Abstracts

Rationale: Drug-resistant epilepsy (DRE) occurs in up to 40% of patients with epilepsy; many are considered for epilepsy surgery. For patients with focal DRE, about 50% require invasive monitoring prior to surgery. Subdural electrodes (SDE) and stereoelectroencephalography (sEEG) are the most common invasive monitoring strategies. While several meta-analyses of observational data demonstrate substantial differences in the morbidity profile of sEEG and SDE, the most cost-effective invasive monitoring strategy remains unknown. This study assessed the cost-effectiveness of invasive monitoring strategies in patients with DRE using a decision analytic model.  

Methods: This economic evaluation used a model to simulate the risks and benefits of SDE and sEEG invasive monitoring strategies, where patients faced differing risks of morbidity, mortality, surgical resection, and seizure freedom depending on which strategy they underwent. Data were collected from meta-analyses of observational data from epilepsy surgery centers internationally and applied as inputs to the model (Table 1). The model considered a base case simulation of an average patient with DRE. The main effectiveness measure was the probability of a positive outcome after invasive monitoring, which was defined as improvement in seizures without a complication. Cost-effectiveness was measured via the incremental cost-effectiveness ratio. Sensitivity analyses were performed to assess the effects of model parameter variations on results.

Results: Invasive monitoring with sEEG had an increased cost of $274 and increased probability a positive outcome by 2% compared to SDE, or  $12,630 per positive outcome obtained (Table 2). Sensitivity analysis revealed that the cost of sEEG and SDE were the two most important determinants of model outcome with the probability of a complication with sEEG and the cost of any complication being the next most influential, with relatively small variations in these parameters causing either sEEG or SDE to be favored. Interestingly, variations in seizure improvement and mortality between sEEG and SDE in the narrow ranges suggested by meta-analysis results did not have a meaningful impact on model outcome. However, a probabilistic sensitivity analysis, varying all parameters simultaneously over distributions, suggests that sEEG would be narrowly favored if the willingness-to-pay per positive outcome gained (or negative outcome avoided) is ≥$20,000 (Figure).

Conclusions: In this analysis, invasive monitoring with either SDE or sEEG were nearly equivalent in terms of cost-effectiveness with SDE or sEEG costs being the most influential determinants of model outcome. This study demonstrates that, when accounting for cost, clinical outcome, and complications wholistically, invasive monitoring strategies are similar with regards to cost-effectiveness. Craniotomy avoidance and a lower overall complication risk may make sEEG preferable to SDE in select patients, as was also suggested when all model parameters were simultaneously varied.

Funding: Not applicable