- Researchers recently created an artificial intelligence model that predicts the onset of diabetes with 12 hours of blood sugar data collected from a wearable device.
- They say their model could aid in the diagnosis of prediabetes and help prevent type 2 diabetes.
- The impact of the AI model on type 2 diabetes screening rates remains to be seen.
Diabetes is a chronic disease
Approximately 463 million people worldwide are diagnosed with diabetes;
Prediabetes, or “intermediate hyperglycemia,” is the high-risk stage before type 2 diabetes, when blood sugar levels are above average but below the diabetes threshold.
Around
Diagnoses of diabetes and prediabetes usually involve blood tests, including the A1C test, a measurement of a person’s average blood sugar over the past 3 months, a fasting blood sugar test, a glucose tolerance test, or a random blood glucose test.
New ways to screen for prediabetes and type 2 diabetes may encourage people to get tested.
In a new study, researchers investigated whether they could use continuous glucose meter (CGM) readings to diagnose prediabetes and diabetes. With just 12 hours of glycemic profile data, the researchers were able to categorize type 2 diabetes, prediabetes, and people without impaired glucose tolerance.
Jouhyun Clare Jeon, Ph.D., principal investigator at Klick Applied Sciences and lead author of the study, said Medical News Today:
“I believe our method has a lot of potential to be used as a new tool to help healthcare providers in their own decision-making, especially for remote or virtual patient care. For the general public, our method could not only be used for monitoring and early detection but alert a patient of their risk of developing diabetes.
The results were recently presented at the 36th Conference on Neural Information Processing Systems (NeurIPS) in New Orleans, LA.
Continuous glucose monitors (CGMs) are portable devices that measure blood glucose every 15 minutes.
CGMs help people with diabetes to regularly monitor their blood sugar.
“Continuous blood glucose monitors (CGMs) are gaining traction for wear in the general non-diabetic population for health or other specific purposes,” said Dr. William Dixon, clinical assistant professor of emergency medicine at the Stanford University and co-founder of Signos, not involved in the study, said DTM.
Dr Dixon added that determining the presence and risk of diabetes based on CGM data can be useful for people who are not routinely tested or screened for the disease.
“There are also signs of impaired glucose tolerance that may be apparent even before average glucose levels reach a range of concern,” Dr. Dixon said.
For the study, the researchers used data from 436 Indian participants.
Each participant wore a CGM device for an average of 12 days and provided data such as gender, age and body mass index (BMI).
The researchers defined participants’ A1C levels of 6.5% and above as type 2 diabetes, 5.5% to 6.5% as prediabetic, and less than 5.5% as healthy.
Of the participants, 172 had type 2 diabetes, 87 had prediabetes and 177 were healthy. The diagnoses were confirmed by doctors.
The researchers created AI prediction models based on different blood glucose durations. They compared models based on data windows of 12, 24, 72, 168 and 288 hours.
They found that CGM data was 1.21, 1.34 and 1.17 times more accurate than demographic data in identifying type 2 diabetes, prediabetes and healthy individuals.
They also found that their 12-hour model was just as effective as longer-lasting models.
After optimizing the 12-hour model, they identified 87%, 84%, and 86% of people with diabetes or prediabetes and healthy people.
The researchers noted that of those in the 12-hour prediction, 23 were misclassified due to unusual 12-hour readings in which they reported the same blood sugar levels over time.
The researchers concluded that CGM systems could allow rapid and accurate screening of diabetes outcomes.
The researchers hope to conduct similar studies on larger cohorts to improve their prediction models.
Asked about the limitations of the study, Dr. Jeon said DTM:
“Our results are developed based on CGM signals from approximately 400 patients. Further evaluation is needed using a larger independent cohort and larger population data to generalize our method. However, we are encouraged by the results. and look forward to our continued work in this area.
Dr. Jeon noted that with Klick’s predictive diagnostics, people could find out their results from home instead of going to a clinic for blood tests and waiting a few days.
Michael Lieberman, Ph.D., general manager of research and development at Klick Applied Sciences, also said DTM:
“From a public health perspective, prediabetes is extremely underdiagnosed. The ability to easily determine with a high degree of probability that a person is prediabetic without a doctor’s visit could be extremely useful in identifying those at risk of becoming diabetic.
Dr. Lieberman added that detecting prediabetes early could give a person’s healthcare team enough time to reverse the course of the disease before it’s too late.
Dr. John Miles, a University of Kansas health system endocrinologist not involved in the study, noted that the practical implications of this study are relatively modest. He noted at DTM:
“I’m not sure we can say at this point that continuous glucose monitoring (CGM), as performed in this study, is an improvement over existing methods of diagnosing diabetes. It is certainly true that some people are unaware that they have diabetes or prediabetes.However, the fact that the CGM can accurately define the category that people are in (diabetes, prediabetes or non-diabetic) as defined by the hemoglobin A1c does not mean it would be a practical alternative to [A1c testing].”
“[A1c testing] the use of a fingerstick blood sample is currently used to screen for diabetes and would be simpler, faster and probably cheaper than CGM in a mass screening programme,” Dr Miles concluded.
#Type #diabetes #model #predicts #onset #hours