Related Experiment Videos
On the number of automated timed manual performance trials to administer
J Busby-Whitehead1, P S Robinson, E S Matteson
1Division of General Medicine and Clinical Epidemiology and the Program on Aging, University of North Carolina, Chapel Hill 27599-7110, USA.
Objective:
To determine the number of trials needed to obtain satisfactory results when evaluating function in older subjects using the automated version of the Williams short (three-door) Timed Manual Performance (TMP) test.
Design:
Administration of from one to five trials in succession on a given test date.
Setting:
A Continuing Care Retirement Community (CCRC), assisted living centers, nursing homes, and a community residence, all located in central North Carolina.
Participants:
The subject population consisted of 182 older volunteers, aged 63 to 100 years. One hundred nineteen lived independently in a CCRC, 33 were assisted-living residents, 29 lived in nursing homes, and one lived independently in the community. Each subject was administered at least three trials in succession on his or her first test date; 23 of the independently living CCRC residents were given three trials on a second test date. The community-dwelling volunteer was administered from one to five trials on each of 26 test dates over an approximately 16-month period.
Measurements:
The time it takes to perform five door opening and closing operations as measured by the three-door Cognatemp Automated Timed Manual Performance (ATMP) system.
Main Results:
Average ATMP time for the subjects living independently decreased approximately 10% between the first two trials but negligibly between trials two and three. The more dependent groups continued to improve between trials two and three. The community-dwelling subject tended to improve in the first three or four trials and to decline by the fifth trial.
Conclusions:
It is recommended that two trials be administered and the best time used; if neither trial results in a time less than 10 seconds, one or two more trials should be administered. It is generally not necessary nor advantageous to administer more than four trials.