A Population-Based Study of School Scoliosis Screening.

Yawn BP, Yawn PA, Hodge D, et al (Department of Research, Olmsted Medical Center; Sections of Biostatistics and Clinical Epidemiology and Department of Orthopedics, Mayo Clinic, Rochester, Minn), JAMA. 1999;282: 1427-1432.

Although there are school scoliosis screening programs in 26 states in the US, few outcome data exist. The purpose of this study was to determine the efficacy of the Rochester, Minnesota, screening program. Nurses screened 2,242 public and private school children entering kindergarten or first grade between 1979 and 1982, and they followed the students until 19 years of age. Beginning in 1984, fifth through ninth graders were examined yearly using visual inspection for asymmetry, the forward bending test, and the scoliometer. Screening was repeated for children with apparent curves or readings exceeding 6 units on the scoliometer. Students with abnormal results were referred for medical evaluation.

The investigators recorded screening results, spinal radiographs, Cobb measurements, curvature progression, greatest degree of curvature before age 19, curves of at least 20 degrees and 40 degrees or more, follow-up recommendations, referrals, and brace and surgical treatment. Statistical analysis included the number of children who tailed the screening exam (ie, showed some evidence of scoliosis) at each grade level, the number of students who failed the examination multiple times, percent positive agreement with future screening results for those students who failed at least once, number seen for follow-up, and number of students diagnosed and treated for scoliosis. The authors calculated the cumulative incidence of scoliosis, number needed to screen, and the positive predictive value (PPV) of the school screening program.

The cumulative incidence for diagnosed scoliosis was 1.8% for curves of more than 10 degrees, 1.0% for curves of at least 20 degrees, and 0.4% for curves of 40 degrees or more. Across consecutive years of screening, the percent positive agreement varied from 7% to 30%, with no increasing or decreasing pattern. Ninety-two of the children screened (4.1%)were referred for additional evaluation. Of those, 68 (74%) subsequently had a chiropractic or medical scoliosis evaluation. School screening identified 5 of the 9 children treated for scoliosis (0.04%) and resulted in referral of many more who were not treated. To identify 1 child who was ultimately treated, 448 children were screened (PPV=0.05).

The authors discussed factors potentially confounding the study. Positive screening tests varied from year to year. The investigators assumed that children who were medically evaluated but dismissed from further care without radiographs did not have scoliosis, possibly lowering the cumulative incidence. Low scoliosis prevalence resulted in a low PPV and high number of students who were needed for screening in order to identify one student requiring treatment.

The authors noted that the school district's high income and parental education levels and the lack of racial diversity might preclude broad generalization of the study. They advised legislators, nonetheless, to consider the data when making decisions about school scoliosis screening programs.

Diane R Katz, PT Tucson, Ariz