New Delhi, Aug 11 : A group of Indian scientists has developed an artificial intelligence (AI) algorithm that is derived from characteristics of individual heartbeats that are recorded in an ECG (electrocardiogram) and is able to accurately predict the presence of the presence of diabetes and pre-diabetes.The team from the Lata Medical Research Foundation in Nagpur included clinical data from 1,262 people.
A regular 12-lead ECG heart trace lasting 10 seconds was recorded for each of the participants.A total of 100 unique functional and structural features on each lead were added to each of the 10461 single heartbeats recorded in order to produce an algorithm that could predict heartbeats.The algorithm was named DiaBeats.
Based on the size and shape of individual heartbeats Based on the shape and size of individual heartbeats, the DiaBeats algorithm was able to quickly detect prediabetes and diabetes with the overall accuracy of 97% and an accuracy of 97 regardless of the influence of other factors, like age, gender and the presence of co-existing metabolic conditions.
Important ECG characteristics are consistently in line with the known biological triggers that are responsible for the heart changes typical of pre-diabetes and diabetes.
If it is validated in bigger studies, this method could be employed to detect the illness in low resource settings, the group stated.
“In the theory of things, our study offers a relatively low-cost non-invasive, accurate, and non-invasive alternative (to current diagnostic methods) that could be utilized as a gatekeeper to effectively detect pre-diabetes and diabetes earlier in its development.”
“Nevertheless the adoption of this algorithm into everyday practice will require rigorous validation using independent, external databases,” they cautioned.
A staggering 463 million adults across the globe had diabetes in the year 2019.Being able to recognize the illness in its early stages is crucial to avoiding serious health problems, however, diagnosis is largely based on the assessment of blood sugar.
It’s not just inconvenient but also difficult to implement as an massive screening test in low resources, as pointed out by the authors in the paper that was published in the journal online BMJ Innovations.
Functional and structural modifications in the cardiovascular system take place at an early stage, before even indicating changes in blood glucose and are visible in an ECG heart trace.
Researchers also noted that participants all had risk of metabolic diseases, making them less likely to be representative of the population as a whole.Also, DiaBeats is slightly less reliable in those who were taking prescription medications for high blood pressure, diabetes and high cholesterol, for example.
There was no data on those who were pre-diabetic or diabetic.This made it impossible to know the effect of the screening process early.
rvt/pgh
