Apple Watch is a black box
Much has already been said about the use of the Apple Watch in medical situations. Some people owe their lives to early detection of all kinds of diseases and medical problems. But to use the Apple Watch in all kinds of medical studies, researchers need to know how the algorithms work and how it is determined whether someone suffers from an irregular heart rhythm, for example. The algorithms that Apple uses for this are not public. Researchers therefore notice that there are inconsistencies in the data they collect, probably because the algorithm has been adjusted in the meantime. But Apple does not provide any insight into this.
The data on heart rate variability, for example, appear to show quite anomalies, because Apple is constantly adjusting the algorithm. Measurements from one period therefore do not correspond to another period. Harvard researcher Jukka-Pekka Onnela addressed the problem:
These algorithms should be seen as black boxes – they are not transparent. So it's impossible to know what's in it.
‘Make more openness or raw data available’
This makes comparing data at different times very difficult. Apple changes the algorithms regularly and without warning. Large differences in measurements were also found in sleep analysis. There is a need for devices specially designed for scientific research, so that the researchers themselves have control and know that all measurements are always interpreted in the same way. Commercial products such as the Apple Watch seem very suitable at first glance, but are not. According to Onnela, Apple should make the raw data available.
The Apple Watch is regularly used for medical studies, including through ResearchKit. Many studies are only open to participants from the US.
In April, for example, Apple announced a partnership with the University of Washington to see if illnesses such as the flu and coronavirus can be predicted. Stanford University used the iPhone and Apple Watch to remotely determine if heart patients are showing signs of weakness. This study was even paid for by Apple.