Article type
Year
Abstract
Background: Overinterpretation of the results by authors in order to make the interventions look favourable is referred to as 'spiná. This has been shown to be common in randomized trials but may also play a role in diagnostic accuracy studies. The clinical use of tests based on inflated conclusions could lead to unnecessary testing and may trigger physicians to make incorrect clinical decisions, compromising patient safety.
Objectives: To explore the mechanisms behind 'spiná in diagnostic test accuracy studies and to estimate their prevalence in journals with a high impact factor.
Methods: We searched MEDLINE for diagnostic test accuracy studies published between January and June 2010 in journals with an impact factor of 4 or higher. A random selection was evaluated with a structured data extraction form by two authors independently. Disagreement was resolved by consensus or by a third party.
Results: Our search yielded 420 eligible articles of which we selected 85. The interrater agreements were 81% and 100% before and after discussion. Of the 74 articles included in the analysis, 66 (89%) contained a form of spin. The mechanisms behind spin included: not stating a prior study hypothesis (85%), reporting positive test recommendations not in line with the test accuracy measures (12%), basing test recommendations on results from sub-group analysis only (12%), conclusions that differed from the aim of the study (12%), an overoptimistic abstract, with selective reporting of favourable results only, or stronger test recommendations compared to the main text (45%).
Conclusions: Overinterpretation of results of diagnostic accuracy studies is common. We hope that highlighting the mechanisms behind spin will enable peer reviewers correctly sieve overoptimistic reports of diagnostic accuracy studies and encourage investigators to be more clear in designing, more transparent in reporting, and more stringent in interpreting test accuracy studies.
Objectives: To explore the mechanisms behind 'spiná in diagnostic test accuracy studies and to estimate their prevalence in journals with a high impact factor.
Methods: We searched MEDLINE for diagnostic test accuracy studies published between January and June 2010 in journals with an impact factor of 4 or higher. A random selection was evaluated with a structured data extraction form by two authors independently. Disagreement was resolved by consensus or by a third party.
Results: Our search yielded 420 eligible articles of which we selected 85. The interrater agreements were 81% and 100% before and after discussion. Of the 74 articles included in the analysis, 66 (89%) contained a form of spin. The mechanisms behind spin included: not stating a prior study hypothesis (85%), reporting positive test recommendations not in line with the test accuracy measures (12%), basing test recommendations on results from sub-group analysis only (12%), conclusions that differed from the aim of the study (12%), an overoptimistic abstract, with selective reporting of favourable results only, or stronger test recommendations compared to the main text (45%).
Conclusions: Overinterpretation of results of diagnostic accuracy studies is common. We hope that highlighting the mechanisms behind spin will enable peer reviewers correctly sieve overoptimistic reports of diagnostic accuracy studies and encourage investigators to be more clear in designing, more transparent in reporting, and more stringent in interpreting test accuracy studies.