Banca de DEFESA: MARCOS FELIPE SILVA DE LIMA

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
DISCENTE : MARCOS FELIPE SILVA DE LIMA
DATA : 22/07/2019
HORA: 14:30
LOCAL: Departamento de Nutrição
TÍTULO:

Assessment of nutritional status, development and validation of equations for the estimation of weight and height among elderly people living in nursing homes.

PALAVRAS-CHAVES:

Cross-sectional studies; Validation studies; Health of the elderly; Nutritional status; Nutrition assessment; Anthropometry

PÁGINAS: 119
GRANDE ÁREA: Ciências da Saúde
ÁREA: Saúde Coletiva
SUBÁREA: Epidemiologia
RESUMO:

Malnutrition in elderly people is related to the fragility, multimorbidity and mortality. In order to identify the risk of deficits early, anthropometric methods based on weight, height, perimeters and skinfolds can be used, which allow for the evaluation of anthropometric indicators. When it is not possible to measure weight and height, estimative equations can be used from these measures. The objective of this study was to evaluate the anthropometric nutritional status, to develop and validate equations for estimating weight and height in elderly living in nursing homes. The study was conducted with elderly people living of nursing homes in Brazil. Anthropometric data (weight, height, body perimeters and skinfolds) were collected for each participant. For analyze the anthropometric nutritional status Principal Component Analysis stratified by sex was performed and the factorial scores of the chosen model were evaluated in relation to the age group, type of nursing home, racial/ethnic identity, schooling, burden of disease and functional capacity. Methods of weight and height estimation were elaborated by linear multiple regression. The regression models developed considered statistical reliability criteria, such as the coefficient of determination (R²), the standard error of the estimate and the Akaike Information Criterion (AIC). The prediction equations were validated by concordance tests such as the Intraclass Correlation Coefficient (ICC) and its respective confidence interval (95% CI). For all analyzes, p values <0.05 were considered statistically significant. Regarding the analysis of main components, the extracted components were denominated "Anthropometric Nutritional State (ANS)" and "Stature (S)". It was verified that the elderly people living in non-profit nursing home (male ANS = -0.855, female ANS = -0.952, male S = -0.919, female S = -0.711), non-white color (male ANS = -0.866, female ANS = -0,816, male S = -0,783; female S = -0,700), low level of schooling (male ANS = -0.973, female ANS = -0.931, male E = -0.846, female E = -0.692), and with a restriction of mobility (male ANS = -0,855; female ANS = -1,076) were the ones that presented the lowest score in the evaluation by the joint indicators. In relation to the development of weight equations, we developed models of predictive equations broken down by the perimeter of the arm, another model adjusted to the statistical criteria, and a third model of equations with measures that are easy to obtain in bedridden or wheelchair-bound elderly. The most appropriate model is calf circumference, knee height, waist circumference, subscapular skinfold, age and sex (ICC = 0.876). Regarding the development of equations of height estimation, models were developed with measurements of knee height, hemi-span, demi-span or ulna length. The most adjusted model uses knee height, age and sex (ICC = 0.863). The components "Nutritional Status" and "Stature" were extracted, which had relation with, especially, the restriction of mobility. The development of weight and height estimation methods followed statistical criteria and the convenience of collecting the component variables of the models.

MEMBROS DA BANCA:
Presidente - 2149611 - CLELIA DE OLIVEIRA LYRA
Interno - 277398 - KENIO COSTA DE LIMA
Externa ao Programa - 1891751 - URSULA VIANA BAGNI
Externo à Instituição - LEANDRO DE ARAUJO PERNAMBUCO - UFPB
Externo à Instituição - RODRIGO PINHEIRO DE TOLEDO VIANNA - UFPB