that they're well cared for. And, as I said, taking your medication appropriately so that things like blood pressure, cholesterol are well controlled.
A lot of people with type II diabetes now can actually reverse their condition. So there's a lot of evidence now emerging around diabetes remission. This is when people can push their diabetes back into remission at a point that their sugar levels are essentially normal, and they don't need any medication to control their diabetes. There's a really strong message for people if you've been fairly newly diagnosed with diabetes in the last few years. And if you really take this lifestyle advice seriously, lose the weight you need to lose, then there is a chance you can remiss your diabetes. And that puts you in a much lower risk group for any complications that might otherwise happen.
We welcome feedback. Please write to firstname.lastname@example.org. We reserve the right to edit responses for language and grammar.