Clinical Characteristics of Patients Who are Lost to Follow-up after Surgical Treatment for Adolescent Idiopathic Scoliosis

Background: Low follow-up rates are a limitation of many long-term studies. Studies on patients after surgical treatment of adolescent idiopathic scoliosis (AIS) rarely have follow-up rates that reach 80% after 5 years or 70% after 10 years. While there is general consensus that a high patient drop-out rate will decrease the accuracy of a study, little is known about the characteristics of those patients who have dropped out. The purpose of the present study was to identify patients who were lost to follow-up after surgical correction of AIS and to compare their clinical characteristics to those of patients who were available for follow-up. Patients and Methods: All members of a popular scoliosis-specific online community were asked to submit information about demographic, surgical, and clinical characteristics. Patients who had surgical treatment for AIS and a minimum of 12 months since surgery were grouped into "lost to follow-up - LTF" and "follow-up - FU" based on whether they had changed their orthopedist or not after surgery. Demographic, surgical, and clinical characteristics with the use of the SRS-22 instrument were then compared. Results: 271 patients submitted information of whom 81 patients (39 FU, 42 LTF) fulfilled all inclusion criteria. No difference was found in terms of age at surgery, surgical approach, number of fused levels, or revision rate whereas major curve correction and time since surgery showed significant differences. A subanalysis that was done after matching patients for time since surgery revealed significant differences for the SRS-22 domains function, pain, mental health, and total score in favour of the FU group. Conclusion: We found significant clinical differences for patients who are lost to follow-up when compared to patients who are available for follow-up. While we are aware of the difficulties in obtaining high follow-up rates for a clinical study, care should be taken in analysing studies with low follow-up rates because results may be skewed.