Perceived weight versus Body Mass Index among urban Aboriginal Australians: do perceptions and measurements match?
Associate Professor Joan Cunningham, Menzies School of Health Research, PO Box 41096, Casuarina, Northern Territory 0811. Fax: (08) 89275187; e-mail: firstname.lastname@example.org
Objective: To examine the relationship between perceived body weight and measured Body Mass Index (BMI) among urban Aboriginal Australian adults.
Methods: We compared responses to a question on perceived weight with BMI based on measured health and weight among 248 Aboriginal volunteers aged ≥15 years who took part in a larger health study in the Darwin area between September 2003 and March 2004. Logistic regression was used to examine associations between socio-economic, demographic and cultural factors and under-assessment of weight.
Results: Being male and having diabetes were significantly associated with under-assessment of weight. Despite under-assessment being common, most participants with a BMI ≥25 – and almost all (>90%) those with BMI≥25 plus high waist circumference – described themselves as overweight.
Conclusions: Study participants with BMI≥25 were generally able to classify themselves appropriately as overweight.
Implications: Lack of awareness of weight is unlikely to represent a major barrier to engaging Aboriginal people. However, other barriers exist, and both individual-level and environmental/structural approaches are required to reduce the burden of obesity among Aboriginal Australians.
Obesity is a growing problem in Australia and one especially prevalent in the Indigenous population.1,2 In 2004/05, 57% of Indigenous people aged ≥15 years in the National Aboriginal and Torres Strait Islander Health Survey were classified as overweight or obese based on self-reported height and weight. After adjusting for age differences, Indigenous adults were about 20% more likely than other Australians to be overweight/obese.2
Obesity is an important risk factor for a range of conditions affecting Indigenous Australians, including cardiovascular disease and diabetes. Together, these two conditions were responsible for an estimated 37% of excess mortality among Indigenous males and 43% of excess mortality among Indigenous females in 1999-2003.3
Reducing the burden of obesity will require environmental/structural changes as well as behavioural changes at the individual level.4 Successfully influencing behaviour is likely to depend in part on individuals’ ability to accurately judge their own body size in order to respond appropriately to targeted public health campaigns.
Studies elsewhere have suggested that misperception of body weight is relatively common and is related to demographic and socio-economic characteristics.5–8 However, little is known about Indigenous Australians’ perceptions of their body weight.
We examined the relationship between perceived weight and measured Body Mass Index (BMI) (specifically, under-assessment of weight) among urban Aboriginal Australian adults who took part in the Darwin Region Urban Indigenous Diabetes (DRUID) Study.
The DRUID study has been described in detail elsewhere.9 Briefly, eligible participants were aged ≥15 years, identified as Aboriginal and/or Torres Strait Islander, had lived in a defined geographic region in the Darwin area for six months or more, and did not live in an institutional dwelling. Eligible participants who gave consent underwent a health examination including blood and urine collection, clinical and anthropometric measurements, and questionnaire administration.
Body weight was recorded to the nearest 0.1kg using a digital scale weighing up to 200kg (Model 767, Seca Deutschland, Hamburg, Germany). Height was recorded to the nearest 0.1cm using a portable stadiometer (Model PE87, Mentone Educational Centre, Moorabbin, Victoria). BMI was based on measured weight and height (kg/m2) and categorised as: underweight (BMI<18.5); normal (≥18.5 to <25); overweight (≥25 to <30); obese (≥30).
Perceived weight was based on the participant's response to the following question: “With regard to your weight, do you consider yourself to be: underweight; the right weight; slightly overweight; very overweight.” These categories were interpreted as equivalent to BMI categories underweight, normal, overweight and obese, respectively.
Waist circumference was measured to the nearest 0.1cm using a two-metre non-stretch fiberglass tape. High waist circumference was defined as ≥88cm (females) and ≥102cm (males).
Information on several demographic, socio-economic, health and cultural variables was also obtained. Those considered in the analysis include: sex; age group; full-time employment status; housing tenure; post-school educational qualifications; private health insurance; current smoking status; current diabetes status (based on oral glucose tolerance test); self-assessed health status; and identification with a clan, tribal or language group.
Data on perceived weight were collected during the first six months of the study (September 2003-March 2004). During this period, 364 people participated in the study, of which 321 completed the relevant questionnaire. Of these, 23 participants were excluded from analysis because they were pregnant (n=2), missing measured height/weight (n=16), or missing perceived weight (n=5). Another 50 people were excluded because they identified as either Torres Strait Islander only (n=11) or both Aboriginal and Torres Strait Islander (n=39). Preliminary analysis suggested that the relationship between perceived weight and measured BMI differed for Aboriginal and Torres Strait Islander participants, but there were too few Torres Strait Islander participants to allow separate analysis. The present analysis thus includes 248 non-pregnant Aboriginal participants with information on BMI and perceived weight.
All analyses were performed using STATA version 9 (Stata Corporation, College Station, TX). Pearson chi-squared tests were used to compare cross-tabulated categorical variables. Simple and multiple logistic regression models were used to assess associations between selected socio-economic, demographic, health and cultural variables and the dependent variable of interest: under-assessing one's weight (defined as having a perceived weight category lower than one's measured BMI).
The study was approved by the Human Research Ethics Committee of the Northern Territory Department of Health and Community Services and Menzies School of Health Research. It was considered and approved by both the Aboriginal subcommittee and the main committee. The study's governance structure included an Indigenous Steering Group and partnerships with key Indigenous organisations.9
The 248 participants included in the analysis ranged from 15-81 years (median=42) and the majority were female (67%). Mean BMI (±SD) was 27.4 (±5.1) for males and 29.8 (±7.8) for females. Mean waist circumference was 98.4cm (±13.3) for males and 98.0cm (±17.4) for females.
A similar proportion of males (72%) and females (71%) was classified as overweight or obese based on measured BMI, but females were more likely than males to be in the obese category (see Table 1). The majority of participants (59% of males and 74% of females) classified themselves as either ‘slightly overweight’ or ‘very overweight’, but there was some disagreement between measured BMI and perceived weight among individuals. Forty-one per cent of males under-assessed their weight, 10% over-assessed their weight, and 49% had agreement between perceived weight and measured BMI. Among females, 27% under-assessed their weight, 14% over-assessed, and 59% were in agreement. Most people who under-assessed their weight were only one category below their measured BMI category; only seven people (2.8%) were more than one category lower. Males with an obese BMI were more likely to describe themselves as ‘slightly overweight’ than ‘very overweight’; this was not true for females (see Table 1).
Table 1. Perceived weight and measured Body Mass Index (BMI) among 248 urban Aboriginal adults.
|Males||n=23 (28.4%)||n=34 (42.0%)||n=24 (29.6%)||n=81 (100.0%)|
| Slightly overweight||8.7||61.8||62.5||46.9|
| Very overweight||0.0||8.8||29.2||12.4|
|Females||n=49 (29.3%)||n=44 (26.3%)||n=74 (44.3%)||n=167 (100.0%)|
| Slightly overweight||34.7||72.7||39.2||46.7|
| Very overweight||0.0||4.5||58.1||27.0|
Despite discrepancies between perceived weight and measured BMI, most males (79%) and females (90%) with BMI ≥25 described themselves as overweight (whether ‘slightly’ or ‘very’). Among those with BMI≥25 plus a high waist circumference, 91% of males and 92% of females described themselves as overweight.
In logistic regression modelling, males were significantly more likely than females to under-assess their measured BMI, as were participants with diabetes (see Table 2). For all other factors examined, odds ratios were neither large (generally between 0.7 and 1.2) nor statistically significant, either before or after adjustment for sex and diabetes.
Table 2. Relative odds of under-estimating measured BMI among urban Aboriginal participants (n=248).
|Male||32.7||1.9 (1.1-3.3)||–||2.6 (1.4-4.9)|
|Age group (years)|| || || || |
| 15-24||14.1||1.1 (0.4-2.8)||0.8 (0.3-2.3)|| |
| 25-34||18.6||1.2 (0.5-2.9)||1.0 (0.4-2.6)|| |
| 35-44||24.6||1.0||1.0|| |
| 45-54||23.8||1.3 (0.6-2.9)||1.1 (0.5-2.6)|| |
| 55-64||12.1||2.5 (1.0-6.2)||2.0 (0.7-5.3)|| |
| 65+||6.8||1.2 (0.4-3.8)||0.7 (0.2-2.6)|| |
|In full-time employment||42.7||0.7 (0.4-1.2)||0.9 (0.5-1.7)|| |
|Housing owned/being purchased by occupants||45.6||0.8 (0.4-1.3)||0.8 (0.5-1.5)|| |
|Has any post-school qualifications||44.8||0.8 (0.5-1.4)||1.0 (0.6-1.8)|| |
|Has private health insurance||25.8||0.8 (0.4-1.5)||0.8 (0.4-1.6)|| |
|Has diabetes||25.0||2.1 (1.1-3.8)||–||2.1 (1.0-4.4)|
|Current smoker||45.3||1.0 (0.6-1.8)||1.1 (0.6-1.9)|| |
|Self-assessed health status poor/fair||28.4||1.0 (0.6-1.9)||0.9 (0.5-1.8)|| |
|Identifies with a clan, tribal or language group||72.5||1.1 (0.6-2.1)||1.3 (0.6-2.5)|| |
Under-assessment of body weight was relatively common among Aboriginal participants in this study, with males and those with diabetes significantly more likely to under-estimate BMI. However, most people with BMI≥25 – and almost all those with high BMI plus high waist circumference – appropriately described themselves as overweight. Given that relatively modest weight loss can reduce diabetes risk,10–12 recognising oneself as being overweight at all is arguably more important than acknowledging the extent to which one is overweight.
The greater likelihood of males to under-assess their weight is consistent with previous studies in the United States (US).5–7 For example, about 12% of obese men and 42% of overweight men in one study believed they were a healthy weight/underweight; corresponding figures for women were 5% and 18%.5
The degree of under-assessment observed in the present study is lower than that previously seen in the general Australian population. Donath found that 51% of males and 28% of females with BMI≥25 in the 1995 National Health Survey did not consider themselves overweight.8 The corresponding figures in the present study were 21% and 10%, respectively. It is not clear whether this disparity reflects changes over time (such as increased awareness of obesity), differences in study methodology or differences in perceptions for Aboriginal and non-Aboriginal Australians. Several US studies have found variation by ethnicity in the relationship between perceived weight and BMI.5–7
Measured BMI has been used in this study as a gold standard, but BMI categorisation is problematic for some body types and ethnic groups, including Aboriginal Australians. A given BMI is associated with a higher proportion of body fat for Aboriginal people from remote communities than for Australians of European descent.13 Lower BMI cut-offs for overweight/obesity have been recommended for some groups, including various Asian populations.14 Although the present analysis was limited to those who identified as Aboriginal but not Torres Strait Islander, it must be recognised that Aboriginal Australians represent a genetically and ethnically diverse population. The appropriate cut-offs for overweight and obesity for the participants in this study are unknown.
The present study was based on a subset of early responders in a larger study. Participants are not representative of the study as a whole and are not likely to be representative of the source population. While estimates of prevalence may therefore be biased, this does not necessarily compromise the internal validity of comparisons between variables.
Preventing and reducing obesity among Indigenous Australians is an important public health challenge. Despite under-assessment of BMI being relatively common, most people who were classified as overweight or obese recognised themselves as overweight. This suggests that a lack of awareness of one's weight is unlikely to represent a major barrier to engaging Aboriginal people, although special efforts may be required for men and people with diabetes. However, other barriers remain. While individual-level approaches are needed to manage obesity and minimise related health risks, societal-level environmental/structural changes are likely to have greater potential for the prevention of obesity in the longer term.4
The authors gratefully acknowledge the support of DRUID study participants, study staff, members of the Indigenous Steering Group, and partner organisations. The DRUID Study was funded by the National Health and Medical Research Council (NHMRC Project Grant /236207), with additional support from the Australian Government Department of Employment and Workplace Relations, the Clive and Vera Ramaciotti Foundation, the Vincent Fairfax Family Foundation, the AusDiab Partnership in Type 2 Diabetes, and Bayer HealthCare. The DRUID Study is an in-kind project of the Cooperative Research Centre for Aboriginal Health. Joan Cunningham was supported by a NHMRC Career Development Award (#283310). Funding sources played no role in the study design, in the collection, analysis and interpretation of the data, in the writing of the manuscript, or in the decision to submit the manuscript for publication.