Methodology

Searching for and scanning trial publications

Skills: Searching
Risk of publication bias and selective outcome reporting bias in ANZCTR registered trials

We need help from people who are competent at searching in Central, Medline and Embase.

Our study consists of taking a sample of RCTs registered in the Australian New Zealand Clinical Trials Registry (ANZCTR) and following them up to see if they are published and whether they switch their primary outcomes. It turns out that our sample is larger than we can handle on our own in the timeframe we have. The tasks we need help with are as follows:

Attempt to locate a publication for trials which have been registered in the Australian New Zealand Clinical Trials registry, through CENTRAL, Medline and Embase, utilizing any useful information which can be found in the registry entry. Record if you cannot. This part, proving a negative, tends to be the most time-consuming and difficult task on the list.

If a publication is found:
Record the publication date of the paper and the completion date, or any information which can be used to infer the completion date.
Copy the passages indicating the primary outcome(s), including timepoint(s).
Make a judgement on whether the primary outcomes in the publication are consistent with those in the registry, according to the following code:
Consistent - If all primary outcomes are described as such in the publication, and no other outcomes are described as primary OR all primary outcomes are included in the paper, and nothing is described as a primary outcome. All the correct timepoints should be included in the publication, and if the timepoints of the primary outcomes are explicitly stated, then all and only the correct timepoints should be stated.
Inconsistent - If the above is not the case. Unless:
Justified Inconsistency - If some reason is offered that purports to justify the inconsistency. You need only look at the points where the registration is mentioned and the inconsistency is described - we are not asking you to trawl through the whole paper looking to prove a negative for this part.

If outcomes or timepoints are spread across multiple publications (eg. for a long follow-up), then assess those publications collectively.

I get through about 3 trials an hour. I imagine someone more experienced might be significantly faster. We are offering acknowledgement for any help and authorship for help with 400 or more trials.

Ideal Applicant

Applicants should be sufficiently confident in their searching ability that we can conclude that if said applicant cannot find a trial, it either isn't published or is so difficult to find that it might as well not be. Location is not an issue. Applicants can request to be assigned a large number of trials upfront or do them in small batches - the only requirement is that any commitments are completed by the deadline.

Think you've got what it takes to get the job done for Sam?

Sign up
or
Log in
to respond