Time to publication among completed diagnostic accuracy studies: associated with reported accuracy estimates

Tags: Oral
Korevaar D1, van Es N2, Zwinderman A1, Cohen J1, Bossuyt P1
1Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Center, University of Amsterdam, The Netherlands, 2Department of Vascular Medicine, Academic Medical Center, University of Amsterdam, The Netherlands

Background: Studies of therapeutic interventions with statistically significant results are published more rapidly than those without, which can lead to reporting bias.

Objectives: We evaluated whether diagnostic accuracy (DA) studies that report higher accuracy estimates are also published more rapidly.

Methods: We obtained all primary DA studies included in meta-analyses of MEDLINE-indexed systematic reviews published between September 2011 and January 2012. For each primary study, we extracted estimates of DA (sensitivity, specificity and Youden’s index), the completion date of participant recruitment, and the publication date. We calculated time from completion to publication and assessed associations with reported accuracy estimates.

Results: Forty-nine systematic reviews were identified, containing 92 meta-analyses and 924 unique primary studies, of which 756 could be included. Study completion dates were missing for 285 (38%) of these. Median time from completion to publication in the remaining 471 studies was 24 months (interquartile range (IQR) 16 to 35). Primary studies that reported lower estimates of sensitivity (Spearman’s rho = -0.14; P = 0.003), specificity (rho = -0.17; P < 0.001), and Youden’s index (rho = -0.22; P < 0.001) had significantly longer times to publication. When comparing time to publication in studies reporting accuracy estimates below versus above the median, the median number of months was 25 versus 23 for sensitivity (P = 0.046), 27 versus 22 for specificity (P = 0.001), and 27 versus 22 for Youden’s index (P < 0.001; Fig). These differential time lags remained significant in multivariable Cox regression analyses with adjustment for other study characteristics, with hazard ratios of publication of 0.81 (95%CI 0.66 to 1.00) for studies reporting a sensitivity below the median, 0.70 (95%CI 0.57 to 0.87) for studies reporting a specificity below the median, and 0.63 (95%CI 0.51 to 0.79) for studies reporting a Youden’s index below the median.

Conclusions: Time to publication was significantly longer for studies reporting lower accuracy estimates, suggesting that reporting bias may also occur in systematic reviews of DA studies.