AI

The length of your upper leg could help detect diabetes

A new home test can detect type 2 diabetes or prediabetes with high accuracy. The test has been developed using artificial intelligence and large amounts of health data – and surprisingly, one of the steps involves measuring the length of your thigh.

Seven measurements need to be entered into a new online test that identifies the risk of type 2 diabetes or prediabetes. Photo: Colourbox

Facts

Prediabetes is a condition in which blood sugar levels are higher than normal, but not high enough to be classified as type 2 diabetes. Pre-diabetes is also referred to as a precursor to type 2 diabetes. There are virtually no symptoms of pre-diabetes. In type 2 diabetes, symptoms may include:

  • Lack of energy – tiredness, nausea, poorer general health
  • Thirst
  • Frequent urination
  • Unexplained weight loss
  • Vision problems
  • Itching around the genitals
  • Infections of the skin and mucous membranes

AI as a tool

The new test is called MEDWACS. It was developed in collaboration between Daniel Yoo and DTU colleague Professor Olivier Jolliet, as well as the Italian doctor Umberto Maggiore, who is specialized in kidney diseases, which are frequent complications of diabetes. Their work has been published in the Journal of Clinical Epidemiology.

MEDWACS was developed using artificial intelligence and large amounts of US health data from the National Health and Nutrition Examination Survey, collected over 30 years for both sexes aged 18 and up.

The researchers have carried out two external validations of the test tool, testing it on data from two populations in the US and South Korea, where the number of people with and without the condition is known.

The validations show that MEDWACS is highly effective at detecting people with prediabetes or undiagnosed type 2 diabetes, performing on par with or exceeding other recognized testing methods.

"The new tool from DTU is an interesting contribution, and it is promising that it has been validated in several populations. In The Danish Diabetes Association, we use the Leicester Risk Assessment, which is a simple and well-tested method for rapid detection. It requires few measurements and is easy to use in the field. It has a precision that is on par with the new tool," says Tanja Thybo.

Measure your thigh

In MEDWACS – as in similar tests – you must answer questions such as age, gender and BMI. These questions are based on seven parameters that the research team selected from nearly 3,700 possible options.

All seven parameters can be measured at home using a tape measure, bathroom scales and a standard blood pressure monitor.

The most surprising parameter is the length of your thigh, which can help reveal your risk of type 2 diabetes or prediabetes. A parameter that initially surprised the research team as well. But Umberto Maggiore, the medical expert, can explain why this curious measurement is interesting in the context of diabetes:

“The length of the femur bears witness to nutrition in early childhood. Poor nutrition in the first years of life slightly inhibits bone growth and is closely linked to an increased risk of diabetes decades later. Furthermore, the body’s largest muscle group is also located in the thighs, and these muscles are responsible for removing sugar from the blood. Overall, shorter legs mean that a person has less muscle mass to absorb that sugar, which increases the risk of diabetes.”

The researchers assume that the thigh measurement is still relevant for Danes, even though it is a parameter based on health data from the US and South Korea.

“The length of the femur bears witness to nutrition in early childhood. Poor nutrition in the first years of life slightly inhibits bone growth and is closely linked to an increased risk of diabetes decades later."
Professor Umberto Maggiore University of Parma

Danes can also use the test

The researchers assume that the thigh measurement is still relevant for Danes, even though it is a parameter based on health data from the US and South Korea.

To be absolutely certain that the test is just as accurate at identifying high-risk patients in Denmark, it must be validated against health data from the Danish population, but the research team behind MEDWACS believes there is a good chance the tool will work here too.

“When MEDWACS performs so well on populations as diverse as the American and South Korean ones, we are confident that the test will also be able to detect Danes with diabetes or prediabetes. But of course, this is something we will need to test to be certain,” says Daniel Yoo.

At The Danish Diabetes Association, they follow the development of new tools closely and are always interested in new solutions, says Tanja Thybo:

"It will be interesting to see whether MEDWACS can be adapted and validated in a Danish population in the long term. There are still 100,000 Danes walking around with undiagnosed diabetes. That's why we are happy about all research that can make us better at finding them."

Facts

  • The MEDWACS test can assess your risk of having type 2 diabetes or prediabetes.
  • The test is designed for adults (aged 18 and over) who have not previously been diagnosed with diabetes.
  • The test has been validated using US and South Korean health data and performs better than similar tests.
  • You can access the test online, where you will be asked to provide your age, gender, BMI, waist circumference, systolic blood pressure, upper arm circumference and thigh length.
  • MEDWACS was developed by postdoc Daniel Yoo and Professor Olivier Jolliet, both from Section for Quantitative Sustainability Assessment at DTU Sustain, in collaboration with the Italian doctor and researcher Umberto Maggiore, University of Parma.
  • The research has been published in the Journal of Clinical Epidemiology.
  • The Novo Nordisk Foundation has funded the project.
  • MEDWACS has paved the way for new applications, and the research team has just put the finishing touches on a testing tool that can identify patients at high risk of kidney disease.

Try the test.

Contact

Daniel Yoo

Daniel Yoo Postdoc Department of Environmental and Resource Engineering

Olivier Jolliet

Olivier Jolliet Professor Department of Environmental and Resource Engineering Phone: +45 45251544