Sickle cell disease: Apple Watch can predict pain
People affected by sickle cell disease experience bouts of pain that come and go seemingly at random. Researchers at the US universities Duke and Northwestern, together with other colleagues, have now demonstrated that the body data recorded by the Apple Watch sensors can be used to predict such flare-ups. This helps sufferers to prepare and better manage their condition. The hereditary disease, also known as sickle cell anemia, affects the red blood cells.
Apple Watch data plus machine learning
In the study, the scientists looked at the development of so-called vaso-occlusive crises (VOCs, also known as sickle cell crises), which often lead to hospital stays with the administration of painkillers. The study, which collected nearly 16,000 data points on the Apple Watch Series 3 between July and September 2021, stored activity data, heart rate variability (HRV) and heart rate to see if they were already suitable as indicators.
The watch data was then combined with patient pain ratings and other vital signs from the clinic. A prediction model was then calculated using machine learning (ML). The accuracy turned out to be good – at best it was at least 85 percent.
predicted pain scores
According to the research, the data of the watch is quite suitable for the purpose. ‘The model’s strong performance in all measures confirms the feasibility of using data collected from a non-invasive device to predict pain scores during VOCs,’ the researchers note. “It’s a novel and viable approach, and the Apple Watch represents a cost-effective way that clinicians and people with sickle cell disease could benefit from managing VOCs.”
At the moment it is still unclear whether the technology can also be implemented in practice. For an officially offered product, the accuracy would have to increase even further. However, the example shows that the data that an Apple Watch can deliver can be used for completely new purposes with clever ML models.