The trouble with activity trackers … or not.

A recent study (1) found that a high level of cardiorespiratory fitness (CRF) was associated with an increased risk for localized prostate cancer.  The reasons for this are unknown.  The researchers speculate that perhaps this group was more likely to undergo preventative screening or detection.  However, higher CRF still showed a 32% decreased risk of cancer specific death for lung, colorectal, or prostate cancers; or 68% decreased risk of death from cardiovascular disease (CVD).  Note: some cancer treatments can be toxic to the heart.

From: Midlife Cardiorespiratory Fitness, Incident Cancer, and Survival After Cancer in Men:  The Cooper Center Longitudinal Study.

Lakoski, S.G., et al. JAMA Oncol. 2015;1(2):231-237

Figure Legend: Cardiorespiratory Fitness (CRF) and Risk of Incident Lung, Colorectal, and Prostate Cancer.  The low CRF group is the referent group relative to moderate and high fitness.  The error bars for moderate and high fitness represent the 95% confidence limits.  Adjusted for age, examination year, body mass index, smoking, total cholesterol level, systolic blood pressure, diabetes mellitus, and fasting glucose level.

The authors distinguish CRF from physical activity (I believe research data for both could be provided by valid and reliable activity trackers):

“Cardiorespiratory fitness is also highly reproducible and objectively assessed via incremental exercise tolerance testing compared with physical activity, which is largely determined by self-report questionnaires [and/or activity trackers?].  A prior study demonstrated that CRF is be a more potent marker of mortality than physical activity.  As such, given the current study findings and prior evidence, we contend that measurement of CRF should be used more frequently in the cancer prevention setting.”

I agree.  Furthermore, I would like to see physical activity, CRF, or aerobic capacity assessed when the cancer diagnosis process begins.  How beneficial would it be to tie fitness to an actual biopsy tissue specimen?  It’s interesting that CRF in the Cooper Clinic Longitudinal Study was assessed by the duration of performance achieved on a maximal treadmill test (2).  Then, based on subjects’ performance time, maximal oxygen uptake (VO2max) and maximal METs achieved were estimated, not measured.  If estimates can be used to assess CRF then it’s possible that some activity trackers could also be used.  Granted, screening patients before a CRF test is recommended, but some activity tracking data may already provide an adequate assessment of CRF.  A few devices already assess VO2max using heart rate, and with acceptable errors (for field measurements) in the 6-7% range (11, 12).  Stratifying data from activity trackers may be an important part of sorting its value: data for showing a training effect requires good accuracy; less accurate data is probably acceptable to assess CRF; and, data for tracking physical activity volume (MET-hours per week, etc.) can perhaps be the least precise of these – particularly since current population research using questionnaires tends to overestimate actual physical activity (13).

In discussing limitations of their study the authors mention something I believe may be significant for exercise-oncology research, and which I think validated activity trackers may be able to provide data for:

“CRF was assessed years prior to a diagnosis of lung, colorectal, or prostate cancer or death in men diagnosed as having cancer.  Thus, it is not known how changes in CRF and related behaviors, such as physical activity from the initial preventive health care to cancer diagnosis as well as changes in CRF and physical activity after diagnosis, may have had an impact on these current findings.”

I believe that exercise during the time from cancer diagnosis until first treatment will be found to have a positive impact on cancer treatments, treatment side effects, and on survival.  Sophisticated activity trackers that also estimate VO2max, or measure heart rate variability (HRV), which is related to CVD, have the potential to provide data in and around the diagnosis/treatment time period.  Furthermore, they can provide data across more cancer types by doing it in a more cost-effective manner than mailing out questionnaires or doing a CRF test on every cancer patient.  One overlooked benefit of activity trackers is that consumers subsidize the data.

Some useable physical activity data already exists in activity tracking databases but sits there underutilized.  Most physical activity data needs standard medical codes to improve its interoperability.  Other data could be retooled by correcting METs, which could provide more accurate estimates of energy expenditure (4, 5, 6, 7, 8), population specific intensity levels (9, 10), and might influence adherence to exercise training programs.  Regarding METs, an issue for some researchers is that the ‘standard’ MET (3.5 ml oxygen/kg/min) was based on the measurements derived from one 70 kilogram, 40-year-old man (5), and then applied to survey research.  Conversely, some activity trackers use ‘standard’ MET values from the Compendium of Physical Activities, which are intended for survey research, to estimate the energy expenditure and exercise intensity for an individual, which the Compendium advises is not its intended purpose.

Besides valid data, another issue activity trackers face is how should data be displayed or reported within an Electronic Health Record (EHR)?  Doctors are already over-worked and many complain about the burden of EHRs, adding physical activity data to their workload and expecting them to do something proactive with it (without reimbursement too) is not going to happen.  Make physical activity data easy for doctors to accommodate: summarize activity tracker data into an indicator of ‘compliance‘ or ‘non-compliance‘ with recommended physical activity guidelines, and provide that to an EHR.  For research, and for the more inquisitive and less time constrained physician, the underlying data supporting a compliance indicator could be accessible via EHR patient portals (e.g. EPIC’s MyChart).

Finally, a new study (3) found the ActiGraph GT3X+ accelerometer not to be very accurate at low and moderate intensity levels.  Of the few validation studies done on accelerometer based activity trackers, some were validated against the Actigraph as the criterion measure.  However, this study itself also missed an opportunity for better measurement when they estimated Resting Metabolic Rate (RMR) using the Schofield equations rather than measuring it with the Oxycon Mobile system they had – RMR is essentially what 1 MET is.  The study’s authors do disclose that they have receive funding support from Bodymedia, which Jawbone recently bought.

There is more to be sorted out in the consumer fitness/activity tracking eco-space.  I think devices and apps that produce valid and reliable data can make an impact in exercise-oncology research, particularly in the time periods surrounding diagnosis and treatment.

1. Lakoski, S.G., et al.  Midlife Cardiorespiratory Fitness, Incident Cancer, and Survival After Cancer in Men The Cooper Center Longitudinal Study.  JAMA Oncol. 2015;1(2):231-237. doi:10.1001/jamaoncol.2015.0226

2. Pollock ML, Bohannon RL, Cooper KH, et al.  A comparative analysis of four protocols for maximal treadmill stress testing. Am Heart J. 1976; 92(1):39-46.

3. Kim, Y., Welk G.J. Criterion Validity of Competing Accelerometry-based Activity Monitoring Devices. Med. Sci. Sports Exerc. 2015 Apr 23. [Epub ahead of print]

4. McMurray, R.G., et al.  Examining Variations of Resting Metabolic Rate of Adults: A Public Health Perspective. Med. Sci. Sports Exerc., Vol. 46, No. 7, pp. 1352–1358, 2014.

5. Byrne, N., et al. Metabolic equivalent: one size does not fit all. J Appl Physiol 99: 1112–1119, 2005.

6.  Kozey, S., et al.  Errors in MET Estimates of Physical Activities Using 3.5 ml·kg–1·min–1 as the Baseline Oxygen Consumption. Journal of Physical Activity and Health, 2010, 7, 508-516.

7. Wilms, B., et al.  Correction factors for the calculation of metabolic equivalents (MET) in overweight to extremely obese subjects.  International Journal of Obesity (2014) 38, 1383–1387.

8.  Hall, K., et al.  Activity-Related Energy Expenditure in Older Adults: A Call for More Research. Med Sci Sports Exerc 2014 Dec;46(12):2335-40.

9. Blair, C.K., et al.  Light-Intensity Activity Attenuates Functional Decline in Older Cancer Survivors. Med Sci Sports Exerc 2014 Jul;46(7):1375-83.

10. Herzig, K-H, et al.  Light physical activity determined by a motion sensor decreases insulin resistance, improves lipid homeostasis and reduces visceral fat in high-risk subjects: PreDiabEx study RCT..International Journal of Obesity (2014), 1–8

11. Montgomery, P.G., et al. VALIDATION OF HEART RATE MONITORBASED PREDICTIONS OF OXYGEN UPTAKE AND ENERGY EXPENDITURE. Journal of Strength and Conditioning Research 23(5)/1489–1495.

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

13.  A systematic review of reliability and objective criterion-related validity of physical activity questionnaires. International Journal of Behavioral Nutrition and Physical Activity 2012, 9:103 pgs 1-55.


5 thoughts on “The trouble with activity trackers … or not.

  1. I have chronic lymphocytic leukemia. Untreated: wait, watch, and: work-out! I’m looking for the “best”, most suitable trackers for me. So I read several of your interesting posts. I wonder what kind of tracker(s) you are using, and which you would advise…

    • Robert,

      I have tried a few activity trackers and apps, opting for the inexpensive route until the more expensive devices or apps show better reliability and accuracy. I’ve found (unscientific) the Misfit Flash Link ($20) to be accurate for distance both indoors and outside. This is important to me since I will soon be running indoors on a treadmill because of the affect that colder weather has on the peripheral neuropathy in my fingers due to chemotherapy. I’ve also found (again, unscientific) that the iphone 6 is accurate for distance both indoors and outside, so I carry that whenever I run and listen to music. The iphone 6 ‘flight of stairs’ measure also seems to be accurate. I will soon be purchasing a new treadmill (Landis next generation commercial model) that is Bluetooth enabled, which should allow my iphone 6 to sync to it and automatically record distance, METs (Metabolic Equivalent of Task – a research measure), calories, speed, and heart rate. However, I don’t think any devices are currently recording calories accurately.

      Apple recently acquired Gliimpse, a company that integrates personal healthcare data using standardized medical IT codes – the same codes that are used in Electronic Health Records (EHRs). Eventually, all my activity tracking data should automatically be entered into an iphone’s (6, +) Health Kit, and from there summarized and entered into my EHR, complete with embedded clinical healthcare IT codes, including research measures like METs (currently Health Kit doesn’t include METs).

      So, IMHO, depending on what you are interested in measuring, if you already have an iphone 6, you probably have what you need for distance. If you don’t have an iphone 6 (or greater), the Misfit Flash Link is an inexpensive and accurate first purchase to record distance. The iphone 6 has an accelerometer, gyroscope, and barometer that Apple has done a good job with combining the raw data into an algorithm that gets distance correct – even indoors on a treadmill (without capturing the treadmill’s distance data). This is significant – I’ve tried a number of different apps (Moves, Argus, Strava, RunKeeper, Runtastic, plus the Moov device & app) on the iphone 6 and they do not get distance correct indoors on a treadmill (although some appropriately disclaim their app for treadmill use). Some of these apps suggest calibrating my steps to distance for accuracy, which I haven’t done, but neither did I do with the Misfit or iphone 6. I think the indoor distance problems are because of a reliance on cell phone tower triangulation, which is inhibited indoors (same with GPS), or else their algorithms need to be tweaked – they should have access to the same iphone 6 raw acceleration data too (although Moov has their own accelerometer). The Moves app (bought by Facebook) typifies this inaccuracy on my iphone. The Moves app also has a problem when I’m running outside around a 400 meter track, it typically records just half the distance I actually run. I’m not sure if Apple’s iphone GPS also triangulates to cell phone towers or if it connects to a satellite, but whatever they are doing they get distance correct on my iphone.

      That said, I’d like to see Apple publish some White Papers on their iphone’s and on the Apple Watch 1 & 2. There are only a few studies so far that have assessed these Apple products, too many people are just assuming that these products are accurate for all the measurements they are recording, healthcare professionals included, and I think accuracy should come first before people start using the data to make any strong inferences about health.

      10,000 steps, a number promoted in the 1960’s by a Japanese pedometer manufacturer, currently has limited clinical value and does not give an indication of the intensity of physical activity. However, I expect that to change as more steps data accumulates (assuming that data is valid). However, if you’re interested in steps now, more than 10,000 steps might be advisable: “… ≥10 000 steps/day indicates the point that should be used to classify individuals as ‘active’. Individuals who take >12 500 steps/day are likely to be classified as ‘highly active’.” How Many Steps/Day Are Enough? Preliminary Pedometer Indices for Public Health. Sports Med 2004; 34 (1): 1-8

      I’m assuming you might have come across my blog through the CLL Society, where I wrote an Exercise and Cancer Survivorship article. If not, FYI, here is the link:

      Here is an article on a recent study that looked at 24-hour activity trackers. Of concern to me, by the time the study was published some of the devices had already been replaced or their software updated. This makes keeping abreast of valid of consumer fitness devices a challenge since the devices change rapidly and not enough manufacturers/developers publish white papers on the validity of their products’ measures.

      Firstbeat is one of the few companies that publishes and lists studies on their products. They also license their heart rate technology, Garmin is one of their licensees. I don’t know how well Firstbeat devices/technology works indoors, but if you are looking for devices that also give VO2 estimates, Firstbeat/licensees should be about as trusted as you can currently get.

      For the Apple iphone 6, from Chipworks: accelerometer & gyroscope – Bosch BMI160 Circuit; Invensense MPU-6500/6515 ASIC Full Analog Circuit; Bosch Triaxial Gyroscope ASIC from BMI055/BMG160 Circuit; Maxim Integrated MAX21000 3-Axis Digital Output Gyroscope ASIC Circuit.

      Disclosure: I receive no compensation or product from any of the companies mentioned above.

  2. Thank you so much for your speedy & comprehensive answer! I really appreciate that.

    I had not seen your CLL article before. You wrote: “Moderate intensity or volume of exercise, including one of the few exercise studies on CLL, has been shown to positively affect proportions of T-cells and the mobilization of Natural Killer cells, both of which can kill abnormal, virally infected, and cancer cells. However, some research found that high volumes and intensity of exercise had a negative effect on immune cells, suggesting that a lot more exercise might not be better.”

    So the challenge for me will be to find the optimal volume and intensity. And absolutely avoid over training, inflammation, airway infections, etc. Regular DIY testing of immunity is impossible. But perhaps I should “follow my heart” and use tools that measure Heart Rate Variability (and Aortic/Arterial Stiffness) on a daily basis.

    I don’t like running (cold/rainy weather, airway problems). What I’ve done & used until now:
    – iPad (filled with health/fitness/movement apps)
    – 2 Moov fitness trackers for acticity tracking (3 active hours per day) + guided cardio boxing/rapid walking/7 minute fitness
    – a mechanical step counter (10.000 + steps…)
    – special breathing techniques for people with cancer
    – some Chinese practices (special qigong styles), said to be effective for people with cancer. This is low intensity exercise, but high volume: proponents advise at least 3 hours per day. I’m looking into thst more.

    With all your information, I have quite some homework to do..! 🙂

  3. Thank you so much for your comprehensive answer!

    I read Firstbeat’s whitepaper that your mentioned, but had not seen your CLL article yet. In it you write: “Moderate intensity or volume of exercise, including one of the few exercise studies on CLL, has been shown to positively affect proportions of T-cells and the mobilization of Natural Killer cells, both of which can kill abnormal, virally infected, and cancer cells. However, some research found that high volumes and intensity of exercise had a negative effect on immune cells, suggesting that a lot more exercise might not be better”.

    So the challenge (here & always): to find the optimal intensity and volume. And absolutely avoid over training, inflammation, airway infections, etc. DIY immunity testing is difficult. Possibly tracking HRV (heart rate variability) on a daily basis helps. And perhaps tracking aortic/arterial stiffness…

    Outside running is not my thing (due to cold/rainy weather & airway irritation). What I’ve been doing/using until now:

    iPad (filled with health/fitness apps)
    2 Moov trackers (for sleep & activity tracking, guided cardio boxing/fast walking/7 minute fitness)
    mechanical step counter (yes, for those 10.000 +)
    lots of short trainings, Tabata interval training, rebounder, body weight, mobility (I particularly like it short with moderate intensity: that gives variety & prevents rsi/injuries/overtraining).
    breathing methods for people with cancer
    some Chinese methods (special qigong styles): low intensity but high volume. For chronic diseases like cancer they advise at least 3 hours of practice daily…

    With all your information I have some nice homework to do… 🙂

    Did you see this recently published article? “People who achieve total physical activity levels several times higher than the current recommended minimum level have a significant reduction in the risk of the five diseases studied.”

  4. Robert, sorry for the delayed response, and thank you for the article link, no, I had not seen that one yet. More questions please use: Here are some responses to the study you might find interesting:

    My take, yes, more is better for RISK reduction, however, there appears to be diminishing returns at high levels of physical activity. I personally wouldn’t classify the studies’ subjects as ‘very high’ either, which I would do for triathletes training for an Ironman. I did the MET numbers on me exercising 6 days a week for just 30 minutes (fairly quick pace 7 min miles and classified as ‘vigorous intensity’) and came up with ~ 2,200 MET-minutes per week, which does not include other physical activities like yard work, household chores, etc. So I would estimate that I’m typically getting 3,000+ MET-minutes Total physical activity per week, but I personally consider that a moderate level of activity even though much of that value is made up of ‘vigorous intensity’ exercise and would be categorized as a high level.

    For cancer patients, not RISK reduction for cancer, my concern (and lack of data) is what effect does high levels of exercise have on immune function, particularly if our immune system is already compromised by cancer? Maybe more is better, but to a point as the study indicates for RISK reduction.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s