Intraoperative neuromonitoring alerts that reverse with intervention: treatment paradox and what to do about it.

Intervention-mediated recovery from adversely changed evoked potential recordings may provide evidence for improved outcomes during neurophysiological intraoperative monitoring. However, these reversible signal changes (RSCs) are ambiguous because the patient's neurologic status cannot be known either at signal decline or after intervention. This article describes methods to reduce this ambiguity. Randomized control trials are not always possible or ethical. Recent thought on grading evidence has acknowledged that guidelines first described by Sir Austin Bradford Hill may support evidence for causation. Causality guidelines identified RSCs most likely to be truly positive in three reported studies. Diagnostic statistics were revised accordingly. A range of revised positive predictive values and likelihood ratios was calculated in the three studies, using causality guidelines. The revised data were similar to those reported for other diagnostic tests used in medicine. The RSCs may be assessed using causality guidelines for more accurate reporting of diagnostic statistics while preserving information related to surgical intervention and recovery that is lost with end of surgery diagnostics or when RSCs are ignored. A method is described for including RSCs in diagnostic statistics. This approach will more readily permit assessment of the value of neurophysiological intraoperative monitoring in prediction and prevention of neurologic deficits.