Most Popular

    Sorry. No data so far.


Winner of Amgen Patients | Choices | Empowerment Competition Emerging Star of HealthCare Engagement Award
Mayo Clinic Award - LeftA winner of the Mayo Clinic iSpot Competition for Ideas that will Transform HealthcareMayo Clinic Award - R

I, Medicine: Predictive Biometrics and Health Optimization

bardThis is a guest post from Bard C., a CureTogether member from Australia and self-tracker who has made some interesting discoveries about his own health. Thanks Bard for sharing your thoughts, perspective, and a link to your own self-tracking system!

——

The most important online applications to be developed since the birth of the Internet are less than a decade away, yet they remain largely off the radar of investors and web-developers. These applications are optimizers for the most important thing we have – our health.

Here’s why…

A couple of years ago IBM released a report that stated that by 2011 the total amount of data in the world may be doubling every eleven hours. I believe that this prediction will be more than met by the surge in self-tracking devices that will be released over the coming years.

The spread of ubiquitous micro-trackers will make health data collection cheaper and easier. The self-tracking hardware and software will no doubt be unified by cloud computing and result in a ‘perfect storm’ of accessible and highly effective health optimizers.

It’s an important convergence of technology and timing.

Imagine if the computer industry was nothing but repairmen that tried to salvage PCs in the last stages of severe virus infections. It would be a pretty bleak situation, yet that is the current state of the health-care system. For the most part, people are only treated once they have become sick and even then with largely disappointing results.
The computer industry, on the other hand, has an endless supply of optimization tools that improve efficiency and user experience long before the computer has become ‘sick’. Defraggers, registry cleaners and driver updaters are just some of the tools in use.
I believe that the coming decade will bring with it a surge in online self-optimization applications. I also believe that the results for people’s health will go far beyond current expectations.

My personal story is that I have an hereditary immunological condition that I have struggled with for some years. Like many people that I have met, I felt dissatisfied with the health-care system and decided to take matters into my own hands to try and improve my health.

I did manage to track down the cause of my health problems and have since made great progress in my treatment. At the same time I developed my own self-quantification system that allowed me greater insight into how the medications and lifestyle choices I was taking were affecting my health.

bard spreadsheet1
What fascinated me was that when I entered one month’s worth of self-tracking data into my system it reached the same result that had taken me over a year to find through old-fashioned trial and error. This translated to a 20-fold improvement in efficiency.
I did feel disappointment that I hadn’t developed this tool years ago but also became excited about what discoveries it would lead me to next. I had become acutely aware of the potential of self-quantification tools.
Our very experience of life is filtered through our current state of physical and mental health. Having less than optimal health fundamentally affects the quality of our lives.
Right now our health-care system is a largely disorganized, anemic beast that advances with painful slowness and inefficiency.
Thankfully, as medicine is gradually transformed into an information technology, we will see an exponential acceleration in advancement.
My concept is to take it one step further and open up our medical laboratories to encompass the great experiment that is the human race. Every day billions of people test out drugs and treatments in their own homes, but this data is largely lost, or relegated to unreliable analogies. I am committed to changing that.
Humans are quite bad at deciphering patterns over time. We can be good at recognizing patterns over short periods but over days or weeks our perception becomes extremely poor. We often miss obvious connections with our health while simultaneously making false positives. For example, someone might swear that a placebo treatment such as homeopathy cured their hay-fever while completely ignoring the fact that three days of rain had washed away the pollen from the air.
A personal biometrics system sees through all the noise and is unswayed by personal biases or wish-thinking. It also has the potential to negate or eliminate the placebo effect from subjective results but this is a deeper area that requires some clever mathematics and development.
The problem with a condition like mine (and conditions such as chronic pain, chronic fatigue, fybromyalgia, migraine, and so on) is that they usually involve multiple organs and can be extremely difficult to diagnose and treat effectively. They can affect the homeostasis of the body’s nervous system leading to debilitating allostasis. I have my own theories about the mechanisms behind many of these conditions, but one man’s opinion is merely conjecture and insignificant in the grand scheme of things.
What is significant are measurable observations and results about people’s health gained from multiple data points. Yes, it is true that the self-quantification concept revolves around gathering self-reporting data, which can be notoriously unreliable from a medical research stand-point, but the era of the Internet brings with it a new paradigm – truly massive numbers. Massive numbers of individuals from which to gather data, and, even more excitingly, massive amounts of data gathered on each individual.
As we can see with sites like Curetogether.com, the results we obtain can be surprisingly accurate and comparable to those obtained through more traditional means.
I believe that in the next decade people will become increasingly comfortable with the idea of self-quantification. In fact, I believe that this paradigm-shift is already happening around us. If you don’t believe me, if you think that people will not be comfortable giving up so much personal information into the cloud, then I remind you that the same contention was raised about Gmail when it was first released.
The truth is, even though it sounds dreadful to have an email service that scans your emails for personalized keyword advertising, in practice it turns out to be a small price to pay for a fantastic free service. I believe the self-quantification market will grow in very much the same way. After all, what better service can you provide than the gift of health?

I am convinced that in a decade or so we will see the refinement and clarification of our current medical knowledge as well as whole new discoveries and new avenues of research opened up by the power of ubiquitous tracking hardware and clever mathematics. Our job as developers and engineers will be to dig up the gold in the mountains of data.

The hardware is well on the way to becoming reality in our lives. From the Fit-Bit device to health-monitoring toilets (as we saw in a final-round proposal for Google’s 10 to 100 project). Now all we need is a comprehensive, well-integrated, robust, cloud-based and highly scalable software solution to take advantage of this opportunity.
That is where my interest lies, and I’m committed to collaborating with others who are heading in the same direction. I feel it has the potential to improve the health of millions, if not billions of people around the world.
By Bard C.

I have been requested to make my personal system available online for others to experience and test out. I have made a demo with some sample data which can be viewed at
http://spreadsheets.google.com/ccc?key=0AoTcUIoZTAbZdHZEdWJOTTFNZ3VybVc1cGNkRllvSWc&hl=en

The form which feeds the data into this spreadsheet can be tested at
http://spreadsheets.google.com/viewform?formkey=dHZEdWJOTTFNZ3VybVc1cGNkRllvSWc6MA

It’s a simple Google docs spreadsheet linked to a live form. bard spreadsheet 3

I apologize for the intense mathematical format of the spreadsheet, but I have not had time yet to develop a proper, simpler user-interface for it. Ultimately, I believe it will have a very slick, minimalist natural-language interface that will hide most of the complicated math.

The correlation table uses the basic correlative coefficient to calculate the strength of the relationship between two series of data. The relationship is either positive (the values rise together, eg: weight gain and chocolate consumed) or negative (the values head in opposite directions, eg: weight gained and minutes exercise).

I use text coloring rules based on the strength of the correlations to draw the user’s attention towards the most important relationships. For example, in my data I quickly found that there was a strong negative correlation between my nasal congestion and how I felt each day. This focused my efforts on reducing my severe allergies and to my surprise I found that when treated effectively, many of my other symptoms, such as asthma, reflux, insomnia and headaches also reduced a great deal. It is these kind of indirect discoveries that we can make using a correlative system like mine.
The fascinating relationship between the autonomic nervous system, allergies, mood and chronic fatigue is the topic of a book I am currently writing.
I should point out that I am not a statistician or trained data-miner, so the algorithms in my system are still rudimentary. I look forward to connecting with others in this growing self-quantification community so that I can further develop my system.

You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

AddThis Social Bookmark Button

4 Responses to “I, Medicine: Predictive Biometrics and Health Optimization”

  1. Great post. I fully agree that we’re headed to great increases in personal health monitoring. But it seems to me to have much impact a lot is going to have to be done to make sure data collected is credible. We’ll need standard measures and calibration. We’ll need open data standards. We’ll need standard protocols for data collection and handling. It seems to me it’s going to be hard to get acceptance of personal data by the medical establishment unless unless there’s some rigor in the collection process. And the work of heading toward this bigger goal needs to start soon before the marketing process turns all this into a hodgepodge of proprietary systems and services.

  2. David, you are absolutely right. You’ve highlighted the number one issue that is likely to face this sector in the years to come – standardization.
    I admire the courage and determination of people like Alexander Carmichael and Jamie Heywood who speak regularly about the significant correlations that their websites have identified – and how they have often been confirmed by subsequent medical studies.
    Ultimately I’m not sure that it’s even necessary that the medical community accepts these technologies. All that matters is that the patients accept them and use them in their lives.
    The proof is in the pudding, and if these systems can deliver even 10% of the results we are claiming they will, their success will be assured.

  3. Thanks David and Bard for your comments! On the other side of the “collecting quality data” question, author and professor Clay Shirky is known to say that “the answer to bad data is more data,” so with enough data + powerful enough algorithms, it doesn’t really matter that much if it’s all standardized or not. I agree that the medical establishment will take a long time to accept this kind of data, but I don’t think their acceptance is necessarily required to start having an impact on helping people.

  4. [...] was the question Bard C. sent in for the QS Scientific Advisory Board. His challenge was met by Neil Rubens, Teresa Lunt and [...]

Leave a comment or question