Response to the letter to the editor: Clarification regarding effect size reporting in “biportal endoscopic versus conventional open spine surgery: a systematic review and meta-analysis”
Article information
Dear Editor,
The authors would like to sincerely thank you for your comments regarding the discrepancy in the reported pooled descriptive statistics and the Cohen’s d for operative time [1]. As requested, the authors have provided:
(1) A transparent extraction table including per-study means, standard deviations, and sample sizes used for the operative-time analysis (Supplement 1).
(2) The analytic workflow (e.g., RevMan file, R syntax, or equivalent) to ensure reproducibility of the following.
(3) The computation of both Cohen’s d in our initial analysis as well as the raw mean difference used for the recalculation of the pooled effect size (Supplement 2).
(4) The computation of the pooled descriptive statistics separately from the meta-analysis using Cochrane’s formula (Supplement 3)
(5) A recalculated pooled effect size, with the raw mean difference in minutes instead of Cohen’s d, with 95% confidence intervals, in the applicable forest plot (Supplement 4).
This has brought to the attention of the authors that there was an error in transcription for the pooled descriptive statistics for the operative time in the biportal decompression subgroup and subsequently the overall biportal cohort operative time. The correct pooled descriptive statistic for the decompression subgroup is 81.92±31.30 minutes and 144.30±73.94 minutes for the overall biportal cohort. The pooled descriptive data were separately and manually calculated from the remainder of the meta-analysis with the script provided in Supplement 2. This was a stand-alone error in manual entry of the data. As this was calculated separately from Cohen’s d and p-values, this error did not affect the conclusion of our meta-analysis.
Moreover, the pooled descriptive statistics described in Table 2 of [1] were calculated using Cochrane’s formula under a fixed effect model as outlined in Supplement 2. We utilized a random effect model in our meta-analysis which did not provide us with pooled descriptive statistics, which is why this was calculated separately. As the pooled mean and standard deviation reported in Table 2 of [1] was the result of a separate calculation and was not used in the meta-analysis, this explains the observed discrepancy with the Cohen’s d.
The authors hope that this information provides clarification to this observed discrepancy, and ensures the reproducibility and credibility of this review’s conclusions.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Author Contributions
All the work for the preparation of this letter was done by all authors.
Supplementary Materials
Supplementary materials can be available from https://doi.org/10.31616/2025.0604.r2.
Supplement 1. Extraction table operative time biportal vs. open.
asj-2025-0604-r2-Supplement-1.xlsSupplement 2. SPSS script for cohen’s d and raw mean difference for operative time.
asj-2025-0604-r2-Supplement-2.docxSupplement 3. Javascript script for cochrane’s formula calculation of pooled descriptive statistics.
asj-2025-0604-r2-Supplement-3.docxSupplement 4. Operative time pooled effect size with raw mean difference.
asj-2025-0604-r2-Supplement-4.pdf