Estimating diagnostic accuracy without a gold standard: A continued controversy.

Area of Science:

• Biostatistics
• Epidemiology
• Diagnostic Test Evaluation

Background:

• Estimating diagnostic accuracy (sensitivity, specificity) without a gold standard is crucial when gold standard data is unobtainable or costly.
• Existing methodologies for diagnostic accuracy assessment without a gold standard are widely used but face significant criticism.
• This field requires careful consideration of methodological advancements and inherent limitations.

Purpose of the Study:

• To review key methodological developments in estimating diagnostic accuracy without a gold standard.
• To discuss the criticisms and challenges associated with these non-gold standard approaches.
• To propose alternative, less controversial study designs and outline future research directions.

Main Methods:

• Literature review of biostatistical and epidemiological studies on diagnostic accuracy estimation.
• Critical analysis of existing methodologies and their limitations.
• Development of alternative study designs and recommendations for practice.

Main Results:

• Identified significant methodological advancements in the field.
• Highlighted substantial criticisms and controversies surrounding current non-gold standard methods.
• Proposed alternative designs and future research avenues.

Conclusions:

• Methods for estimating diagnostic accuracy without a gold standard offer practical advantages but require careful application due to criticisms.
• Alternative designs may provide more robust and less controversial solutions.
• Recommendations are provided for the judicious use of these methods in biostatistical and epidemiological research.