School-based scoliosis screening among 73,509 students in Zhangye, Northwest China: epidemiological characteristics, 7-day non-completion in the radiographic diagnostic cascade, and an exploratory case–control analysis

Background School scoliosis screening is effective only when students who screen positive complete diagnostic confirmation. We described a citywide program in Zhangye, Northwest China, quantified 7-day non-completion in the radiographic diagnostic cascade, and explored characteristics associated with adolescent idiopathic scoliosis (AIS) case status. Methods From September to October 2025, all 66 primary and secondary schools in Zhangye City participated. Students aged 10–18 years underwent postural inspection, the Adams forward bend test, and angle of trunk rotation (ATR) measurement. ATR > =5 degrees defined a screening-positive result. Screening-positive students were advised to undergo standing full-length spine radiography within 7 days. Two experienced spine surgeons, blinded to ATR results, independently measured the maximum Cobb angle; their mean was used. AIS required a mean Cobb angle ≥ 10 degrees and exclusion of known secondary causes. We analyzed the screening population, diagnostic cascade, and an unmatched exploratory case–control sample of 250 AIS cases and 250 screening-negative controls using multivariable logistic regression. Results Among 73,509 students, 1,451 screened positive (1.97%, 95% CI 1.88–2.08). The rate was higher in girls than boys (2.59% vs. 1.49%), increased after age 12, and ranged from 1.50 to 3.08% across counties/districts. Only 348 screening-positive students completed radiography within 7 days (23.98%, 95% CI 21.86–26.25); 1,103 had no recorded completion within this window (76.02%, 95% CI 73.75–78.14) reflecting 7-day non-completion rather than permanent non-completion. Among those imaged, 250 had a mean Cobb angle ≥ 10 degrees and were diagnosed with AIS, representing 71.84% (95% CI 66.89–76.31) of those imaged and a minimum confirmed detection proportion of 0.34% (95% CI 0.30–0.38) among all screened students. In the complete-case model ( n = 452), female sex, increasing age, sometimes rather than frequently maintaining an upright classroom posture, and left-side rather than supine sleeping were associated with AIS case status; BMI was not independently associated. Conclusion The principal implementation gap occurred between screening-positive identification and timely radiographic confirmation. The 0.34% proportion is a minimum confirmed detection estimate, not population prevalence. Closed-loop referral pathways are needed, and the exploratory case–control associations should not be interpreted causally.