With the help of the identification, the scientists from the Weizmann Institute hope, it will be possible to prevent the development of gestational diabetes through changes in diet or lifestyle
A computer algorithm will be able to identify in advance, that is, already at the beginning of pregnancy or even before it begins, which women are at high risk of developing gestational diabetes - this is according to a study by Weizmann Institute of Science scientists which was published today in the scientific journal Nature Medicine. With the help of the identification, the scientists hope, it will be possible to prevent the development of gestational diabetes through changes in diet or lifestyle.
For the purpose of the study, the scientists analyzed data from the largest health organization in Israel - Klalit Health Services - on almost 600 pregnancies. "Our goal was to help the health system prevent cases of gestational diabetes," says Prof. Eran Segal, whose laboratory in the Department of Computer Science and Applied Mathematics and the Department of Molecular Cell Biology led the research.
Gestational diabetes is characterized by high blood sugar levels during pregnancy without a history of diabetes. It appears in 3%-9% of pregnancies and involves significant health risks for the mother and the fetus. Today, gestational diabetes is usually diagnosed between week 24 and week 28 through a sugar loading test: drinking a sugar extract and then a blood test that reveals how quickly sugar is removed from the blood.
In the new study, Prof. Segal and his colleagues first analyzed, with the help of machine learning methods, data from the general databases of about 450 thousand pregnancies between 2017-2010; In about 4% of these pregnancies, gestational diabetes was diagnosed through a sugar loading test. After the algorithm developed by the institute's scientists analyzed a huge amount of data - more than 2,000 characteristics for each pregnancy, including the results of the pregnant woman's blood tests and data on her and her family's medical history - it was discovered that nine of these characteristics are sufficient to accurately identify which women are at high risk of diabetes pregnant The nine parameters included, among others, the woman's age, her body mass index (BMI), family history of diabetes and the results of blood sugar tests from previous pregnancies (if there were any).
In the second phase of the study, the scientists wanted to make sure that the nine parameters did indeed successfully predict the risk of gestational diabetes. To this end, they examined 140 pregnancies that were not included in the original phase of the study. The results confirmed: the nine parameters successfully identified the women who developed gestational diabetes.
The findings show that using a short questionnaire it will be possible to determine in advance if the woman may develop gestational diabetes. Early detection of the risk of developing gestational diabetes, at the beginning of pregnancy or even before the woman became pregnant, will allow pregnant women to reduce the risk of gestational diabetes through exercise or a special diet. Also, the questionnaire may save women who are at low risk of developing gestational diabetes, the trouble and the costs associated with a sugar loading test.
On a more general level, this study shows how the analysis of "big data" - in this case, the analysis of medical records - can lead to personalized recommendations for treating or preventing diseases.
Similarly, Prof. Segal's laboratory initiated the 10K project - a long-term observational study to collect information on lifestyles and diseases in the Israeli population that combines innovative medical tests and advanced artificial intelligence methods for the purpose of predicting future medical conditions. The ambition is to develop methods that will make it possible to prevent these situations, improve the quality of life and extend their lifespan. These days, the project is recruiting participants.
The research was led by research students Nitzan Shalom Artzi, Dr. Samdar Shiloh and Hagi Rosman from Prof. Segal's laboratory, who collaborated with Prof. Eran Hadar, Dr. Shiri Barbash-Hazan, Prof. Avi Ben-Harosh and Prof. Arnon Viz Nitzer from the Rabin Medical Center in Petah Tikva, and with Prof. Ran Blitzer and Dr. Beka Feldman from Klalit Health Services.
