Article Text

Validation of two activity monitors in patients with COPD
  1. D Langer1,2,
  2. R Gosselink1,2,
  3. R Sena2,
  4. C Burtin1,2,
  5. M Decramer1,2,
  6. T Troosters1,2
  1. 1
    Respiratory Rehabilitation and Respiratory Division, UZ Gasthuisberg, Leuven, Belgium
  2. 2
    Faculty of Kinesiology and Rehabilitation Sciences, Katholieke Universiteit Leuven, Leuven, Belgium
  1. Dr T Troosters, UZ Gasthuisberg, Respiratory Division, Herestraat 49, B-3000 Leuven, Belgium; thierry.troosters{at}med.kuleuven.be

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Physical activity in daily life is increasingly used as an outcome measure in chronic respiratory disease.1 Valid and user-friendly instruments are needed to quantify daily activity.1 2 The DynaPort activity monitor (McRoberts, The Hague, The Netherlands) has been validated and used in patients with chronic obstructive pulmonary disease (COPD).3 4 The device is, however, technically difficult to handle and, due to its size (12.5×9.5×3 cm, 375 g), it is always noticeable. We therefore validated two smaller activity monitors in a sample of 10 patients with COPD (mean (SD) forced expiratory volume in 1 s 49 (16)% predicted; mean (SD) age 65 (8) years) and 10 healthy elderly volunteers (mean (SD) age 65 (9) years). Detailed characteristics of the study subjects are summarised in table 1 in the online supplement.

All patients simultaneously wore the DynaPort, the SenseWear Pro (SenseWear, Body Media, Pittsburgh, USA) activity monitor (8.5×5.0×1.5 cm, 85 g), the DynaPort Minimod (Minimod, McRoberts, The Hague, The Netherlands) activity monitor (8.5×5.0×1.0 cm, 70 g) and a portable metabolic system (VmaxST 1.0, Viasys, MEDA, Belgium) during a 53 min protocol including different postures and activities (see table 2 in online supplement for a detailed description of the protocol). In a first analysis the accuracy of the Minimod and DynaPort to detect time spent in walking and time spent in different postures was validated against video analysis; the step count of the three activity monitors was then validated against manual step counting and, finally, estimates of energy expenditure from both the Minimod and the SenseWear were validated against indirect calorimetry. Additional information on material and methods is available in the online supplement.

The Minimod was as accurate as the previously validated DynaPort in detecting time spent in different postures and in walking (see table 3 in online supplement). During the protocol, a mean (SD) of 1976 (220) steps were manually counted. Excellent agreement with the manual step count was observed for the Minimod (mean (SD) step count 1891 (363) steps; fig 1A), with the exception of a large underestimation in one patient who walked slowly (2.5 km/h compared with other slow walking speeds of 2.8–4.5 km/h). This outlier was excluded from the statistical analyses (see details in online supplement). Agreement between the manual step count and the SenseWear monitor (mean (SD) step count 1512 (517) steps) was worse. The SenseWear step count differed significantly from the manual step count (difference between means −465 steps (95% CI −717 to 213), p<0.001; fig 1B).

Figure 1

Mean sum of steps plotted against differences between sum of steps of (A) the Minimod monitor and manually counted steps (Step Count): mean bias −43 steps (95% CI −146 to 60) and (B) the SenseWear monitor and Step Count: mean bias −486 steps (95% CI −1278 to 306). *Subject walking slower than 2.5 km/h.

No significant differences were found between energy expenditure estimates from indirect calorimetry (mean (SD) 144 (5) metabolic equivalents (MET)-min), the Minimod (mean (SD) 153 (6) MET-min; +6%) and the SenseWear (mean (SD) 139 (6) MET-min; −4%). Correlations and agreement between energy expenditure estimates from activity monitors and indirect calorimetry are shown in figs 1 and 2 in the online supplement.

Of particular interest for physical activity intervention studies is the ability of devices to detect minutes of moderate intense physical activity (⩾3 METs). Important health benefits have been described mainly for activities that are performed at moderate intensity (⩾3 METs).5 6 No significant differences between the SenseWear and the Minimod were found for the ability to detect minutes spent sedentary (<3 METs) (84 (11)% agreement vs 84 (7)% agreement with minutes classified as sedentary by indirect calorimetry). The SenseWear, however, detected significantly more minutes of moderate intense physical activity than the Minimod (80 (11)% vs 70 (13)% agreement with minutes classified as moderately intense activity by indirect calorimetry; p = 0.03).

We conclude that the SenseWear and the Minimod provide complementary information on habitual physical activity and could be useful both as outcome measures and for self-monitoring of daily activities in physical activity intervention studies in COPD. The Minimod is a very accurate instrument for detecting postures, walking and steps. The SenseWear does not provide information on time spent in postures and walking, and step counts are not accurate. It was, however, the better of the two devices in identifying minutes of moderate intense physical activity.

REFERENCES

View Abstract

Supplementary materials

Footnotes

  • ▸ Additional data are published online only at http://thorax.bmj.com/content/vol64/issue7

  • Funding: This research was funded by Research Foundation Flanders grants G0523.06 and G05989.09

  • Competing interests: None.

  • DL and CB are doctoral fellows of the Research Foundation Flanders.