Symptoms of intermittent claudication might influence daily life activities. Moreover, IC may have profound consequences for exercise ability, behavior and levels of physical activity (PA). Findings of this 'real life' design study indicate that activities detected by the DP showed high sensitivity and specificity values for most activity categories except for "standing" and "shuffling".
Measuring inter-rate reliability (IRR) characteristics is an important step in the evaluation of a 'gold standard' that is used to validate novel tools such as the DP. Strong correlations concerning the IRR were found in 6 out of 7 activity categories. In contrast,
a poor IRR was found for shuffling (ICC of 0.38). High IRRs were previously found in a similar but smaller video study of 5 elderly patients (ICC of 0.95, 0.78, 0.99 and 0.98 for walking, sitting, standing and lying, respectively) (12). The poor ICC for shuffling is probably due to ambiguities in our annotation protocol (table 1). Furthermore, transitions per se were difficult to score using a 'per second' time window analysis method, whereas the duration of a particular transition between activities is often a matter of seconds, a phenomenon that was also reported previously (11).
A poor sensitivity of 46.2% of the DP was found for the activity "standing" (PPV: 47.8). These results are somewhat lower as compared to other studies with community-dwelling patients and Parkinsons disease patients (80.2% and 81.4%, respectively) (10). Additionally, we found a large standard deviation (19.0%) for the "standing" activity, which can be explained by 10 aberrant measurements. It is thought that the DP missed several transitions between sitting and standing that may have considerably influenced outcomes, especially due to increased durations of activities. The low PPV of 'shuffling' (3.7%) is particularly striking and induced by a low IRR and a difference between the annotation protocol and the algorithms that were used by the DP in allocating transitions to specific activities. According to our protocol, observers assessed transitions between activities as part of the previous activity, while in contrast the DP scored them as shuffling. Therefore conclusions regarding shuffling cannot be drawn from this study. However, it should be appreciated that shuffling activities were just observed during a relatively short period (1% of total time recorded). Therefore, the effect of shuffling on overall PA performance is very limited.
An overestimation of gait duration (10.7%) and underestimation of number of steps (7.4%) in Parkinsons disease patients was previously found (12). However, these studies were performed in a laboratory setting with relatively short distances and just 236.8 minutes of video observation per study. Despite that walking patterns are altered in IC patients (16-18), the present study that was executed under daily life conditions using prolonged video observation periods demonstrated that the DP can accurately measure number of steps in a IC population.
Studying PA levels and ambulatory activities in daily life may yield important information regarding quality of life, health status or mortality in chronic diseases such as COPD and PAD/IC. There is an increasing interest in understanding associations between sedentary activities and their impact on quality of life (20-22). As a consequence, research interest may shift from focussing on measurement of exercise capacity (maximal or pain-free walking distances) to determining daily activities over a prolonged period of time. Our study shows that the DP can correctly measure most types of ambulatory activities in daily life. Future research should aim to determine the value of PA as a novel outcome measurement in IC related topics such as in the comparison between IC treatment modalities.
Video-observers and the DP scored shuffling differently. In retrospect, positional change would have been marked as 'shuffling'. Furthermore, reference values of physiological or walking intensity parameters were not obtained although the DP is capable of determining the intensity of periods of locomotion (movement intensity) (23). Therefore, a validation of energy expenditure values could not be made. Although movement intensity may be of importance in IC patients, the present study focused on validation of periods of activities rather than intensity. The assessment and the effect of treatment strategies on physical activity and daily activities in PAD patients should be subject of future research.