Medical Wearables – Quest for the grail.

The pilgrims are still far away from the holy land.

• Much is being made for the potential for the Galaxy Watch 4 and the Apple Watch 7 to measure blood glucose but their inability to conquer the easier task of blood pressure highlights just how hard these goals are to reach.
• The realistic outlook is that these devices will provide very vague wellness and fitness indicators at best.
• Two of the biggest lifestyle-related conditions that cause ill-health and death today are hypertension (high blood pressure) and Diabetes (inability to control blood sugar levels).
• Furthermore, the monitoring of these conditions is invasive and uncomfortable in both cases making constant monitoring over a long period of time impractical.
• Constant monitoring results in by far the best outcome for the patient rather than monitoring at specific times of the day as a much more accurate picture of how the patient responds to his or her environment is revealed.
• Hence, a non-invasive system for measuring blood pressure and blood glucose to a medical standard is a holy grail that will reward both the inventor and the health of those that suffer from these conditions.
• Many claims have been made, but as of yet, none have made the grade.
• For example, the gold standard of blood pressure is a measurement that repeatedly gets within 8mm of mercury (Hg) when compared to the benchmark.
• The benchmark is a trained physician with a stethoscope listening for the characteristic sound when the vein collapses under the pressure from the cuff.
• Not even the digitised cuffs sold by Omron and others are as good as this but for all intents and purposes, they are good enough.
• No one is reliably close to this for a number of reasons:
  • First, Correlation: Most companies are trying to correlate fluctuations in green light absorption by capillaries with blood pressure and blood glucose levels.
  • This is easy to do for pulse where absorption is a direct consequence of the heartbeat but the other two are degrees of freedom away.
  • This makes correlation very difficult and it has yet to be achieved to a consistent and reproducible medical grade.
  • Second, Biological variation: Humans are different from one another and respond to stimuli differently.
  • This means that the absorption of green light and how that correlates to both blood pressure and blood glucose varies substantially from one person to the next.
  • This complicates the situation significantly although some progress has been made in using machine learning to correlate factors such as age, weight and sex to green light absorption and blood pressure.
  • This has been affected with some success by Valencell, but it is still very far from what could be considered as a solution to using cuffs.
  • Samsung’s offering requires calibration using an arm cuff every week or so as the accuracy of its measurement deteriorates quickly with time.
  • Furthermore, none of these measurements makes the grade when it comes to offering a useful alternative to the current invasive techniques.
  • Third, wrist: From the user’s perspective, the wrist is the most convenient place to affect this monitoring.
  • This is because it causes the least inconvenience, as well as users, are quite used to wearing devices in this location for long periods of time, if not constantly.
  • However, compared to the outer ear or the fingertip, this is a suboptimal place to monitor given the extra signal noise created by external hairs (men mostly), ligaments, tendons and other internal characteristics that are in constant motion.
  • Furthermore, arm movements also change the characteristics which make separating the signal from the noise even more challenging.
  • Hence, I suspect that the ear with a sensor integrated into inner ear headphones could be one of the first to produce a really good prototype.
  • Fourth, data mining: When one is looking for signal in a large amount of noise, data mining becomes a potential problem.
  • This issue refers to the outcome where one keeps on digging and eventually one ends up with a completely spurious relationship because one has been digging too long.
  • There is a lot of this going on by those seeking to find correlations in the data they are generating and consequently, I suspect that there will be a number of spurious relationships generated.
• The net result is that the medical device companies are not going to go out of business in the near future.
• However, they should make hay while the sun shines because sooner or later, these issues are going to be solved and then the business of charging vast amounts of money for medical equipment will come crashing to a halt.
• This sector has a bright future, but I still have no clear idea of when the sun is going to poke its head above the horizon.