Predictors of publication of randomized controlled trials in nephrology

Abstract

Background and aims: Changes in clinical practice based on research findings are dependent on the dissemination of information, which is a goal at national meetings. Given that evidence for many interventions in nephrology is lacking, it is important to determine predictors of publication of randomized controlled trials (RCTs) presented as abstracts at national conferences.

Materials and methods: All abstracts submitted to the American Society of Nephrology (ASN) meeting 2005 were reviewed to identify completed RCTs. Univariate logistic regression was used to compare characteristics between published RCTs until June 2010 and those not published. Time to publication was calculated.

Results: 73 completed RCTs were presented out of 4,280 abstracts. 53% of these were published; median time to publication was 24 months (13.36 IQR). Oral presentations were published more frequently (OR 3.66, p = 0.01) as well as those with stated industry funding (OR 2.92, p = 0.03) and larger sample sizes (OR 2.1, p = 0.01). Blinded RCTs had ~ 4 times the odds of being published. Only ten RCTs used clinical endpoints such as death, hospitalizations etc.; however, all reached publication.

Conclusions: Almost 50% of RCTs presented at the ASN 2005 meeting were not published within the studied time-frame. Well-conducted RCTs enhance the body of evidence needed to care for patients, if published. It is important to identify characteristics associated with non-publication, as above, to help us perform and report our studies better.