Acceso usuarios
Blanca Lumbreras Lacarra 
, Ildefonso Hernández Aguado , Inmaculada Jarrín
Introduction: Advances in genomic analysis technologies have led to the development of new diagnostic tests with clinical application. Therefore, as in other diagnostic fields, awareness of the methodological limitations of genetic investigation will facilitate the application of the results.
Methods: 44 articles which studied the diagnostic accuracy of genetic, molecular and proteomic tests, and published in JAMA, Lancet, New England Journal of Medicine, Cancer Research y Clinical Cancer Research from 2002 to June 2005 were analysed. 24 methodological criteria of the STARD guide (Standards for Reporting of Diagnostic Accuracy) were applied.
Results: The mean number of methodological criteria satisfied was 9.8 (95%, CI 8.8-10.6), with the greatest deficiencies observed in the aspects related to the description of patient selection, 9 (20%), the treatment of indeterminate results, 5 (11%), and the determination of the technique's reproducibility, 6 (13%). In contrast, a high frequency was observed in the description of the reference standard, 39 (87%), and the method used, 28 (62%).
Discussion: The articles evaluated fail to fulfil many of the quality requirements laid out in the STARD proposal, with the methodological quality being lower than in other diagnostic fields. The aspects most in need of improvement are those related to the description of patient selection and the determination of reproducibility. Research and progress in new genetic-molecular technologies require improved fulfilment of the epidemiological and clinical standards which are already applied by other diagnostic fields.
