Previous studies generally used cross-sectional data and focused on under-five children to assess the risk factors for malnutrition among Indian children. Some recent studies have reported that recovery from or faltering in malnutrition is possible after five years of age, but socio-demographic subgroup disparities have not been explored. This study aims to find the longitudinal disparity in height-for-age Z-scores (HAZ) and body-mass-index-for-age Z-scores (BMIAZ scores) across various sub-groups of a cohort from childhood to adolescence. This study used a cohort from the Young Lives Survey, which followed children aged of 1-15 years between 2002 and 2016-17 in the states of Andhra Pradesh and Telangana, India. Mixed-effect models were applied to find the main, time, and interaction effects of HAZ scores and BMIAZ scores. In addition, an extended Kitagawa-Oaxaca-Blinder decomposition approach to assess group-based differences over time was used. The cross-sectional prevalence of stunting reduced across all subgroups, while thinness rose during the same period. The interactions of child sex, mother's education, place of residence, wealth index, and antenatal care with time were statistically significant at p <0.05. The gender disparity in adjusted HAZ score decreased from 0.214 units at 1 year to 0.011 units at 15 years, whereas BMIAZ score differential increased from 0.106 to 0.538 units over same timeframe. Disparities in scores were also observed across rural-urban, maternal education, social group, religion, socioeconomic status, maternal age at birth, antenatal care, and premature birth status. The study sheds light on the nuanced dynamics of paediatric growth, emphasising the importance of longitudinal approaches in understanding and addressing the health disparities across different stages of childhood and adolescence.