How do they claim this works?
In their own words
The History and Technology Behind Healbe GoBe:
Algorithm owes its success in part to the active use of Inventive Problem Solving Theory (TRIZ) in its innovative design process. All researchers working with Algorithm are professionals in TRIZ application. Moreover, the personnel of Algorithm and GEN3 Partners currently include 9 TRIZ experts.
That actually explains a lot.
Alas it does not explain how anything might work in the product.
Healbe applies ground-breaking technology to automatic
calorie intake measurement
Various attempts to measure glucose concentration through noninvasive methods have been covered in thousands of papers and articles. There are hundreds of projects devoted to this topic.
During our time with Algorithm, we combined the advantages of several project groups from various disciplines. This allowed us to integrate into a single whole a keen understanding of such physiological process as glycometabolism, along with the most informative and accessible measurements of health parameters. That is how our water-balance measurement technology was born. This method produced consistent results in the course of testing.
Instead of entering the market with a noninvasive blood glucose meter, Healbe chose to use this technology to develop a personal device for monitoring energy intake, consumption, and balance, which is especially important in the process of weight management.
So they do have a technology that claims to measure blood glucose concentration and that while being non-invasive. No more blood pricking. Yay. Quite an achievement! That alone should be a godsend to all those insulin users out there.
Seriously, very seriously: Why not bring that fantastic technology to the market as soon as the patent is granted?
That tech is needed! (And others already pointed that out.)
A Series of Doubts:
Amidst all the PR in the above document we find the first part of the answer:
The accuracy rate of glucose concentration calculating methods that this hardware and software system’s computing model is based on (15-20%) are quite enough, in my point of view, for implementation of a device
intended for day-to-day applications.
A nice opinion, but at least they concluded sensibly:
This is one of the basic technologies that we have developed. GoBe does not display glucose concentration to the user, as our device is not intended for medical use, and doesn’t meet medical accuracy requirements.
Admitting that the method for their device is too inaccurate for medical use means it is just unreliable overall for measuring blood glucose.
Much worse than the above is what this means for the the accuracy of the device measuring the calorie intake non invasively:
The current dietary guidelines for the US say that you should eat a varied diet. That is low on added sugar and including fibre, fat and protein. If you eat adequate amounts of carbohydrates the intake is usually not converted into blood glucose, which seems to be the only parameter measured. But these ingredients or meal constituents have a caloric impact!.
Unless they demonstrate in an independent study that different diets are estimated by such a device only halfway convincingly, that is high carb as well low carb for example: this is very likely complete bogus.
A niche application might be: for kids on the night of Halloween that only eat hard candy (chocolate has way too much fat content) this might work within the error range of 15% off.
From a review:
For one week, I wore the band every day and simultaneously logged every food I’ve eaten in an old-school calorie-counting app called Lose It. The results were startling: The GoBe was within 10 percent of my calorie tracking estimates every single day of the seven straight days tested, and the average deviation was under 6 percent.
It’s worth noting that manual calorie counting isn’t perfect either: When I enter five strips of bacon into the app and it tells me how many calories that food is “supposed” to have, it’s a guess. Does every strip really have exactly 46 calories? Who knows? So GoBe’s deviation from manual calorie counting doesn’t necessarily reflect its accuracy, just how far it is from another flawed metric, which happens to be the next best option for regular folks.
GoBe deviated further from this method when it came to the exact grams of protein, fat, and carbohydrates I ate each day, to the tune of 20 to 25 percent on average. For carbs and protein, GoBe typically thought I ate more than Lose It, but for fat, GoBe thought I ate less.
Are they nuts?
Touting these claims might gt you into legal trouble. How do they prepare for that:
Why is there a significant difference between the calories I manually calculated and the calories GoBe automatically tracked?
It is important to remember that digestion is personal. In fact, some of the food you consume is not digested by your body at all. This largely depends on personal specifics
I understand that my body’s glucose levels vary based on what proportion of carbs, protein, and fat I eat. If GoBe uses glucose levels to measure calorie intake, will the accuracy be affected by whether I eat a lot of carbs, protein, or fat?
GoBe accurately measure calorie intake from carbs, proteins, and fats. It uses FLOW Technology to analyze changes in your body’s glucose concentration–not by measuring absolute glucose levels. As glucose concentrations rise, cells absorb glucose and release water. GoBe uses an impedance sensor that sends high and low frequency signals through your tissue to measure the fluid moving in and out of your cells. FLOW Technology uses an advanced algorithm to analyze this data and determine calorie intake.
They have tech that is too imprecise to use as a medical device and apply some Big Data calculations to refine these very unreliable measurements. If results are off, it's clearly your own fault. They say so.
At least they are now wise enough to not answer curious emails from skeptical inquirers, any more. Because they apparently did one time too many:
For example: I asked him how GoBe accounts for the fact that so many calories we consume are locked up in fats and proteins. “Protein is more like ‘construction material’ for cells,” Shipitsin replied in an email, “and without diets, contribution of protein depend on climate zone and average for U.S. is near 15 to 20 percent... If user using special nutrition like ‘low-carb’ diets, he or she must set up accordingly flags in app, GoBe change the algorithm accordingly.”
Wait, what? What’s all this talk about construction materials and climate zones? Let’s concede there may be a language barrier problem in Shipitsin’s email, and continue on.
I also challenged Shipitsin to respond to Robinson’s observation that the relationship between glucose intake and cellular glucose isn’t proportional. To this Shipitsin replied, “Yes, of course, your doctor’s absolutely right. You see the main idea—we have exactly ‘physiological technology.’ GoBe is following the process inside the body and we absolutely aligned with doctors knowledge. We preparing content to Indiegogo’s page with description of technology and internal tests of accuracy. I guess you understand, not only consumers asking us about technology.”
Even in it's current iteration there are systemic and structural limitations of how this might work? As a very large error prone estimate it is a nice idea. Trusting this device is foolish.
With more sensors, more accurate sensors, to be precise, this might go somewhere. Why not just add an accelerometer to measure your food intake?
A New Method for Measuring Meal Intake in Humans via Automated Wrist Motion Tracking Two tests of the accuracy of our device in counting bites found that our method has 94 % sensitivity in a controlled meal setting and 86 % sensitivity in an uncontrolled meal setting, with one false positive per every 5 bites in both settings. Preliminary data from daily living indicates that bites measured by the device are positively related to caloric intake illustrating the potential of the device to monitor energy intake.
Simply estimating the caloric intake seems, for example by just counting bites and swallows, no less precise than these products. But wait a minute there is more:
Scaling up Dietary Data for Decision-Making in Low-Income Countries: New Technological Frontiers (Advances in Nutrition, 2017):
The GoBe wristband, which needed to be tapped before eating to activate it, used an impedance sensor, an accelerometer, and a pressure sensor.
So they do use an accelerometer! Way to go to measure what a friendly person holds to your mouth and you get to swallow with a straw.
Unfortunately, the reliability of these technologies are not really assessed:
Third, an assessment of the established validity of these technologies would have been useful but was outside the scope of this review because of space limitations.
In summary, although none of the technologies reviewed was immediately fit for “prime time” application in large- scale 24HDR surveys in LICs, the assessment criteria suggested that the computer- and tablet-based programs came closest to fulfilling the criteria set forth by this study.