Cheap Pedometers Don't Step Smartly

GHENT, Belgium, June 23 ? Inexpensive pedometers can't be counted on to keep step accurately, researchers reported here, and may send the wrong message in fitness-walking programs.

Three out of four inexpensive pedometers tested failed to keep accurate count, and almost two-thirds of the devices overestimated the number of steps taken, according to a study reported online by the British Journal of Sports Medicine.

Pedometers have become popular as monitoring and motivational tools in promoting physical activity in adults, said Katrien De Cocker and colleagues in the department of movement and sports sciences at Ghent University here.

However, tests have found considerable variation in accuracy between different types and brands of step counters. For this reason, the accuracy of the pedometers used in public health interventions must be accounted for, the Ghent researchers said.

In preparing a large community-based physical activity intervention, the 10,000 steps/day campaign, the researchers decided to use a widely accessible, inexpensive model, the Stepping Meter (.20 each), which could be handed out free. But first, they tested the accuracy of the device.

In a convenience sample of 35 healthy adults (nine men, 26 women; 20 to 60 years old), the validity of more than 1,000 inexpensive Stepping Meters was tested against a gold standard, the Yamax Digiwalker SW-200 (available in the U.S. for ), known to be one of the most reliable and accurate pedometers, the researchers said. A variation in step count of no more than 10% from the Digiwalker was considered acceptable.

Each volunteer was given 30 cheap pedometers, one Digiwalker and an automated step-count log. The participants wore five different cheap pedometers on the right and then the left side of the body every day for six days, recording the number of steps counted each day in the log. At the end of the study, the researchers calculated the step-count differences between the Stepping Meters and the Digiwalker.

Only 25.9% (252) of the Stepping Meters met the 10% criterion, whereas 74.1% (721) over- or under-counted by more than 10%. For more than one third (36.6%) of the Stepping Meters, the deviation was greater than 50%.

Almost two-thirds (64.8%) of the invalid pedometers overestimated the actual steps taken, the researchers reported. The magnitude of error, expressed as absolute percentage error (regardless of direction) was 60.4%.

The large range of deviation, without any consistency, is significant, the researchers said, because an error of 20% in 10,000 daily steps is 2,000 steps. Thus either 8,000 or 12,000 steps are recorded, significantly changing the activity level of the user.

Participants mentioned that the Stepping Meter also registered non-stepping movements such as sitting on a chair, twisting the hip, bending over, and kneeling. Furthermore, some said, the device could unintentionally be reset during the day by accidentally pushing the exposed reset button. The Digiwalk has a cover to prevent accidental resetting, the researchers noted.

Previous studies had suggested that the sensitivity of the internal mechanism may have been the main cause of the counting inaccuracy. The threshold needed to trigger a step may be too sensitive and may be receptive to non-stepping movement, the study authors wrote. However, the devices also undercounted, perhaps resulting from inadvertent use of the unprotected reset button, they said.

The results of this pilot study indicate that the inexpensive Stepping Meters are inappropriate for physical activity promotion targets, the researchers wrote.

The wide accessibility of pedometers, as a way of increasing fitness or encouraging weight loss, needs encouragement, but because of validity problems, the use of an untested inexpensive pedometer is not recommended and may give users the wrong message, they said.

"The introduction of a quality label given to all valid, accurate, and reliable pedometer types and brands would be useful for researchers and individual users," they concluded.