Please use this identifier to cite or link to this item:
Title: Construction of Anthropometric Growth Charts for Pakistan Using the LMS and Quantile Regression Approach
Authors: Asif, Muhammad
Keywords: Social Sciences
Issue Date: 2020
Publisher: Bahauddin Zakariya University Multan
Abstract: Construction of Anthropometric Growth Charts for Pakistan using the LMS and Quantile Regression Approach Muhammad Asif In the epidemiological area of study, physicians and clinicians are interested in monitoring growth of their people. Growth charts are considered as a renowned and worldwide applicable tool for doing this, elucidating how the distribution of growth measurement changes according to some time covariate for a particular population. A growth chart, actually consists of a series of smooth centile curves sowing how selected centiles of the growth measurement change when plotted against a time covariate i.e., age. These curves are based on representative sample from a reference population. For the present research work, a cross-sectional data on different anthropometric measurements i.e., height, weight, head circumference (HdC), neck circumference (NC), chest circumference (ChC), waist circumference (WC), head circumference (HC), midupper arm circumference (MUAC) and wrist circumference (WrC) of the Pakistani children and adolescents, aged 2-18 years, were obtained. The second sample was taken as secondary data from Pakistan Panel Household Survey-2010, consisting of data about weights, heights and other socio-demographic related information for the adult population of Pakistan. Both data sets were utilized for this research work. The main focus of the present research work is to construct the age-and gender-specific growth charts of above stated anthropometric measurements using the lambda-mu-sigma (LMS) and quantile regression approaches. Moreover, receiver operating characteristics (ROC) analysis was used to determine the optimal cut-off points of different anthropometric measures in order to identify the studied subjects with different levels of obesity. Initially, a total of 11,281 participants were measured in Sample-I (2-18 years old) and Sample-II contained 16,271 participants, aged 19 to 105 years. After removing the outliers in the data sets, nutritional indicators were studied. The results of these analyses vii showed that for overall sample of children and adolescents, aged 2 to 19 years, 2.3% of children were stunted (2.2% male and 2.3% female) and 2.0% children were thin. More female children were stunted in 2-5 years age-group. Thinness was higher (2.5%) in 2-5 years age-groups that decreased after five years age. In addition, 2.3% children were underweight in the overall children sample. More (3.1%) children were underweight in 25 years age-groups than the other ages. While for the adult sample, majority participants 5971(59.3%) were having normal weight. Overweight and obesity prevalence was reported to be 2296 (22.8%) and 512 (5.1%), respectively. For the pediatric sample, the mean height, weight, HdC and NC and MUAC of boys was significantly greater than that of girls with few exceptions. Whereas, the girls on average had more BMI than boys. Also, the MUAC and NC indicators both had robust diagnostics performance to discriminate children with or without elevated BMI (i.e., BMI > 2SD). By using the LMS and quantile regression techniques, a full set of smoothed percentile curves of all studied anthropometric parameters were also constructed by age and gender. Gender specific comparison of growth reference values showed that male children and adults had more values than females except for few ages. The growth charts constructed by both developed techniques were also compared with the raw percentiles to see which methods produces centiles with the best fit. The data structure shows that the both studying techniques is shown to produce the best fit, and these techniques have been used to construct the growth charts of different anthropometric characteristics by different researchers, worldwide. Comparison of our references with the WHO and USCDC derived growth references showed that our centiles are substantially below their corresponding centiles and verified the concept that growth references vary with geographical variations. In view of the difference between the Pakistani reference values and the WHO reference values, it is concluded that reference values presented here should be opted by local practitioners as a new reference values for the Pakistani population to perform high-quality clinical practices. viii Acronyms BF Breastfed BMI Body Mass Index ChC Chest Circumference CMR Child Mortality Rate FATA Federally Administered Tribal Area HdC Head Circumference HDI Human Development Index HPG Hypothalamic-pituitary-gonadal IDF International Diabetes Federation IMR Infant Mortality Rate LBW Low Birth Weight LMS Lambda-Mu-Sigma MUAC Mid-upper-arm Circumference NBF Non-breastfed NC Neck Circumference NCDs Non-Communicable Diseases NCHS National Centre for Health Statistics NHANES National Health and Nutrition Examination Survey NHSP National Health Survey of Pakistan NNS National Nutrition Survey PDHS Pakistan Demographic and Health Survey PHC Primary Health Care PIDE Pakistan Institute of Development Economics PNHDR Pakistan National Human Development Report PPHS Pakistan Panel Household Survey PSLM Pakistan Social and Living Standards Measurement SD Standard Deviation SES Socio-economic status UFM Under 5-years mortality USCDC United States’ Centers for Disease Control WC Waist Circumference WHO World Health Organization WHR Waist-to-hip Ratio WHtR Waist-to-height Ratio WrC Wrist Circumference
Gov't Doc #: 20298
Appears in Collections:PhD Thesis of All Public / Private Sector Universities / DAIs.

Files in This Item:
File Description SizeFormat 
Muhammad Asif (Statistics)-2020 BZU prr.pdfphd.Thesis3.86 MBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.