Abstract
Dear Editor,
We read with interest this recent article that sought to examine the rate of quotation errors arising from articles referencing the Distal Radius Acute Fracture Fixation Trial (DRAFFT) (Costa et al., 2014). The authors defined a major quotation error as ‘one that was not substantiated by, was unrelated to or contradicted the findings of DRAFFT’. We thank the authors for highlighting two articles from our centre that they defined as containing major quotation errors.
The articles from our centre were study protocols for prospective randomized trials of operative versus non-operative management, one relating to fractures of the humeral diaphysis (the HUmeral shaft fracture FIXation [HU-FIX] study) (Oliver et al., 2019) and the other to associated medial malleolus fractures in unstable fractures of the ankle (the Medial malleolus: Operative Or Non-operative [MOON] trial) (Carter et al., 2019). Both studies are deliberately pragmatic, leaving specific management decisions (relating to operative technique, postoperative immobilization and physiotherapy input) to the discretion of the treating surgeon so that the results might be generalizable. As Cay and colleagues observed, both the HU-FIX Study and MOON Trial protocols reference DRAFFT after a summary of the pragmatic elements of those studies. We feel it was appropriate to cite DRAFFT in this way, given that DRAFFT is a large, randomized trial that also incorporated pragmatic elements into the design to improve the external validity of the findings. Indeed, we would commend the DRAFFT article to Cay and colleagues as an excellent overview of the value of pragmatic randomized trials like the HU-FIX Study and MOON Trial. The article details the pragmatic nature of DRAFFT, with key aspects such as ‘the details of the surgery’, ‘the use or otherwise of a cast’ and ‘other rehabilitation input beyond the written information sheet’, all being left to the discretion of the treating surgeon (Costa et al., 2014). The DRAFFT Group explained their reason for doing so was ‘to ensure that the results of the trial could be generalized to as wide a group of patients as possible’ (Costa et al., 2014).
The citation of the DRAFFT article in our study protocols could not have been interpreted as relating to the clinical findings of DRAFFT, given that the reference was made in the context of study design in both instances. We would invite Cay and colleagues to consider that it is possible to cite a reference precisely because it relates to study design (or any other aspect of a study), even if it does not relate to the study findings. Moreover, while it is unrealistic to expect quotation errors will never occur in academic literature, we suggest authors may wish to communicate with other centres directly if they are uncertain about the context or nuance of a cited reference. This may be preferable to inappropriately ascribing major quotation errors and then publicizing those assertions to the wider orthopaedic community.
References
Brook Leung
Keegan Curlewis
Andrew Stone
Tom Roper
David Ricketts
Trauma and Orthopaedics Department, Royal Sussex County Hospital, Brighton, UK
*Corresponding author:
Dear Editor,
We thank Mr Oliver and colleagues for their letter. The aim of our article was to examine the rate of quotation errors related to DRAFFT (Costa et al., 2014). We tried to avoid comment on the content or validity of individual articles. In our study, each of the studied articles was assessed independently by two reviewers, with a third independent review if there was disagreement between the two initial reviewers. In addition, we discussed each potential error identified and how it should be classified.
As we stated, a major error was defined as one that was ‘not substantiated by, was unrelated to or contradicted the findings of DRAFFT’. We note the point relating to the fact that the specified reference in both articles (Carter et al., 2019; Oliver et al., 2019) was discussing the pragmatic nature of the study protocols. In our opinion, in neither case was it made clear that it was the pragmatic nature of the study protocol that was related to DRAFFT and why it had been referenced as there was no mention of other trials. Therefore, we classified this as a major error; namely one that is unrelated to DRAFFT.
We would suggest that the inclusion of a comment explicitly linking to the study design would have avoided this. For example: ‘pragmatic study design, such as used in similar multicentre trials, will reproduce day-to-day trauma care and improve the external validity of study results’.
In any subjective assessment of references there is the possibility for disagreement over the nature of an error or whether a given reference is incorrect. We feel that our interpretation represents what a reasonable clinician would assume.