Publication Bias: Association of Diagnostic Accuracy in Radiology Conference Abstracts with Full-Text Publication

Abstract

eporting bias in biomedical research results in delays and leads to a lack of publication of insignificant study results. It is well documented that trials of therapeutic interventions with insignificant or negative results are less likely to reach full-text publication in peer-reviewed journals; this is typically called publication or reporting bias (1-5). This may lead to bias in syntheses of available literature, such as systematic reviews, leading to overestimates of the effectiveness of interventions (2,(6)(7)(8).In contrast to the well-documented phenomenon and impact of publication bias in therapeutic interventional studies, publication bias in diagnostic test accuracy (DTA) research is not well understood. More than 50% of conference abstracts presented at imaging research meetings do not go on to full-text publication (9-12). Imaging researchers often evaluate DTA, which does not generate a significant result. Instead, these studies produce measures of DTA, such as sensitivity and specificity, which can be considered as a proxy for statistical significance, as has been done in previous evaluations of publication bias in DTA research (13)(14)(15)(16)(17).A recent study showed that imaging studies with higher DTA are published faster than those with lower DTA ( 14); however, it is unknown whether such studies are more likely to be published. A Cochrane systematic review found that conference abstracts of interventional studies that reported at least one significant result were 30% more likely to reach full-text publication than were those that did not (18). However, similar studies of DTA research presented

Searching practices and inclusion of unpublished studies in systematic reviews of diagnostic accuracy

“…Whether this phenomenon also occurs among diagnostic accuracy studies is largely unclear. Some evidence is hinting towards similar selective reporting practices, although other studies could not confirm this . In our study, among the three systematic reviews that made a comparison between published and unpublished diagnostic accuracy studies included in the meta‐analysis (or performed a sensitivity analysis excluding unpublished studies), none found a significant difference.…”

“…In recent years, a number of evaluations assessed publication rates among completed diagnostic accuracy studies, identifying that approximately a quarter to half of them failed to reach full-text publication in a peer-reviewed journal. [5][6][7][8][9] Two studies evaluated time from study completion to publication among published diagnostic accuracy studies, both concluding that those reporting higher estimates of diagnostic accuracy were published more rapidly. 10,11 It seems plausible that studies reporting higher estimates of diagnostic accuracy also more often reach publication, although this has yet to be demonstrated.…”

“…For diagnostic accuracy studies, evidence of similar selective publication practices is still scarce, yet increasing. In recent years, a number of evaluations assessed publication rates among completed diagnostic accuracy studies, identifying that approximately a quarter to half of them failed to reach full‐text publication in a peer‐reviewed journal . Two studies evaluated time from study completion to publication among published diagnostic accuracy studies, both concluding that those reporting higher estimates of diagnostic accuracy were published more rapidly .…”

“…Two studies evaluated time from study completion to publication among published diagnostic accuracy studies, both concluding that those reporting higher estimates of diagnostic accuracy were published more rapidly . It seems plausible that studies reporting higher estimates of diagnostic accuracy also more often reach publication, although this has yet to be demonstrated …”

Tweeting Bias in Diagnostic Test Accuracy Research: Does Title or Conclusion Positivity Influence Dissemination?

“…5 Interestingly, a multivariable analysis identified no association between magnitude of diagnostic test accuracy (DTA) and higher likelihood of publication. 6 As diagnostic imaging studies have a tendency to ''spin'' their results referring to the reporting practice of accentuating optimism regarding their study topic beyond what the results justify, this citation bias could result in the overinterpretation of accuracy of imaging modalities. [7][8][9][10] This subsequently can lead to their increased and inappropriate use in clinical practice.…”