190. Predictions for who are lost to follow-up in spinal cord injury clinical trials: an opportunity to enhance trial design and execution

Abstract

<h3>BACKGROUND CONTEXT</h3> Despite advancements in research and promising preclinical findings we have yet to isolate any new practice-changing interventions for the treatment of spinal cord injury (SCI) in clinical trials. This raises concerns about the methodology and design of clinical trials in SCI. One of the concerns in SCI trials is with high rates of missing data. Combined with difficulty in recruiting patients and the cost associated with clinical trials in SCI strategies need to be implemented to reduce the rates of loss to follow-up (LTFU) in trials <h3>PURPOSE</h3> Identify factors associated with LTFU in SCI clinical trials. <h3>STUDY DESIGN/SETTING</h3> Multicenter, prospective data from the STASCIS study and the NACTN registry database were analyzed for this study. <h3>PATIENT SAMPLE</h3> All adult patients with SCI from the STASCIS and NACTN registries were selected. <h3>OUTCOME MEASURES</h3> LTFU was defined to be in patients whodid not come to follow up at 12 months and previously did not come to follow-up at least 3 months after baseline ASIA Impairment Scale assessment (≤72 hours post-injury). <h3>METHODS</h3> Missing data were imputed with multiple imputations. Univariate analysis was performed for each independent variable, and predictors with p < 0.2 were included in the final multivariate model for backward stepwise logistic regression based on AIC values. Odds ratios with 95% confidence intervals were calculated with a significance level set as p < 0.05. <h3>RESULTS</h3> A total of 885 patients (46.5±17.6, 45.5% female) were eligible for inclusion. On univariate analysis, 23 variables were significant and included for feature selection. In the final model, patients were less likely to miss follow-up if they had higher baseline lower extremity motor scores (OR: 0.99[0.98-0.99], p=0.021), received early surgery (OR: 0.49[0.35-0.69], p < 0.001), Asian (OR: 0.31[0.18-0.54], p < 0.001) or Black race (OR: 0.63[95% CI 0.45-0.88], p=0.006) compared to White Caucasian, or had a systemic infection complication (OR: 0.35[0.17-0.70], p=0.003). Independent predictors of LTFU were postoperative DVT (OR: 3.28[1.64-6.55], p=0.001) and hematological complications (OR: 1.67[1.07-2.60], p=0.025), and a history of malignancy (OR: 3.65[1.01-13.15], p=0.048) and smoking (OR: 1.55[1.15-2.08], p=0.004). Higher baseline GCS was associated with making follow-up (OR: 0.94[0.88-1.00], p=0.053) and diabetic history with LTFU (OR: 1.56[0.97-2.52], p=0.068), but were not statistically significant in the final model. <h3>CONCLUSIONS</h3> Patients that had a greater motor disability, late surgery and associated comorbidities (such as the history of malignancy, smoking, DVT and hematological complications), are at greater risk for LTFU and may require concerted patient engagement approaches to ensure comprehensive spinal care and track recovery trajectory. <h3>FDA DEVICE/DRUG STATUS</h3> This abstract does not discuss or include any applicable devices or drugs.