Activity trackers … beyond just us fitness geeks

walk-500-3

This video is how I see exercise data in oncology.  Of all the training I’ve done, which I recorded but which no one else did, and of all my data sitting in Garmin, Moves, RunKeeper, Moov, and MiFitLife databases, not one byte of it is in my EHR (Electronic Health Record).  However, during my stay in a hospital isolation ward for my allogeneic stem cell transplantation (donor HSCT) my bowel movements were recorded and entered into my EHR.  Exercise affects gut microbiota (bacteria) and immune function (1), and diversity of gut microbiota correlates to improved survival from a donor HSCT (2).  Did the exercise I did prior to and throughout my stem cell transplantation preserve my gut microbes, affecting my immune system, which helped me breeze through stem cell transplantation?  Who knows?  But since the data for all this exists, it sure would be nice if it were gathered together in the same place (EHR) in order for me and other patients to find out.

At home, my treadmill is Bluetooth compatible, but none of the workout miles I’ve done on it are in my EHR either.

Outside, sigh … you get the picture.  Fitbit, Nike, Jawbone, Garmin, or ‘fitness this’ and ‘fitness that’ app/device, please do not spend one more dime on another advertisement touting your latest model or bells and whistles.  Instead, hire more (or better) IT people proficient in medical coding/HIT, exercise science, and data interoperability, and free our data from your proprietary databases so that it can be used, if we agree, within EHRs for research.  Looking at a graph of the last few months of my training does not benefit others, no matter how inspiring some may think my training to be.  We are losing too many people like Laurie Becklund, and activity tracker data on us may be significant toward improving survival from cancer.  Exercise decreases cancer metastasis, in mice studies anyway (3, 4, 5).  Is the data right at our wrists for metastatic significance in humans?

ResearchKit, can data flow back into subjects’ EHRs?  Exercise researchers, how much of the study data generated from subjects goes back to them so that it can enter their EHR?  EHR vendors, do your patient portals even accommodate exercise data should it become useable?  National Coalition for Cancer Survivorship, can we improve the Journey Forward, Survivorship Care Plan tool’s small section on exercise and populate it with valid activity tracker data similarly to how it is populated with cancer registry data?

This is not rocket science, neither is it a billion dollars in new drug development that marginally improves survival for a few months, this is already here, we just have to capture it in ways that can actually be used to save lives.

There’s a lot of interest in tracking physical activity, ranging from simple weight loss tools to more high-tech gadgets for mountain climbing or ultra-endurance events.  Altitude, distance, speed, calories, heart rate, steps, type of activity, intensity, and even time spent sitting, are just some of the variables being recorded.  Smartphone ‘apps’, watches/bands, heart rate straps, pocket/clip-on devices, ear buds, and web-based diaries are able to collect these variables.  However, few of these devices, including ‘apps’, have been validated – scientifically tested for accuracy against a criterion (standard).  This is important if we want to use consumer fitness data (Patient-Generated Health Data – PGHD) for more than just personal curiosity, which typically wanes after a few months.  If you haven’t validated your device, do it.  If you fear the results, then improve your product so that it accurately records valid fitness measures.

For exercise-oncology research, and exercise research in general, in many ways PGHD from validated activity trackers can be more accurate than Patient Reported Outcomes from validated questionnaires.  Either way, data comes from patients, but some fitness trackers are as accurate as the criterion (6).  This I like because it will require other fitness trackers to improve accuracy.  Developers unwilling to validate and improve their device will be relegated to the Big Data sidelines, if they survive at all.  Regardless of how sophisticated and proprietary a devices sensors are, most end measures will be the same as from other devices – energy expenditure – kcal, estimated VO2, METs.  No longer should an app be able to use a metabolic equation inaccurately and have their data be relevant.

Validated activity trackers have the potential to expand physical activity related observational research to every Electronic Health Record (EHR) – this is a big number, over a billion, which is much larger than the fitness geek marketplace.  For small population cancer types, which get little, or no, exercise research due to their inability to recruit enough patients from single or multiple healthcare facilities, this could do wonders for statistical power.  Might we find similar benefits for exercise among other cancer types as have been found in brain, breast, colorectal, and prostate cancers?  Will we discover more information about intensity, duration, frequency, and type of exercise regarding proximity to diagnosis and treatment?

For cancer patients, having our physical activity automatically tracked, medically coded, encrypted, summarized, and made available for upload into our EHRs, this may be the least invasive thing involving our body, and with the lowest cost per survival outcome.

 

1. The microbiota: an exercise immunology perspective. Bermon, S., et al.  Exercise Immunology Review  2015;21:70-9.

2. The effects of intestinal tract bacterial diversity on mortality following allogeneic hematopoietic stem cell transplantation. Taur, Y., el al.  Blood, 14 August 2014 x Volume 124, Number 7.

3. Effects and potential mechanisms of exercise training on cancer progression: A translational perspective.  Allison S. Betoff, Mark W. Dewhirst, Lee W. Jones. Brain, Behavior, and Immunity 2013 Mar;30 Suppl:S75-87.

4. Exercise modulation of the hosttumor interaction in an orthotopic model of murine prostate cancer.  Jones, LW., et al. J Appl Physiol (1985). Jul 15, 2012; 113(2): 263–272.

5. Exercise modulation of the host-tumor interaction in an orthotopic model of murine prostate cancer.  Jones, LW., et al.  J Appl Physiol 113: 263–272, 2012.

6. Earbud-based sensor for the assessment of energy expenditure, HR, and VO2max.  Lebouf, SF., et al. Med Sci Sports Exerc 2014;46(5):1046-52.