Journal of Intellectual Disability Research 134
doi: 10.1111/j.1365-2788.2004.00617.x
pp – Blackwell Science, LtdOxford, UKJIRJournal of Intellectual Disability Research-Blackwell Science Ltd, 2134143Original ArticleUnderweight,
obesity and exerciseE. Emerson
Underweight, obesity and exercise among adults with intellectual disabilities in supported accommodation in Northern England E. Emerson Institute for Health Research, Lancaster University, Lancaster, UK
Abstract Background Significant deviation from normal weight (obesity and underweight) and lack of physical exercise have been identified as three of the most significant global behavioural risks to health. Methods Body mass index (BMI) and levels of physical activity were measured in a sample of adults with intellectual disabilities (ID) receiving supported accommodation in nine geographical localities in Northern England. Comparative population data were extracted from the Health Survey for England and . Results Men and women with ID living in supported accommodation are at increased risk of being significantly underweight and physically inactive. Women with ID living in supported accommodation are at increased risk of obesity. Within the population of people with ID living in supported accommodation increased behavioural health risks are associated with gender, severity of ID, age and location. Conclusion Significant deviation from normal weight and lack of physical exercise are significant behavioural risks to health among people with ID.
Correspondence: Eric Emerson, Institute for Health Research, Lancaster University, Lancaster LA YT, UK (e-mail:
[email protected]).
© Blackwell Publishing Ltd
Keywords obesity, physical activity, underweight
Introduction Increasing concern has been expressed about mortality, morbidity and the behavioural determinants of health among people with intellectual disabilities (ID) (Prasher & Janicki ; Sutherland et al. ; Walsh & Heller ). Significant deviation from normal weight (obesity and underweight) and lack of physical exercise have been identified as three of the most significant global behavioural risks to health (World Health Organization ; see also Ezzati et al. ). To date, however, relatively little is known about either the prevalence of obesity, underweight or physical activity levels or the personal and environmental risk factors associated with these health behaviours among people with ID. The available evidence does suggest, however, that people with ID may be at increased risk for each of these three behavioural determinants of physical ill-health. Deviations from ‘ideal’ weight are most commonly defined in terms of body mass index (BMI). BMI is calculated by dividing weight (in kg) by the square of height (in metres). BMI scores are conventionally categorized as: underweight (BMI < ); normal (BMI –); overweight (BMI .–); obese (BMI > ). It has been consistently reported
Journal of Intellectual Disability Research 135 E. Emerson • Underweight, obesity and exercise
that women, and men, with ID are at significantly greater risk of obesity (Gravestock ). Using BMI-based definitions Beange et al. (b) reported prevalence rates of obesity of % for men and % for women among a community-based sample of adults with ID in New South Wales, Australia (compared with % of men and % of women in the general population). More recently, Hove () reported prevalence rates of obesity of % for men and % for women among a community-based sample of adults with ID in western Norway (compared with % of men and % of women in the general population). Robertson et al. () reported prevalence rates of obesity of % for men and % for women among a sample of adults with ID in various forms of supported accommodation in the UK (compared with % of men and % of women in the general population at that time: National Centre for Social Research ). Increased prevalence rates of BMIdefined obesity, although no figures were given, have also been reported among a community-based samples of adults in the UK by Wells et al. () (data also reported in Martin et al. ). Using a different cut-off point for obesity (. for men, . for women), Rubin et al. () reported prevalence rates for obesity of % of men and % of women in a community-based North American sample of adults with Downs syndrome (compared with general population rates of % among men and % among women). Alternative approaches to defining obesity are based on percentage deviation from ideal weight. Studies using such an approach have reported increased prevalence rates for obesity among school children with developmental disabilities in Japan (Takeuchi ) and among women, but not men, with ID living in group homes or with their families in North America (Rimmer et al. ). Among adults with ID, increased risk of obesity has been associated with: female gender (Rimmer et al. , ; Beange et al. b; Fujiura et al. ; Rubin et al. ; Robertson et al. ; Hove ); less severe ID (Rimmer et al. ; Robertson et al. ; Hove ); less restrictive living environment (Rimmer et al. ); living in a family setting (Rubin et al. ); lower senior staffing ratios (Robertson et al. ); and Down’s syndrome (Robertson et al. ; Hove ).
Much less attention has been paid to the behavioural health risk of being significantly underweight. The data that are available, however, suggest that certain groups of people with ID may be at increased risk of being underweight (BMI < ). Hove () reported prevalence rates of underweight of % for men and % for women among a community-based sample of adults with ID in western Norway (compared with % of men and % of women in the general population). Robertson et al. () reported prevalence rates of underweight of % of men and % of women among a sample of adults with ID in various forms of supported accommodation in the UK (compared with % of men and % of women in the general population: National Centre for Social Research ). Beange et al. (a) identified % of an institutionalized sample of adults with profound multiple disabilities as being underweight. Similarly, Kennedy et al. () reported that % of men and % of women in a sample of institutionalized people with ID who could feed independently were underweight. However, no clear differences in prevalence rates for being underweight between people with and without ID are apparent in the data presented visually by Wells et al. (). In general the available evidence suggests that adults with ID are significantly more likely to lead sedentary lifestyles than adults who do not have ID. Beange et al. (b) reported that % of men and % of women in their community-based sample had not engaged in any moderate to vigorous exercise in the preceding weeks (compared with % of men and % of women in the general population). Wells et al. () reported that % of people with ID in their community-based sample had not engaged in any moderate to vigorous exercise in the preceding month (compared with just % of the general population). In the UK, physical inactivity has been defined as participating in activity of moderate or vigorous intensity less than times in weeks (a level of inactivity that represents a risk factor for cardiovascular disease: Bennett et al. ). According to this definition, % of men and % of women in England were reported to be inactive in (National Centre for Social Research ). Using this definition, reported rates of inactivity among people with ID have ranged from –% (Turner ; Messent et al.
© Blackwell Publishing Ltd, Journal of Intellectual Disability Research , –
Journal of Intellectual Disability Research 136 E. Emerson • Underweight, obesity and exercise
) to –% (Robertson et al. ). In contrast, Draheim et al. () reported that levels of exercise among a community-based sample of adults with mild to moderate ID were similar to those found in a nonintellectualy disabled comparison group (see also Temple et al. ). Within the population of people with ID increased inactivity has been associated with more severe ID (Robertson et al. ), type of living situation (Rimmer et al. ; Robertson et al. ) and older age (Robertson et al. ). However, the majority of these studies may be criticized on the grounds of relatively small and/or unrepresentative samples (for an exception see Beange et al. b) and their reliance on simple bivariate analytic procedures (for an exception see Robertson et al. ). The aim of the present paper is to present data derived from a relatively large-scale study of weight and exercise undertaken among adults with ID in supported accommodation in Northern England.
Method Sampling The data were collected between and across nine geographical localities in Northern England in the context of audit-based reviews of the quality of supported accommodation (Bliss et al. ). Supported accommodation included all forms of support provided to enable people with ID to live outside of their family home. It included examples of supported living, group homes, hostels and village communities. Of the nine areas, six fell within the % most socially deprived districts in England, two within the –% most deprived and two within the –% most deprived (Department of the Environment, Transport, and the Regions ). Within each area sampling strategies were determined by local managers. These included sampling all people with ID receiving supported accommodation from a particular provider organization and a variety of random and nonrandom sampling strategies. Information was collected on a total of adults with ID. The number of participants per locality ranged from to . Given sampling was undertaken by local organizations in the context of service audit, it was not possible to determine response rate.
Measures NW Audit of Quality in Residential Supports (Bliss et al. ) involves the collection of basic information from a key informant (e.g. keyworker, first line manager) on the characteristics of people with ID and the nature of the supported accommodation they receive prior to a visit by an external audit team. The previsit information includes: height and weight (from which BMI is calculated), the Physical Activity Scale used in the Health Survey for England between and (e.g. Bennett et al. ) and the Learning Disability Casemix Scale (Pendaries ) to measure overall level of adaptive and challenging behaviour. Height and weight data were extracted from existing service records. The Physical Activity Scale collects information on the number of times the participant engaged in moderate to vigorous physical activity during the weeks preceding interview. Sports and exercise activities were only included if they lasted for min or more. All other moderate or vigorous activities were included irrespective of duration. The Health Survey for England defines people as ‘inactive’ if they have engaged in fewer than bouts of moderate or vigorous activity in the preceding weeks. The Learning Disability Casemix Scale (Pendaries ) is a simple -item behaviour rating scale from which levels of adaptive behaviour (most able, moderately able, least able) and challenging behaviour (no/little, moderate, severe) can be derived. The scale has acceptable psychometric properties including good convergent validity with the AAMR Adaptive Behavior Scale (Nihira et al. ) and acceptable levels of inter-informant and test– retest reliability (Comas-Herrera et al. ).
Comparative data In order to make comparisons with adults who do not have ID, data on BMI were obtained from the Health Survey for England (National Centre for Social Research ) and data on physical activity were obtained from the Health Survey for England (National Centre for Social Research ), the last year in which the Physical Activity Scale was used in the form used in the present study. The Health Survey for England contains information on BMI for a nationally representative sample of people in England aged or over. The Health Survey
© Blackwell Publishing Ltd, Journal of Intellectual Disability Research , –
Journal of Intellectual Disability Research 137 E. Emerson • Underweight, obesity and exercise
for England contains information on physical activity for a nationally representative sample of people in England aged or over.
Results Participants Information on the characteristics of the participants and the nature of the supported accommodation they received is presented in Table .
Obesity Information on BMI was available for (%) of participants. Overall, % of participants were
underweight, % overweight and % obese. These data (and comparative data from the Health Survey for England : National Centre for Social Research ) are presented in Table . For adults with ID, prevalence rates of obesity were greater for women at the – years (Fisher’s exact P = .), – years (Fisher’s exact P = .) and – years (Fisher’s exact P = .) age groups than for men. When compared with men in the general population, prevalence rates for obesity were lower among men with ID only within the – years age group (z = ., P < .). When compared with women in the general population, prevalence rates for obesity were higher among women with ID within – years (z = ., P < .) and – years age
Table 1 Characteristics of participants and settings
n Age (years) (3% missing data; mean = 49.3, SD = 15.5) 16–24 25–34 35–44 45–54 55–64 65–74 75+ Gender (2% missing data) Men Women Ethnicity (2% missing data) White South Asian Black Other Adaptive behaviour (10% missing data) Most able Moderately able Least able Challenging behaviour (26% missing data) No/little Moderate Severe Sensory impairment (33% missing data) Yes Epilepsy (4% missing data) Yes
%*
72 199 337 328 298 164 93
5 13 23 22 20 11 6
824 693
54 46
1485 9 10 10
98 1 1 1
504 443 441
36 32 32
646 377 143
55 32 12
221
18
434
29
n Size of setting (number of co-residents; 7% missing data) 1 2–3 4–6 7–9 10 + Type of setting (12% missing data) Participant home owner Participant holds tenancy Small residential home Residential home (4 + places) Nursing home NHS provision Adult placement Dispersal/location (9% missing data) Dispersed in community Campus/cluster housing Type of prior residence (6% missing data) Residential child setting Family Group home Hostel Residential/village community Institution Other
NHS, National Health Service. * Because of rounding errors some percentages may add up to slightly more or less than %.
© Blackwell Publishing Ltd, Journal of Intellectual Disability Research , –
%*
66 494 743 57 74
5 35 52 4 5
5 835 102 270 53 42 52
