Getting clearer on overdiagnosis

Wendy A. Rogers, Yishai Mintzker

Research output: Contribution to journalArticlepeer-review

42 Scopus citations

Abstract

Overdiagnosis refers to diagnosis that does not benefit patients because the diagnosed condition is not a harmful disease in those individuals. Overdiagnosis has been identified as a problem in cancer screening, diseases such as chronic kidney disease and diabetes, and a range of mental illnesses including depression and attention deficit hyperactivity disorder. In this paper, we describe overdiagnosis, investigate reasons why it occurs, and propose two different types. Misclassification overdiagnosis arises because the diagnostic threshold for the disease in question has been set at a level where many people without harmful disease are nonetheless diagnosed. We illustrate misclassification overdiagnosis using the example of chronic kidney disease. Misclassification occurs in diseases diagnosed using biomarkers or based on patient reported phenomena. Maldetection overdiagnosis arises because, at the time the diagnosis is made and despite the presence of a ‘gold standard’ diagnostic test, it is not possible to discriminate between harmful and non-harmful cases of the index disease. We illustrate maldetection overdiagnosis using the example of thyroid cancer. While there is some overlap between misclassification and maldetection overdiagnosis, this conceptual analysis helps to clarify the phenomenon of overdiagnosis and is a necessary first step in developing strategies to address the problem.

Original languageEnglish
Pages (from-to)580-587
Number of pages8
JournalJournal of Evaluation in Clinical Practice
Volume22
Issue number4
DOIs
StatePublished - 1 Aug 2016

Bibliographical note

Publisher Copyright:
© 2016 John Wiley & Sons, Ltd.

Funding

FundersFunder number
Australian Research CouncilFT130100346

    Keywords

    • diagnosis
    • epistemology
    • philosophy of medicine

    Fingerprint

    Dive into the research topics of 'Getting clearer on overdiagnosis'. Together they form a unique fingerprint.

    Cite this