Peer review reduces spin in PCORI research reports

Table 2 Characteristics of included PCORI studies and Draft Final Research Reports (DFRR)

Characteristics	N (%)
Accepted version of DFRR (N = 64)
1st revision	20 (31%)
2nd revision	31 (48%)
3rd revision	11 (17%)
4th revision	2 (3%)
Did the reviewers comment about spin in the DFRR? (N = 64)
Yes	55 (86%)
No	9 (14%)
Did the project have one or more journal articles *accepted or published* before receiving PCORI peer review feedback? (N = 64)^a
Yes	58 (90%)
No	5 (8%)
Can’t tell	1 (2%)
In DFRRs with spin, which type of spin was identified? (N = 55)
Reporting bias	46 (84%)
Inappropriate interpretation	40 (73%)
Attribution of causality	5 (9%)
Inappropriate extrapolation of results	15 (27%)
Among DFRRs with spin identified, were comments about spin addressed in last version of the DFRR? (N = 55)
Yes, the authors changed the DFRR to address this issue	38 (69%)
Yes, the authors provided a satisfactory response explaining why a change is not required^b	9 (16%)
No	8 (15%)
Can’t tell	0 (0%)
Of the DFRRs in which spin was identified, did at least one article include the spin identified in the DFRR? (N = 55)
Yes, one or more articles included spin	21 (38%)
No, none of the articles included spin	13 (24%)
Not applicable, none of the associated articles reported results related to comments about spin	21 (38%)
Can’t tell	0 (0%)

^aWe could not determine the exact date of acceptance for 88 articles, so we compared the date of publication with the date of the PCORI peer review feedback. Of those, 67 articles were published before sending PCORI peer review feedback and therefore must have been accepted before PCORI peer review feedback was sent. Two (2) articles were published within 28-days of sending PCORI peer review feedback and we assumed the articles were accepted before PCORI peer review feedback was sent. Of the remaining articles, 15 were published more than 98 days after PCORI peer review feedback was sent and we judged them likely to have been accepted for publication after the authors had received PCORI peer review feedback. We could not determine whether 4 articles published between 29 and 97 days of sending PCORI peer review feedback had been accepted before or after sending PCORI peer review feedback; three (3) of these articles were associated with DFRRs for which there was at least one article accepted prior to receiving PCORI peer review feedback
^bIn all cases for which assessors coded “provided a satisfactory response explaining why a change is not required,” the associate editors accepted the investigators’ reports without their previously recommended changes

ISSN: 2058-8615