This study showed the beneficial effect of using templates for RCT protocols tailored to the SPIRIT guidelines on the completeness of reporting of RCT protocols developed by master’s students. Specifically, 8.24 out of the 10 core SPIRIT items were reported to be used in RCT protocols where the templates were used; this number represents 1.79 items more (0–10 scale, 95% CI 0.90 to 2.85) than in protocols where the templates were not used. We observed the greatest differences for items 16b (allocation concealment mechanism), 16c (implementation), 17a (blinding), and 22 (harms).
These findings are in line with those of previous studies focused on CONSORT guidelines that show that the most successful strategies for improving adherence are those focused on helping authors at the writing stage of the manuscript and those involving reporting guideline experts in the peer review process. Barnes et al. performed an RCT in which RCT manuscripts were developed using the CONSORT-based online writing aid tool (COBWEB) reported an average of 2.1 CONSORT items (0–10 scale, 95% CI 1.5 to 2.7) more than did those that did not use that tool (13). In the context of a dentistry journal, a cross-sectional survey revealed an increase of 1.52 CONSORT items (0–10 scale, 95% CI 1.05 to 2.0) in articles conforming to a subheading system such as the one proposed in our study. Additionally, another RCT showed a difference of 1.78 CONSORT items (0–10 scale, 95% CI 0.39 to 3.23) between the manuscripts that received an additional review by a reporting guideline expert focused on 8 core CONSORT items and those that underwent usual peer review (12). However, other strategies focused on the manuscript submission, peer review, and manuscript revision stages have been proven unsuccessful: requesting authors to submit a checklist together with the manuscripts (9), asking standard peer reviewers to check specific reporting guideline items (3), and implementing a web-based tool at the manuscript revision stage (25).
Regarding the reporting quality of each item, our results match those of previous studies that have shown remarkable improvements in key methodological items that are common to SPIRIT and CONSORT and that are usually poorly reported in RCT reports: outcomes, blinding, or allocation concealment mechanism (12). For this reason, helping authors properly report these items in RCT protocols could have a remarkable impact on the reporting quality of the final RCT reports. Notably, it is surprising that in both periods, there was a low proportion of manuscripts (47% and 53%, respectively) that correctly reported the item Outcomes. This was mainly due to the lack of explicit differentiation between the primary and secondary outcomes and the failure to include their specific measurement variables. Additionally, less than half of the manuscripts in the post-intervention period (47%, 8 of 17) and none in the pre-intervention period included an adequate description of item 22 (Harms). We hypothesize that the reason for this is that, unlike many medical or pharmacological RCTs, most physiotherapy interventions are not considered potentially harmful. This could make authors less prone to report the absence or presence of harms, even though SPIRIT guidelines indicate so.
Strengths and limitations
The strengths of the study include that the intervention was implemented in a real setting with no disruption to usual procedures in the master’s program. Also, the intervention evaluated has no cost, and it could be easily implemented in different contexts (journals, ethics committees, or education) and for other reporting guidelines. In addition, the study outcomes were assessed in duplicate by two blinded assessors.
We also mention several limitations. First, we did not use an RCT design, which may have affected the validity of the study results. For this reason, factors other than the intervention might have influenced the completeness of reporting of the RCT protocols included in the study. However, our results are similar to those of previous studies that evaluated other interventions focused on the CONSORT guidelines (12, 13). Second, our participants were master’s degree students, who are not necessarily representative of typical authors who write RCT protocols. Third, we included study protocols from only one master’s program in the field of physiotherapy. Furthermore, our intervention focused only on 10 items of the SPIRIT guidelines, and the results could be different if the whole checklist or other guidelines were considered. Finally, there is no validated outcome measure that evaluates the completeness of reporting of research manuscripts. For this reason, we used the SPIRIT checklist, which is not intended to be an evaluation tool but rather just guidance for reporting (26). However, this decision is consistent with the evaluation strategy of previous research in this field.
Implications
This is the first intervention focused on the protocol writing stage that has ever been evaluated as to whether it improves the completeness of reporting. Furthermore, this is also the first intervention that consists of training biomedical students on the practical use of reporting guidelines (7). The results shown here should stimulate the implementation of this and other research-based educational interventions to help students acquire competencies regarding research reporting and methodology (20, 21). Some facilitators of this strategy are that it has no cost, it would be easy to implement in different contexts (e.g., education, ethics boards, or journals), and it could be followed for other reporting guidelines. Interestingly, future research should evaluate, using an RCT design, whether similar benefits can be obtained from other common reporting guidelines, such as CONSORT, STROBE or PRISMA, and in other contexts, such as those mentioned above.
Improving adherence to SPIRIT guidelines is fundamental for different reasons. First, it makes RCT protocols more transparent and complete, allowing readers to fully understand the rationale, methods, and ethical aspects of RCTs. Second, as the background and methods sections of CONSORT are very similar to those of SPIRIT, improving adherence to SPIRIT makes it easier to comply with CONSORT requirements. Finally, even though SPIRIT provides reporting rather than methodological guidance, using SPIRIT makes authors aware of certain methodological aspects that they need to consider when carrying out an RCT, which can improve the study’s conduct.