T. G. Lohman

The purpose of this study was to assess the predictability of body fatness as determined by three separate approaches: (1) body density from hydrostatic weighing and functional residual volume alone (PFD); (2) body density and water (PFDW); (3) body density, water and bone mineral (PFDWB). Bone mineral was estimated from radius and ulna measurements using photon absorptiometry, and total body water was estimated using deuterium oxide dilution. N ine skinfolds were taken with a Harpenden caliper. A 4 X 2 X 2 factorial plan was used representing four maturation groups (based on secondary sex characteristics) males and females, black and white subjects. The sample consisted of 310 subjects children and adults ranging in age from 8-29 years. In the prepubescent and pubescent groups of children, systematic differences were found among methods with the PFD alone producing higher mean density values than the other two estimates. Constants were fitted by the least squares method in a step down multiple regression analysis to determine which skinfold measurements most accurately predicted PFDWB. The systematic effects of racial group, sex, and maturation were also included in the analysis. The R2 and SEE from sum of triceps and calf and sum of triceps and calf squared plus design variables in prepubescent group were substantially lower for each of the three PFD estimates compared to the other three groups. When PF estimate was determined from PFDWB rather than PFD alone the R2 increased from 62% to 77% and the SEE dropped from 4.5 to 3.9%, lending support to the concept that constants used to estimate fat in adults tend to overestimate body fatness in children. Curvilinear equations for the prediction of PFDWB are presented for triceps and subscapular skinfolds and linear equations for triceps and calf skinfolds taking into account the effects of maturation level, race and sex. These equations may provide more accurate estimates of PF than those currently available, since they take into consideration the use of a multicomponent approach to body composition and account for the chemical immaturity of children.

Prompted by the large heterogeneity of individual results in obesity treatment, many studies have attempted to predict weight outcomes from information collected from participants before they start the programme. Identifying significant predictors of weight loss outcomes is central to improving treatments for obesity, as it could help professionals focus efforts on those most likely to benefit, suggest supplementary or alternative treatments for those less likely to succeed, and help in matching individuals to different treatments. To date, however, research efforts have resulted in weak predictive models with limited practical usefulness. The two primary goals of this article are to review the best individual-level psychosocial pre-treatment predictors of short- and long-term (1 year or more) weight loss and to identify research needs and propose directions for further work in this area. Results from original studies published since 1995 show that few previous weight loss attempts and an autonomous, self-motivated cognitive style are the best prospective predictors of successful weight management. In the more obese samples, higher initial body mass index (BMI) may also be correlated with larger absolute weight losses. Several variables, including binge eating, eating disinhibition and restraint, and depression/mood clearly do not predict treatment outcomes, when assessed before treatment. Importantly, for a considerable number of psychosocial constructs (e.g. eating self-efficacy, body image, self-esteem, outcome expectancies, weight-specific quality of life and several variables related to exercise), evidence is suggestive but inconsistent or too scant for an informed conclusion to be drawn. Results are discussed in the context of past and present conceptual and methodological limitations, and several future research directions are described.

The purposes of this study were to develop and cross-validate the "best" prediction equations for estimating fat-free body mass (FFB) from bioelectrical impedance in children and youth. Predictor variables included height2/resistance (RI) and RI with anthropometric data. FFB was determined from body density (underwater weighing) and body water (deuterium dilution) (FFB-DW) and from age-corrected density equations, which account for variations in FFB water and bone content. Prediction equations were developed using multiple regression analyses in the validation sample (n = 94) and cross-validated in three other samples (n = 131). R2 and standard error of the estimate (SEE) values ranged from 0.80 to 0.95 and 1.3 to 3.7 kg, respectively. The four samples were then combined to develop a recommended equation for estimating FFB from three regression models. R2 and SEE values and coefficients of variation from these regression equations ranged from 0.91 to 0.95, 2.1 to 2.9 kg, and 5.1 to 7.0%, respectively. As a result of all cross-validation analyses, we recommend the equation FFB-DW = 0.61 RI + 0.25 body weight + 1.31, with a SEE of 2.1 kg and adjusted R2 of 0.95. This study demonstrated that RI with body weight can predict FFB with good accuracy in Whites 10-19 yr old.

Social-ecological (SE) models are becoming more widely used in health behavior research. Applying SE models to the design of interventions is challenging because models must be tailor-made for each behavior and population, other theories need to be integrated into multi-level frameworks, and empirical research to guide model development is limited. The purpose of the present paper is to describe a SE framework that guided the intervention and measurement plans for a specific study. The trial of activity for adolescent girls (TAAG) is a multi-center study of interventions to reduce the decline of physical activity in adolescent girls. The TAAG framework incorporates operant learning theory, social cognitive theory, organizational change theory and the diffusion of innovation model in a multi-level model. The explicit and practical model developed for TAAG has already benefited the study and may have elements that can generalize to other health promotion studies.