The role of sampling duration on basal metabolic rate measurement error

This study estimated the reliability and the measurement error associated with measuring BMR in elderly men and the effect of within-day and day-to-day variance on measurement accuracy. The BMR of each patient was measured following a standardized protocol, with indirect calorimetry. Each patient was tested on two days within a seven-day period.; The subjects of this study were 35 long-term impatient rehabilitation patients from the Veterans Affairs Hospital, Houston Medical Center. A 12-minute data (24, 30-seconds intervals) were analyzed for both VO2 (ml/min) and BMR (kcal/day). A generalizabilfly design with two facets was used to calculated reliability coefficients and standard errors of measurement (SEM). A repeated measure ANOVA found no significant differences between days, among intervals. The mean VO2 for day 1 was 196 ml/min (ñ7.52) and for day 2 was 201.89 ml/min (±8.6). The mean BMR for day 1 was 1376.7 kcal/day (±311.7) and for day 2 was 1415.9 kcal/day (±360.1). The G-study found that about 32% the total variance associated with between subjects variation, 20% was associated with the interaction of subjects by days, and almost half of the variation was associated with the three-way interaction of subjects, days and intervals for both VO2 and BMR analysis.; D-study estimated the reliability and SEM for various sampling intervals. The sharp decrease in SEM for a single day slowed down after few minutes and reached a plateau at about 20-minutes. Increasing the number of intervals from the interval facet did not increase the G-coefficient more than 0.61. The reason is that the largest source of error was due to day-to-day variance. Increasing the number of days that BMR is measured increase the reliability further.; The SEM for various sampling durations and days can be used to recommend an empirically based measurement protocol that optimizes the accuracy of measuring BMR. An appropriate test protocol is for a 20-minute duration on a single day. This would yield a reliability estimate of about 0.60 with a standard error of about 210 kcal/day. If greater accuracy is needed, BMR will need to be measured on two or more days.