The prediction model of social alienation trajectory in adolescent idiopathic scoliosis patients based on social ecosystem theory: a prospective cohort study

Social alienation can adversely affect the psychological well-being, social functioning, and longer-term adjustment of adolescents with idiopathic scoliosis (AIS). However, longitudinal patterns of social alienation after diagnosis and prediction frameworks integrating multilevel factors remain insufficiently studied. This study aimed to identify distinct social alienation trajectories during the first year after AIS diagnosis and to develop a prediction model for trajectory-group membership. In this single-centre prospective cohort study, 276 patients with AIS completed assessments at diagnosis and at 3, 6, and 12 months. Growth mixture modelling (GMM) was used to identify social alienation trajectories. Trajectory classes were subsequently grouped to define a binary outcome for prediction modelling. Candidate predictors underwent univariable screening followed by forward likelihood-ratio logistic regression. Internal validation was performed using 1000 bootstrap resamples, with the selected predictor set held fixed and the regression coefficients re-estimated in each resample. Three trajectories were identified: lower-score increasing ( n = 144), moderate-score increasing ( n = 91), and persistently higher-score ( n = 41). The latter two classes were combined as the elevated-score trajectory group ( n = 132), with the lower-score increasing class as the reference. The final model retained seven predictors: age > 12 years (OR = 6.461; 95% CI, 2.019–20.678), presence of comorbid health conditions (OR = 8.535; 95% CI, 1.803–40.394), positive self-esteem (OR = 0.876; 95% CI, 0.771–0.995), negative self-esteem (OR = 1.243; 95% CI, 1.117–1.383), intra-family support (OR = 0.829; 95% CI, 0.747–0.920), extra-family support (OR = 0.891; 95% CI, 0.826–0.962), and economic discrimination (OR = 2.003; 95% CI, 1.420–2.824). The apparent AUC was 0.913 (95% CI, 0.878–0.948). Fixed-predictor-set bootstrap correction yielded an AUC of 0.902, a Brier score of 0.124, a calibration intercept of − 0.004, and a calibration slope of 0.907. Participants with AIS followed heterogeneous social alienation trajectories during the first year after diagnosis. The model with seven predictors showed promising performance after bootstrap correction and may inform further research on psychosocial risk assessment. However, predictor selection was not repeated during bootstrap validation, and independent external validation is required before routine clinical use.