BACKGROUND: We examined whether and to what extent different strategies ofdefining and incorporating quality of included studies affect the results ofmetaanalyses of diagnostic accuracy.METHODS: We evaluated the methodological quality of 487 diagnostic-accuracystudies in 30 systematic reviews with the QUADAS (Quality Assessment ofDiagnostic-Accuracy Studies) checklist. We applied 3 strategies that varied both in the definition of quality and in the statistical approach to incorporate thequality-assessment results into metaanalyses. We compared magnitudes ofdiagnostic odds ratios, widths of their confidence intervals, and changes in ahypothetical clinical decision between strategies.RESULTS: Following 2 definitions of quality, we concluded that only 70 or 72 of487 studies were of "high quality". This small number was partly due to poorreporting of quality items. None of the strategies for accounting for differencesin quality led systematically to accuracy estimates that were less optimisticthan ignoring quality in metaanalyses. Limiting the review to high-qualitystudies considerably reduced the number of studies in all reviews, with widerconfidence intervals as a result. In 18 reviews, the quality adjustment wouldhave resulted in a different decision about the usefulness of the test.CONCLUSIONS: Although reporting the results of quality assessment of individualstudies is necessary in systematic reviews, reader wariness is warrantedregarding claims that differences in methodological quality have been accountedfor. Obstacles for adjusting for quality in metaanalyses are poor reporting ofdesign features and patient characteristics and the relatively low number ofstudies in most diagnostic reviews.
Impact of adjustment for quality on results of metaanalyses of diagnostic accuracy
Rutjes A;
2007-01-01
Abstract
BACKGROUND: We examined whether and to what extent different strategies ofdefining and incorporating quality of included studies affect the results ofmetaanalyses of diagnostic accuracy.METHODS: We evaluated the methodological quality of 487 diagnostic-accuracystudies in 30 systematic reviews with the QUADAS (Quality Assessment ofDiagnostic-Accuracy Studies) checklist. We applied 3 strategies that varied both in the definition of quality and in the statistical approach to incorporate thequality-assessment results into metaanalyses. We compared magnitudes ofdiagnostic odds ratios, widths of their confidence intervals, and changes in ahypothetical clinical decision between strategies.RESULTS: Following 2 definitions of quality, we concluded that only 70 or 72 of487 studies were of "high quality". This small number was partly due to poorreporting of quality items. None of the strategies for accounting for differencesin quality led systematically to accuracy estimates that were less optimisticthan ignoring quality in metaanalyses. Limiting the review to high-qualitystudies considerably reduced the number of studies in all reviews, with widerconfidence intervals as a result. In 18 reviews, the quality adjustment wouldhave resulted in a different decision about the usefulness of the test.CONCLUSIONS: Although reporting the results of quality assessment of individualstudies is necessary in systematic reviews, reader wariness is warrantedregarding claims that differences in methodological quality have been accountedfor. Obstacles for adjusting for quality in metaanalyses are poor reporting ofdesign features and patient characteristics and the relatively low number ofstudies in most diagnostic reviews.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
