A recent study by researchers at the Duke University (North Carolina, USA) showed that the combination of data collected by a wearable (like the Apple Watch) and techniques of machine learning can help predict pain levels in people with sickle cell disease (sickle cell disease).
Caused by a genetic alteration of red blood cells, this disease — the most common being sickle cell anemia — is associated with serious complications such as chronic anemia, stroke, and especially vaso-occlusive crises (CVOs) — which are the main cause of hospitalization of patients due to their great unpredictability and difficulty of treatment.
Aiming to determine the feasibility of using the Apple Watch to predict the level of pain in people with the disease, the researchers gave the watch to patients who were admitted to the university hospital due to episodes of VOCs. Used during the visit, the device helped to collect data such as heart rate level and variability and calories — as detailed in the MyHealthyApple.
At the same time, researchers collected patients’ pain levels and vital signs from the electronic medical record to cross-reference the data, which were analyzed using three models of machine learning (multinomial logistic regression, gradient boosting e random forest) and two null models.
With both models overcoming nulls — including, with that of random forest achieving 84.5% accuracy — it was concluded that it is possible to use data collected from a non-invasive device such as the Apple Watch to predict pain levels during CVOs.
This is an innovative and low-cost method that has the potential to benefit both physicians and people suffering from sickle cell disease, as it would be useful to avoid possible future hospitalizations due to vaso-occlusive crises.
Apple Watch Ultra
Apple Watch Series 8
Apple Watch SE
TRANSPARENCY NOTE: The MacMagazine You receive a small commission from sales completed through links in this post, but you, as a consumer, pay nothing more for products you purchase through our affiliate links.
via 9to5Mac
