– The Results section states that …. only 8 studies ….. met the inclusion criteria and were synthesized. Yet the Discussion section cites dozens of studies (e.g., 21, 22, 23, 24, 25, 26, 27, 28, 29, etc.) that were not listed in the included studies in Table 2. This suggests the authors used the systematic review format to lend methodological legitimacy but then abandoned their own inclusion criteria to discuss a much broader, unselected body of literature.
How do you justify presenting a systematic review with only 8 included studies, while your Discussion draws extensively from studies outside this set, thereby making your conclusions unverifiable and essentially a narrative review in disguise?
– No formal risk-of-bias or quality appraisal tool (e.g., QUADAS-2, ROBINS-I, CASP) was applied to the 8 included studies. Without this, there is no way to determine whether the reported findings (e.g., diagnostic accuracy, bias in algorithms) are reliable or at high risk of bias. The claim of rigorous synthesis is therefore unsupported.
What validated quality assessment tool did you use to appraise the 8 included studies, and how did you account for potential biases in those studies before concluding that AI outperforms clinicians or introduces algorithmic bias?