Minor effect of loss to follow-up on outcome interpretation in the Swedish spine register

Abstract

Background and purpose: Loss to follow-up in observational studies may skew results and hamper study reliability. We evaluated the importance of loss to follow-up in the Swedish spine register. Patients: Patients operated in the lumbar spine and scheduled for a postal questionnaire follow-up during part of 2016 were identified. Out of the 351 patients, 203 had responded. After multiple attempts, 115 of the 148 non-responders were reached; 68 returned the complete questionnaire; and 47 answered a brief questionnaire by phone. Analyses were made with the Chi-square test, analysis of covariance or logistic regression. Some analyses were adjusted. Results: At baseline, the non-responders were younger than the responders (55 vs 61 years, p < 0.001) and had higher Oswestry Disability Index (ODI) (54 vs 48, p = 0.003), lower SF-36 physical component summary score (PCS) (36 vs 40, p = 0.011) and lower EQ-5D (0.17 vs 0.27, p = 0.018). Mean back pain, leg pain, ODI, EQ-5D, SF-36 mental component summary score (MCS) improved significantly in both groups (all p < 0.001). SF-36 PCS did not improve in the non-responder group (p = 0.063). Non-responders perceived less improvement in back pain (global assessment back 60% vs 72%, p = 0.002). At follow-up, there were no differences in patient-reported outcome measures between the groups (all p ≥ 0.06), with the exception of a lower SF-36 MCS among the non-responders (p = 0.015). Interpretation: After surgery for lumbar spine degenerative disorders, non-responders achieve similar outcome as responders in the Swedish spine register, with the exception of a lower mental health and less perceived improvement in back pain. Graphic abstract: These slides can be retrieved under Electronic Supplementary Material. [Figure not available: see fulltext.]