Researchers Blame ‘Black Boxes’ for Unreliable Apple Watch Health Data

by -64 views

Apple Watch is known for its health-centric features. Over the past years, there have been several instances wherein Apple Scout helped diagnose impending health conditions. Newly added features like ECG, fall detection, oxygen sensor, and exercise tracking have aided health studies. Researchers from Harvard and the University of Michigan have at present raised red flags almost relying on Apple tree Watch for health studies.

Researchers are concerned about how Apple Watch algorithms for studies create “black boxes.” JP Onnela, associate professor of biostatistics, highlights the issue with using Apple tree Watch for health studies. The professor has noticed a example wherein heart charge per unit variability data sourced from Apple Scout is inconsistent. Furthermore, the unreliable information tin also affect the findings of other health studies based on the Apple tree Lookout man.

Typically Apple updates wellness algorithms regularly. Due to the algorithm changes, the data is jump to alter. In some cases, the changes could affect the entire health written report. In other words, “the data from the aforementioned time period tin alter without alarm.”

These algorithms are what nosotros would phone call black boxes — they’re non transparent. So information technology’southward incommunicable to know what’s in them,” JP Onnela, acquaintance professor of biostatistics at the Harvard T.H. Chan School of Public Health and programmer of the open up-source information platform Beiwe, told The Verge.

Devices like Apple Watch are aimed at consumers and not exactly a neat fit for enquiry. This is the reason Onnela uses research-grade devices that supply raw output data. On the reverse, the data obtained from Apple Watch is being processed through algorithmic filters and is likely to vary significantly.

So, they checked in on heart rate data his collaborator Hassan Dawood, a enquiry boyfriend at Brigham and Women’s Hospital, exported from his Apple tree Watch. Dawood exported his daily heart rate variability information twice: once on September 5th, 2022 and a 2nd fourth dimension on April 15th, 2022. For the experiment, they looked at data collected over the same period — from early December 2022 to September 2022.

Unacceptable Variances in Apple Lookout Data

Onnela used Apple Watch in one of the studies to highlight differences in the same data exported at two instances. The variances showed huge differences and an unimpressive linear correlation of 0.67.

To be clear, these data comprehend the same appointment range, so they should be identical. In fact, their means are very similar, 52 vs. 55 for the get-go and second export, respectively, simply their variances are very different: 1240 vs. 572. To get some further insight into this, I made a scatter plot of the values of one fourth dimension series against the other. The dashed identity line is where nosotros’d like to see the points fall if they were identical, every bit we’d promise. Instead, there’s a lot of scatter in the information, and their Pearson linear correlation coefficient is just 0.67. That’s not a very high correlation.

The researcher explains how Apple Watch information could help those who want to track their health. When information technology comes to inquiry, the differences are so huge that it is unacceptable. There is a big question marking on reliability, especially while studying participants wearing smartwatches from different brands. Academy of Michigan sleep researcher Olivia Walch agreed with Onnela.

Constantly changing algorithms makes it almost prohibitively hard to use commercial wearables for sleep research, Walch says. Sleep studies are already expensive. “Are you lot going to be able to strap iv FitBits on someone, each running a dissimilar version of the software, and so compare them? Probably not.

The report underlines take a chance of using the Apple Watch or other wearables, especially for research purposes. Apple is nonetheless to comment on this matter.

[via TheVerge]


Posted by: