Noel-Storr and colleagues’ maternity-care report (26 April 2021) used the review team’s 50 included studies as its reference, not scientific truth. Each had at least one individual potentially relevant or unsure judgment; the first-round resolver making final decisions on disputed or uncertain records nevertheless rejected eight. A repeat used 85 new contributors and two independently judging resolvers; disagreements kept records in. It retained 48 of 50 included studies: two were lost before resolution, none at resolution. The body reports 96%; the abstract says 94% for that count. Contributors and the decision rule changed together, so improvement cannot be attributed to adding a resolver alone. In this single review, individual detection was undone at final escalation.