Portal de Periódicos da CAPES

Trial registration records, updates, and protocols

2016; Elsevier BV; Volume: 388; Issue: 10042 Linguagem: Inglês

10.1016/s0140-6736(16)30965-5

1474-547X

Lisa Askie,

Cardiac, Anesthesia and Surgical Outcomes

Scrutiny of the GAS trial1Davidson AJ Disma N de Graaff JC et al.for the GAS consortiumNeurodevelopmental outcome at 2 years of age after general anaesthesia and awake-regional anaesthesia in infancy (GAS): an international multicentre, randomised controlled trial.Lancet. 2016; 387: 239-250Summary Full Text Full Text PDF PubMed Scopus (596) Google Scholar by the COMPare project team and the subsequent correspondence2Slade E Drysdale H Heneghan C on behalf of the COMPare project teamThe GAS trial.Lancet. 2016; 387: 1613Summary Full Text Full Text PDF PubMed Google Scholar, 3Davidson AJ Disma N de Graaff JC Withington DE McCann ME The GAS trial—authors' reply.Lancet. 2016; 387: 1615Summary Full Text Full Text PDF PubMed Scopus (3) Google Scholar highlight important issues regarding measures to minimise selective outcome reporting in clinical trials. The GAS trial was registered with ANZCTR in 2006 and ClinicalTrials.gov in 2008. Since then, registries have improved their outcome reporting mechanisms, and most registries now require separate listings for each outcome and more specific information regarding their method and timepoints of measurement.4Zarin DA Tse T Williams RJ Califf RM Ide NC The ClinicalTrials.gov database—update and key issues.N Engl J Med. 2011; 364: 852-860Crossref PubMed Scopus (503) Google Scholar Additionally, ANZCTR now allows PDF files of full protocols and other relevant documents to be uploaded. This functionality was not available when the GAS trial was initially registered. Given the inability to upload a full protocol to ANZCTR when the GAS trial was registered, why did the The Lancet publish only a summary protocol in 2009? Had the full version been published as submitted by the trialists, none of this Correspondence would likely have been necessary. In ANZCTR, trialists can include up to 43 outcomes (three primary and 40 secondary outcomes), so space is not an issue.3Davidson AJ Disma N de Graaff JC Withington DE McCann ME The GAS trial—authors' reply.Lancet. 2016; 387: 1615Summary Full Text Full Text PDF PubMed Scopus (3) Google Scholar Information submitted to a trial registry is the responsibility of the registrant, who agrees before registration that they are an appropriate representative of the trial sponsor. Hence, trial registries cannot “request”3Davidson AJ Disma N de Graaff JC Withington DE McCann ME The GAS trial—authors' reply.Lancet. 2016; 387: 1615Summary Full Text Full Text PDF PubMed Scopus (3) Google Scholar that all outcomes are listed in the registration record because there is no way of verifying what outcomes have been prespecified when full protocols are not made public. Trialists should thus give due attention to their trial's registration record and update it regularly. Too often, this task is seen as administrative rather than scientific and is frequently delegated to junior team members, increasing the risk of important details, such as the inclusion of all relevant outcome information, being inadvertently missed. It is the trialists' responsibility to regularly update their registration record, ensuring that they adhere to current standards of disclosure over (the often long) trial timeline. Trial registries and the COMPare project are both trying to reduce the potential bias introduced by selective outcome reporting. The GAS trialists have conceded that they did not fully update their registration record over the 10 years of the trial to include the level of detail now expected. We have acknowledged that ANZCTR's standards for outcome reporting are now higher than they were when the GAS trial was first registered in 2006. I hope the COMPare team will also concede that although their approach to expose outcome switching undoubtedly has high impact, the collateral damage caused to well meaning trialists along the way is something worth considering. I am the ANZCTR Manager. I declare no other competing interests.