Corroborating evidence by exploring sources of bias in observational spinal cord injury studies

Abstract

Observational studies investigating large real-life datasets are a valuable resource in clinical research. Understanding the imperfect nature of clinical data, statistical approaches factoring in known confounders are instrumental for rigorously addressing bias. 1 Our recent work identifying pneumonia and postoperative wound infections (Pn/Wi) as risk markers for impaired long-term functional recovery and survival after spinal cord injury (SCI) 2 was considered as a strong statistical analysis. 3 However, some unexplored putative confounders in terms of nonrandom loss to follow-up, temporal changes in clinical practice, and exclusion criteria were discussed. 3 In order to evaluate and objectivize for the probability of attrition and temporal and selection bias, we apply and discuss an array of analytical tools extending beyond the format of the original publication. 2

Prognostic value of early leukocyte fluctuations for recovery from traumatic spinal cord injury

“…For the validation (i.e. Berlin) cohort, possible patient selection bias was examined first using a logistic regression model with the availability of WBC counts as the dependent variable and other inclusion criteria and relevant patient characteristics (AIS grade at admission, neurological level of injury, age, body mass index (BMI), sex and accompanying injury) as independent variables (Table S1B) 27 . No evidence for selection bias was found based on the relatively small sample of SCI patients with available WBC data, nor was there any association for any of the baseline parameters with an increased risk of exclusion from analysis.…”

Prognostic value of early leukocyte fluctuations for recovery from traumatic spinal cord injury

“…Berlin) cohort, possible patient selection bias was examined first using a logistic regression model with the availability of WBC counts as the dependent variable and other inclusion criteria and relevant patient characteristics (AIS grade at admission, neurological level of injury, age, BMI, sex and accompanying injury) as independent variables (Supplementary e-Table 1B). 27 No evidence for selection bias was found based on the relatively small sample of SCI patients with available WBC data, nor was there any association for any of the baseline parameters with an increased risk of exclusion from analysis. Linear mixed effects models with random intercept were then used to describe temporal changes for log transformed WBC data relative to the time between injury and Jogia 12 blood collection; all independent variables were used as fixed effects.…”