INTRODUCTION
Lack of compliance with follow-up can threaten the validity of outcomes reported in clinical studies. The baseline characteristics of "drop-outs" differ from those of responders (1) and these same characteristics are often also predictors of outcome. We evaluated the impact of attrition by comparing the predicted outcomes of patients failing to return a questionnaire 12 months after surgery (non-responders) with the outcomes of those that did return one (responders).
METHODS
The data of all patients with degenerative spinal disorders of the thoracic or lumbar spine included in our in-house spine outcomes registry from 1.1.2006 to 31.12.2017 were analysed. Using the data of responders (8374/9189 (91%) cases), a model was developed to predict the multidimensional outcome of surgery (COMI, and its domain scores) at 12 months' post-surgery based on 20 key baseline variables (Müller et al 2021). This was used to predict the outcome of both responders and non-responders. The groups were compared using paired t-tests (for the predicted vs actual scores in those that responded, to check the fidelity of the model) and unpaired t-tests (for the predicted scores of responders and non-responders).
RESULTS
The predicted outcome scores of the responders did not differ significantly (p>0.05) from their actual outcome scores, suggesting the model was sufficiently accurate. The mean baseline scores of non-responders were significantly (p < 0.05) worse than those of respondents for most domains. The predicted outcome of non-responders at 12mo FU was significantly (p< 0.001) worse than that of the responders, for all domains.
DISCUSSION
Non-response at follow-up introduced significant and consistent bias in reported outcomes. Although the size of the effect was small in this particularly compliant cohort (with 91% follow-up at 12mo), when evaluating healthcare providers with less good follow-up rates the bias may be sufficiently large to threaten the validity of comparisons. The bias would overestimate the performance of hospitals with lower follow-up rates (perhaps also failing to detect poorly performing hospitals) and underestimate that of hospitals with high follow-up rates. If using spine surgery registries to perform benchmarking activities, the difference in follow-up rates between hospitals must be considered and adjusted for.  
1) Mannion AF, O'Riordan D, Fekete TF, Porchet F, Kleinstueck F, Jeszenszky D, Loibl M, Reitmeir R, Haschtmann D. Who is lost to follow-up and does it matter? A study of over 15'000 patients in a local spine surgery registry. EUROSPINE, e-Congress. 06-09.10.2020