Stroke is a leading cause of mortality, morbidity and disability worldwide. Saying that mechanical thrombectomy is the standard of care in patients with acute large vessel occlusion may be clinically obvious and almost unquestionable nowadays [1, 2], but can we do in such a sound statement from a cost-effectiveness point of view?

Based on the systematic review by Xiao Wu et al [3] published in this issue, all the included cost-effectiveness analyses (CEAs) concluded indeed that mechanical thrombectomy is cost-effective. This is by itself an important conclusion, as systematic reviews are considered a high level of evidence in scientific literature. Nonetheless, this was not the main aim of this paper; authors went a step deeper and analysed how this conclusion was reached in each study, by focusing on the methods used to account for costs and utilities.

Clinicians may not always be familiar with CEAs and their terminology. In a very summarised way, healthcare CEAs do evaluate the costs and effects of a concrete medical intervention. Costs are usually measured in US dollars and should include direct and indirect, as well as acute and long-term medical costs, but also indirect costs from a societal perspective. On the other hand, quality-adjusted life-years (QALY) is a routinely used summary measure of effect [4]. QALY corresponds to the product of multiplying utility to years. Utility is a value between 0 (death) and 1 (perfect health), although in some circumstances could be negative reflecting a health state “worse than dead” [5]. Utility values can be obtained by various different methods [6], and in stroke patients are often treated closely related to modified Rankin Scale (mRS) [7]: in general terms, the lower the mRS, the higher the utility, and vice versa.

In the analysed CEAs, regarding utilities, the authors found important heterogeneity mainly in (1) the methods used to measure utility, (2) the utility values given to different concrete mRS, (3) the pooling of patients according to mRS (single values or ranges), (4) the statistical distributions assigned to utility parameters, and (5) the incorporation of recurrent stroke. The reader may wonder if this could be explained by a practical lack of consensus recommendations for stroke-specific utility calculations.

On the other hand, regarding costs, (1) very few studies included social perspectives and (2) almost none achieved to include all the indirect cost components. Also, (3) heterogeneity was found around the distribution used for cost parameters. These facts were demonstrated both before and after the publication of the Second Panel Recommendations for Cost-effectiveness Analyses in 2016 [4]. At this respect, the rationale of lacking guidelines may seem questionable for papers published after 2016, but it should be noted that the Second Panel Recommendations are not stroke-specific, and some particular cost components are not directly approached.

Along the same line, this heterogeneity in cost and utility methodologies carries direct impact in measures such as the willingness to pay (WTP) and the incremental cost-effectiveness ratios. Moreover, the WTP thresholds were also found to be not homogeneous between studies.

The lack of stroke-specific consensus guidelines for CEAs, but also an unstrict adherence to available recommendations, may result in the not uncommon practice of using previously published CEAs’ methodology, and this approach could ultimately potentially perpetuate some inaccuracies.

Closing the circle, even if it seems quite unarguable that mechanical thrombectomy is cost-effective, the exposed facts could cast doubts on the accuracy and generalizability of the different CEA results, and from an extreme hypothetical point of view: could even cause to somebody any doubt on its actual cost-effectiveness?

If this humble editorial comment has raised your interest, you must definitely read the full paper and make your own conclusions.

Finally, and seizing the opportunity, the readers may kindly perform critical thinking on the following questions at three different levels:

  1. 1.

    Should stroke researchers, in concrete, demand stroke-specific guidelines for CEAs?

  2. 2.

    Should authors, in general, be stricter following available consensus guidelines?

  3. 3.

    Should scientific-literature publications, in general, uniformly demand more rigour following consensus guidelines?