Accuracy Of Self-measured Height, Weight, Waist And Hip Circumference In Women With Musculoskeletal Diseases

Introduction
A variety of anthropometric measurements are routinely made in clinical settings [1]. However, self reported height,
weight, waist and hip circumference may be more convenient in large epidemiological studies [2]. This investigation
examined the accuracy of these self-reported measurements in comparison with the same variables measured by an
accredited anthropometrist, amongst women with established musculoskeletal diseases.

Methods
Self reported and measured height, weight, waist and hip circumference were obtained from 131 Caucasian women
aged 61.3y (6.1). Participants comprised healthy controls (n=35), or patients with established osteoporosis (n=36),
rheumatoid arthritis (n=35) or osteoarthritis (n=25). Subjects reported height and weight, and were given a flexible tape
to measure waist and hip circumference, following detailed instructions. Within a few days subjects had values remeasured
by an accredited anthropometrist following a standard protocol [3]. Derived variables of body mass index
(BMI = weight (kg) / height2 (m2)), waist to hip ratio and conicity index (CI = waist girth (m) / 0.109 Ö(body mass
(kg) /ht (m))), based on both self-reports and anthropometrist measurements were calculated. Error of self-reported
compared to measured values was quantified in inter-tester percent of technical error of measurement (%TEM). After
checking the data for normality, pearson correlation coefficients between self-reported and measured values and derived
indices were calculated, and paired samples t-tests were used to compare any differences in the means of the selfreported
and measured values.

Results
Pearson correlations between self-assessed and anthropometrist measurements were 0.92 for the hips, 0.94 for the waist
and height and 0.99 for the weight. For derived variables the highest agreement was observed for BMI (r = 0.98), while
WHR and CI were correlated less strongly (r = 0.69 and 0.66, respectively). All correlations were found to be
significant at the 0.01 level.

Discussion/ Conclusions
Although there were significant correlations between self-reported and measured values, large individual variations
were observed. Reliable inter-tester %TEM for trained technicians should be less than 1.5% [3], but weight, waist and
hip values exceeded this. The lack of a statistically significant difference between measured and self-reported values
does not render the measurements valid, as this may be due to relatively poor precision of self-reported measurements.
These findings indicate that direct anthropometric measurements should be performed whenever possible for optimal
measurements in clinical practice and research.

References
[1]. Engstrom J.L. et al. (2003) J Midwifery Womens Health, 48, 338-45.
[2]. Weaver T.W. et al. (1996) Int J Obes Relat Metab Disord 20, 644-50.
[3]. International Society for the Advancement of Kinanthropometry. International standards for anthropometric
assessment. Underdale, South Australia: ISAK. 2001

NOTA: O texto com a iconografia está no anexo.

Arquivo