Endocrine Diagnostic Systematic Reviews: An Audit of Methods, Reporting, and Performance

Presentation Number: SUN 292
Date of Presentation: April 2nd, 2017

Gabriela Spencer-Bonilla*1, Naykky Maruquel Singh Ospina2, Rene Rodriguez-Gutierrez3, Juan Pablo Brito1, Patricia J Erwin1, Nicole M Iniguez-Ariza1, M. Hassan Murad1, Shrikant Tamhane1 and Victor M Montori1
1Mayo Clinic, Rochester, MN, 2University of Florida, Gainesville, FL, 3University Hospital “Dr. Jose E. Gonzalez”, Monterrey, Mexico


By rigorously summarizing the existing body of evidence, systematic reviews (SRs) provide clinicians and policymakers with estimates of the accuracy of diagnostic tests and of their usefulness in clinical practice. These summaries prevent overreliance on the largest, the latest or the most well-known study rather than on the body of evidence. The quality of diagnostic SRs in endocrinology, however, is unclear.

We audited all available SRs of endocrine test performance to describe their methods, reporting, and results.

We searched for SRs reporting measures of diagnostic test performance in endocrinology published until December 2015. Trained reviewers, working independently and in duplicate, identified eligible studies and extracted data about methods, reporting, and test performance.

We identified 83 SRs, including 42 of tests for thyroid conditions. Most reviews adequately reported how studies were identified and selected and how their trustworthiness was judged. Only 1 in 3 reported an overall judgment about trustworthiness. Only 1 in 6 reported using adequate meta-analytic methods. Pooled sensitivity and specificity and summary ROC curves were the most commonly reported summary statistics. Four in 10 reviews assessed for publication bias and 2 out of 10 reported contacting authors for further information. Half of the tests were classified as helpful when positive (i.e., likelihood ratio of a positive test result >10); and one fourth as helpful when negative (i.e., likelihood ratio of a negative test result <0.01). The most useful test, when positive or negative was the composite of free cortisol, salivary midnight cortisol and dexamethasone suppression test for ruling in or out Cushing Syndrome.

Up to half of the diagnostic endocrine tests evaluated formally appear to be useful when positive as do one-quarter when negative. Unfortunately, the trustworthiness of these encouraging results relies on reviews of inconsistent methodological quality, many using inadequate statistical approaches, and infrequently reporting about the quality of the studies pooled. Substantial efforts seem necessary to improve the quality of diagnostic SRs in endocrinology as well as the ability of peers to referee them, and of users to consume them.


Nothing to Disclose: GS, NMS, RR, JPB, PJE, NMI, MHM, ST, VMM