Physical activity assessment – Combination heart sensors

The measurement of physical activity by combined sensors is a developing area of research. In this section of the toolkit, discussion centres on combined heart rate and motion measurement; this has been relatively well validated compared to other motion sensors which are combined with physiological measures e.g. temperature, blood pressure.

Over twenty years ago, a study demonstrated that the combination of heart rate and motion registration improved estimates of oxygen consumption and carbon dioxide production (Avons et al, 1988). This study showed that the addition of movement registration to heart rate monitoring, the more common field approach at the time, improved estimates of energy expenditure. Haskell et al (1993) performed multiple regression analyses to predict energy expenditure from arm and leg motion sensors and measured heart rate during a variety of laboratory activities. The addition of the motion sensors increased accuracy of the physical activity energy expenditure estimate (Haskell et al, 1993). This work was replicated in a combination of treadmill and simulated daily living activities conducted in a laboratory (Luke et al, 1997). The addition of motion sensors again improved accuracy, especially during activities of daily living. Further combined monitoring studies have been undertaken in adults (Rennie et al, 2000; Strath et al, 2001; Strath 2002) and children (Treuth et al, 1991; Moon et al, 1996; Eston et al, 1998). All these studies demonstrated increased accuracy when using combined measures rather than each of the measures in isolation.

Combining motion and heart rate monitoring utilises the unique advantages of each method, thereby negating some of the disadvantages of each method used alone; the measurement error from the two methods is not positively correlated (Brage et al, 2004). At lower levels of intensity, heart rate is less accurate at estimating energy expenditure; this is the level that accelerometer models have low error. Conversely, activities performed at high intensity, especially biomechanically diverse activities, are assessed with great uncertainty with accelerometry but measured well with heart rate monitoring, in particular if this information is individually calibrated. Activities not measured well by waist-mounted accelerometers such as cycling, walking on incline, carrying weights and activities involving predominantly upper-body work will be captured well by heart rate monitoring (Brage et al, 2005; Strath et al, 2005). Additionally, non-wearing time segments in non-labelled activity records from free-living are more easily determined from physiological signals than when using an accelerometer alone, since “no motion” looks the same regardless of the monitor is worn or not. However, attention must be given to handling measurement noise in long-term recordings obtained during free-living, a phenomenon which is more common in physiological signals such as heart rate (Stegle et al, 2008).

Heart rate data and accelerometer data have been captured simultaneously using two separate devices and it is important to remember that the principles of combined monitoring can be applied regardless of the capturing device used. The preliminary testing of a single-piece combined sensor was reported in 2000 but this device was never available commercially (Rennie et al, 2000). Conditional modelling underpinned the analysis of the acquired data. Rapid advances in technology have resulted in more sophisticated devices which are capable of monitoring heart rate using digitalised ECG signal and simultaneously measure motion by an integrated accelerometer; the output can then be integrated to provide an estimate of physical activity energy expenditure. The first commercially available combined sensor is the Actiheart sensor (Brage et al, 2005) which:

has a main component 7mm thick with a 33mm diameter and houses a movement sensor, a rechargeable battery, a memory chip and other electronics
weighs 8g
does not require a chest strap; instead electrodes are attached to the chest.
is waterproof; indivduals only remove it to replace pads that have perished
offers a choice of epoch length-15s, 30s, 60s
the memory capacity allows 11-days of continuous monitoring using a 60 s epoch (or 15 s epochs in newest generation which has 4 times enhanced memory capacity)
allows collection of additional heart rate variability (HRV) information during free-living which provides useful information on the quality of data
provides advanced analyses to estimate energy expenditure more accurately

The initial reliability and validity study of the Actiheart (UK) was undertaken in 2005 (Brage et al, 2005). In this study, the intra- and inter-instrument reliability and validity of the heart rate and motion sensor during electronically simulated heart rate, mechanical shaking, and also during rest, walking and running activities were demonstrated (Brage et al, 2005). Other combined heart rate and motion sensors are available, e.g. the ickal but there is not yet much literature describing this device (Berntsen et al, 2009). In general, combined heart rate and motion sensing allow relatively accurate estimation of activity energy expenditure across the intensity range. For example, low to moderate activities performed by adults in a laboratory setting were well-captured by branched equation modelling of the two sources of information (Thompson et al, 2006). In a study in children, the validity and predictive accuracy of combined sensing was investigated in 39 children during laboratory based treadmill walking and running. Physical activity energy expenditure was measured continuously using indirect calorimetry. The activity energy expenditure from the combined model had the strongest agreement with measurements from indirect calorimetry and accounted for 86% of the variance (Corder et al, 2005). Another evaluation of branched equation modeling (using Actiheart, US version) validated against indirect calorimetry during a wide range of activities in a laboratory setting also reported that this particular technique produce valid estimates (Crouter et al, 2007).

Calibration
A study investigating a range of calibration techniques with decreasing levels of complexity showed that simple calibration techniques (walk and step test) achieve acceptable levels of accuracy for the combined technique to be considered as an objective measure in population studies (Brage et al, 2007).

A flexible but consistent approach to calibration has been suggested which spans individual calibration over a range of activities at different intensities to static calibration which accounts for parameters know to affect the heart rate – energy expenditure relationship such as age, sex or sleeping heart rate (Strath et al, 2005).

The software of the Actiheart (distributed by Camntech,
Cambridge, UK) incorporates a step test that permits individual HR-VO₂ calibration. Alternatively, group HR-VO₂ calibration equations may be used with the in-built algorithms to estimate physical activity energy expenditure. Individual calibration may, however, be performed with any device which measures heart rate, for example using the same protocol (step frequency prompt available from www.mrc-epid.cam.ac.uk). All that is required is that the raw data are exported and analysed in standard statistical packages with appropriate programming.