(ThyBlackMan.com) Smart wearables are good as long as they keep you updated with accurate and realistic data. But what if one day you wake up and your Apple Watch tells you that you have diabetes? With the help of machine learning, people have been predicting market sentiments and the rate of cash flow, but now when more than 8 percent Americans own an Apple Watch, companies are ready to turn your smartwatch into a fortune teller.
Apple Watch comes in different models and each model has different capabilities to keep you fit and healthy by providing different metrics including heart rate, step count, calories etc. Now a company called Cardiogram is working with the University of California San Francisco and claiming that machines can easily read signs of diabetes by just analyzing heartbeat patterns and step count. It means that if this technology is implemented into a fitness tracker like Apple Watch, it can detect diabetes in your body.
The Science Behind This Tech
This is very surprising because we all know that to detect diabetes, one has to check sugar level in blood. But here, the team of researchers is just programming a machine so that it can learn to read patterns, match them with the heartbeat patterns of other diabetic people and give you a health report. Can you trust such indications given by your smartwatch?
The researchers claim that a deep neural network called Deep Heart can easily distinguish between diabetic and non-diabetic people. The accuracy rate of this network is claimed to be 85 percent. It’s really impressive that our engineers are working hard to make sure that the technology can keep people alert about their health conditions, but the question remains the same, can you trust such report produced by a smartwatch? Even today, we haven’t achieved accuracy in counting steps, heart rate, and calories.
The Cardiogram team asked around 14000 Apple Watch and Android Wear enthusiasts to be a part of the study. The data collected by the team was used to teach the neural network about these patterns. This research is based on several studies that suggest that when diabetes starts developing in people, their heart rate variability pattern changes.
In simple words, if such variability is detected, your Apple Watch will inform you about it and you might be visiting a doctor soon. Cardiogram doesn’t explicitly predicts diabetes, but it gathers data and can inform users after reading the patterns. The team also said that using the Apple Watch, they have detected arrhythmia with 97 percent accuracy. The company in fact successfully detected sleep apnea and hypertension with 90 percent and 82 percent accuracy, respectively.
The big questions
This is undoubtedly a great step forward but there are some questions that need to be answered first. What would be the ratio of accurate and false positive/negative results? If there’s one thing that bugs us the most right now, it’s advertising campaigns and data privacy. What will happen if such data is compromised, hacked or sold? Many insurance companies and industries are just waiting to irritate you with their so-called personalized advertisements, and all they need is your data so that they can scare you and sell insurance policies.
It is really good that soon these fitness trackers will get more features and it will be very easy for people to learn more about their health, but this experiment, particularly the diabetes detection through heart rate and step count, sounds very impractical. Physically, we humans are different from each other in many ways. Genes, lifestyle, eating and drinking habits, activity, fitness level, and medication play a major role and we can’t just compare our body’s capabilities with someone else and come to a point where we can start relying on such devices.
Imagine someday you get a notification on your Apple Watch saying symptoms of diabetes are developed in your body, you carry the thoughts in your mind until you visit the doctor and get a negative report. Prevention is always better than cure, but why can’t we just schedule a monthly or quarterly appointment with our doctor and get a full body check up with accurate reporting?
Staff Writer; Corey Shaw
Have any Tech Tips? News? Hit up our Tech Guru at; CoreyS@ThyBlackMan.com
