My week with the FitBit Ultra Tracker

Anyone familiar with my research program will know that I'm always looking for new and innovative ways to capture data in clinic that is normally so resource-intensive it is usually reserved for labs or highly specific and well-funded applications.  Sometimes this is through simplified patient self-report measures, sometimes through simplified clinical tests, or sometimes through less expensive alternatives to high-precisions instruments.  An example of the latter is the new FitBit Wireless Activity tracker available from www.fitbit.com.  This thing is loaded with all sorts of technical goodness, including an accelerometer and an altimeter, which means it can not only count the steps you take in a day (a la a pedometer), but also how fast you were moving, and whether or not you were moving up or down hill (or up or down stairs as it were).  It then corrects your calorie usage through knowledge of these other parameters, making it far superior to a standard pedometer.  It also has a 'sleep tracking' mode, which has you move the little USB-drive-sized device from your pocket (where it normally clips happily and unobtrusively throughout the day) to a specially-designed comfortable wristband for sleeping.  While on your wrist, it basically tracks how restless you were throughout the night by monitoring the amount your arms have flailed about.  It will then send all of this information wirelessly through a USB-connected base station/charger any time you're close to your computer, and you can track your calorie expenditure and sleep habit either online or through handy smartphone apps.  So far pretty cool, right?

Pictures are taken directly from the FitBit.com website.

What drew me to this is that, if we believe the cognitive-behavioural model of chronic pain development, the crux of that model is that catastrophizing and fear of movement lead to disuse and deconditioning, and hence chronic pain and disability.  People have tried in the past to monitor change in activity using a variety of approaches, many of them through use of pedometers, accelerometers, GPS, personal digital assistants, one-shot fitness indices (body fat, strength) or plain old diaries.  So far the results are mixed, with a paper from Dr. Vlaeyen himself suggesting that physical acitivity level doesn't change appreciably over the course of chronic pain development.  Colleagues of mine have tried more advanced approaches (unpublished) through use of a wearable vest that incorporates motion sensors, heart rate sensors, and other such technologies that cost several thousand dollars.  This is the 'resource-intensive' bit I was talking about in the opening sentence.

Enter the FitBit.  While it won't provide heart rate data, it will give total motion data, including speed of movement and change in altitude, and an indicator of 'restfulness' while asleep.  All for $99.  So I had to give it a go to determine whether it would be something I could distribute in a prospective cohort study to evaluate overall movement from say weeks 1 and 2 to weeks 51 and 52 after an acute injury.  Here's my experience:

Like any movement sensor, it's possible to 'game' the device by sitting on your couch and shaking it around, but who are you really cheating there?  Setting it up is not hard, but it does require the download of a driver which some people who are a bit more computer-phobic might not like (I'm thinking about my standard subject here).  As a researcher, I love the fact that I can pull the data from the website any time I want by logging into the FitBit 'Dashboard' and seeing total calories spent and monitoring sleep effectiveness.  This also gives me real-time information about compliance, which is a HUGE bonus.  Here's what a fairly typical, perhaps slightly above-average day looked like for me:

Actually, that's probably considerably above average, considering I generally sit in front of a computer most of the day.  The dashboard also gives a sort of summary activity level: 

Which would be difficult to analyze as is but could be tailored I suspect.  And finally, a sleep pattern:

This could come in particularly handy if we believe, as I do, that disturbed sleep is a key indicator of the development of chronic pain.

As far as a clinical/research device, it definitely holds potential.  Some of the possible challenges that would need to be overcome are really just logistic - I occasionally found myself forgetting to put the tracker into the wrist band before bed, then forgetting to take it out and clip it back onto my pajamas when I awoke, and then forgetting to take it off my pajamas and put it onto my pants when I got dressed.  I suspect subjects who aren't as fully engaged as I am would probably forget this even more often.  Similarly, I occasionally forgot to put it into sleep mode before I went to bed, or to take it out of sleep mode when I got up.  This is in no way a criticism of the device, rather when looking at this is as a potential research tool, one needs to consider anything that could bias your results.  That said, it is far superior to either a pedometer or GPS, and is smaller than some other motion sensors I've seen in the past, so definitely has a lot going for it.  The cost is another big plus.  For $5000 I could get 50 of these on a nice prospective cohort and track movement (calories would probably be the key indicator here) over time.  As a clinical tool, these also hold promise for tracking change in activity over time, especially for people with chronic disabilities.  

Now I have to mention one more thing, and this is both a pro and a con.  The device is small and unobtrusive which is great, but as it turns out is also easy to lose.  As I did.  After only 2 weeks.  D'oh!  Now I have to say that the customer service at FitBit are AMAZING.  They have a discounted replacement program, and even offered to replace mine for free (the first time), as long as I promised not to be such a moron again.  Well, they didn't say that exactly, but I read between the lines.  Again, AH-MAZING.  However, this does raise some concern as far as using this as a research or clinical tool - it could become pricey if they keep getting lost.  I will from now on be including a label with my phone number on any ones that I get.  But I would also like to see something like maybe a user-configurable 'alert' system, in which the device can be set to start beeping after some specified period of complete immobility, say 12 hours.

All in all, I love the fact that we are seeing high tech gagets at consumer-accessible prices.  Now, evaluation of patient activity level doesn't have to be confined to the clinic.  I continue to be excited by the next advancement in health technology.  Can't wait to see what else comes down the pipes.

Pressure Pain Threshold for Android

These links are for beta versions of two apps for the Android operating system.  These are currently free, and are just two applications of what we hope will be many for facilitating clinical practice.

Note that these apps will currently only work for the Android mobile operating system.  There are a variety of reasons for this decision that I won't get into here.  Suffice it to say, it is both easier and cheaper to start on Android, and will work on porting apps to other platforms (ie. iOS) assuming there is positive feedback from the Android users.

These first two apps are relatively simple in their implementation.  One is for measuring pressure pain threshold at the angle of the upper fibres of the trapezius muscle (about mid-way between the supero-medial angle of the scapula and the ssecond cervical vertebra).  

The second is for the same purpose over the muscle belly of the tibialis anterior.  

These are the two points we've been using for the past several years for evaluating prognosis in acute whiplash associated disorder (WAD).  We've found that people who are sensitive in both the local (trapezius) and distal (tibialis anterior) areas are at greater risk for long-term problems.  We're not quite sure of the mechanism behind this finding just yet.  It may be an early sign of central sensitization of the nociceptive system, or it may be a function of fear of pain, or perhaps even something else (ie. genetic pain vulnerability).

The numbers used in the app are based on a database of over 300 subjects with neck pain who have been measured using the Wagner FDX-25 digital algometer.  We have previously found this to be a valid and reliable tool that is not overly expensive (about US$395 as of this writing).  That said, it is probably reasonably safe to assume that the numbers should be similar for other digital algometers, but we can't say so with any great confidence as we haven't tested them.  The exception here is the Somedic algometer which is well-regarded as a scientifically rigorous, but pricey, tool.  NOTE: We at cWhIP have ABSOLUTELY NO affiliation with Wagner Instruments in any way.  We use their devices because they're cost-effective and appear scientifically sound.

To install the app, first make sure that your phone is set to allow the installation of third party apps.  To do so, go to settings -> security and make sure that the 'unknown sources' box is checked.  You will likely have to accept a security warning.  

Then you just need to get the .apk file from the links below to your phone.  You can do this by:

1. Saving the file to your computer, then sending it to your phone via usb, dropbox or email.  

2. Opening the browser on your phone, navigating to this page (www.whipresearch.squarespace.com/cwhip-apps), and tapping the download link directly from your phone.  

Most phones running Android 2.2 ('froyo') or higher should be able to install it automatically, but some might requrie an app installer application, such as one called app installer that you can get for free from the Google Play Shop.

To use the app, first choose the units in which you are measuring (pounds, kilograms or newtons) and the sex of the patient.  You MUST make both of these two choices before you can enter data into the other fields.  Once you've done so, perform your first PPT measurement and enter the number into the first box.  Since we recommend you always do every measurement at least twice, separated by a minimum of 30 seconds (1 minute is more desirable), you will also enter a recording in the second box.  After you do so, the app will tell you whether you should do a third in order to get a more accurate mean. This decision is based on the difference between your first two measurements - if the first two are within the standard error of measurement of the device, the app will tell you that 2 measurements are enough.  If they are beyond the standard error, it will suggest you do a third.  You don't have to, it will still calculate a mean with just two, but it at least gives you a little guidance.

After you've entered your measurements, hit the 'calculate' button.  The app will then calculate the mean value, and will compare that to a database of currently 227 females and 82 males, which is constantly growing and will be updated at appropriate intervals.  The app will tell you whether the mean readings are in the first quartile (most sensitive) of the database, the second, third or fourth.  You're generally pretty happy with the third quartile and arguably the fourth.  The second quartile is below the mdedian value and might be a bit more sensitive than you would like, but is generally acceptable.  The first quartile represents the most sensitive 25% of the database.  If you only see these values at the local (traps) area, you might be seeing what is termed primary hyperalgesia, and may or may not be anything to worry about (you would of course interpret all findings in light of all other findings you get from the assessment).  If you also see these sensitive (1st quartile) values in the distal region (tibialis anterior), these are the folks at highest risk of long-term problems.  Again, you would interpret this in light of all of your other findings in order to determine the best course of action, but it gives you at least a quantifiable indication of pain threshold (but not necessarily objective - be aware of the difference!). 

Once again, this is a beta and I am continuing to work on it.  I will include a video in the app shortly to show you precisely how to perform the testing.  I am also working on a system that will allow you to ask the patient's permission to add their anonymized data to the database.  That will take some time to navigate the ethical logistics but isn't technically difficult.  In the meantime, I would love it if you have an Android phone and would give this a go.  Let me know your thoughts in the comments section below.

Here are the links for upper traps (UFT) and tibialis anterior (TA):

UFT_PPT.apk

TA_PPT.apk