Background Few research studies have been conducted designed specifically to examine the correlates of physical activity behaviour for people with RA. Most of this work has been conducted assessing the psychosocial correlates of physical activity in this population, with perceived benefits and self efficacy shown as strong predictors of higher physical activity levels. However, the impact of demographic or RA specific health related factors has been less well investigated despite the frequent assessment of demographic factors in other populations.
Objectives The aim of this study was to examine demographic and RA specific health related correlates of energy expenditure in the RA population and determine a model which best predicts energy expenditure behaviour in this population.
Methods Analysis was conducted on 59 RA subjects, who were recruited from outpatient rheumatology clinics at the Mid Western Regional Hospitals, Limerick. All subjects wore a SenseWear Armband on the right upper arm and energy expenditure was estimated over seven consecutive days. This tool has previously been proven to be valid to measure energy expenditure in the RA population.
The following demographic variables were recorded:
Gender, Smoking status, Employment status, Age, Body mass index (BMI).
The following RA specific health related variables were recorded:
Medication use (Biologic use, DMARD use, NSAID use, steroid use), Tender and Swollen Joint Count, Visual Assessment Score for Global Health, Health Assessment Questionnaire and Disease Activity Score-28.
Statistical Analyses was conducted using SPSS v.18 for Microsoft Windows. Normality was assessed by use of boxplots, histograms, Quantile-Quantile plots and Kolmogorov-Smirnov statistic where appropriate. A number of variables were found to be non normally distributed and thus were transformed to a normal distribution to allow for parametric analyses.
In order to determine if a relationship existed between the demographic and RA specific health factors and the amount of energy expended, point biseral correlations for dichotomous variables and Pearson’s correlation coefficients for continuous variables were conducted.
In order to develop a subset of independent variables which best predicted each of the dependent variables statistical (stepwise) multiple regression analyses was utilised. Significant (p<0.05) contributors identified through bivariate analyses were introduced into a stepwise multiple linear regression analysis.
Results This study found that consistently moderate (r = 0.30 - 0.49) to strong (0.50 - 1.00) relationships existed between age, gender, BMI and employment status and energy expenditure.
The model of best fit for Total Energy Expenditure included the variables gender, age, BMI and employment status and had an adjusted r2 value of 0.679. The variables which produced the model of best fit for Energy Expenditure related to Physical Activity included gender, employment, smoking status and average HAQ score and had an adjusted r2 value of 0.557.
Conclusions The present study has identified a number of significant correlates of energy expenditure in the RA population as well as determining a model of best fit for both total energy expenditure and energy expenditure related to physical activity in the RA population.
Future interventions aimed at increasing energy expenditure levels in this population group should be informed by evidence of the present study.
Acknowledgements Marie Tierney is supported in her postgraduate studies by scholarships sponsored by the Irish Research Council, Intel Ireland and the Central Remedial Clinic.
Disclosure of Interest None Declared