Journal article
An original approach was used to better evaluate the capacity of a prognostic marker using published survival curves.
- Dantan E Department of Biostatistics, Pharmacoepidemiology and Subjective Measures in Health Sciences, EA 4275, Nantes University, 1 rue Gaston Veil, 44035 Nantes, France. Electronic address: Etienne.Dantan@univ-nantes.fr.
- Combescure C CRC & Division of Clinical Epidemiology, Department of Health and Community Medicine, University of Geneva and University Hospitals of Geneva, rue Gabrielle Perret-Gentil 4, 1211 Geneva, Switzerland.
- Lorent M Department of Biostatistics, Pharmacoepidemiology and Subjective Measures in Health Sciences, EA 4275, Nantes University, 1 rue Gaston Veil, 44035 Nantes, France.
- Ashton-Chess J TcLand Expression, 21 rue Lanoue Bras de Fer, 44200 Nantes, France.
- Daguin P Institute of Transplantation, Urology and Nephrology (ITUN), CHU Nantes and INSERM U1064, 30 bd Jean-Monnet, 44093 Nantes, France.
- Classe JM Department of Surgical Oncology, Institut de Cancérologie de l'Ouest-Cancer Center René Gauducheau, Boulevard Jacques Monod, 44805 Saint-Herblain, France.
- Giral M Institute of Transplantation, Urology and Nephrology (ITUN), CHU Nantes and INSERM U1064, 30 bd Jean-Monnet, 44093 Nantes, France; CIC Biotherapy, CHU Nantes, 30 bd Jean-Monnet, 44093 Nantes, France.
- Foucher Y Department of Biostatistics, Pharmacoepidemiology and Subjective Measures in Health Sciences, EA 4275, Nantes University, 1 rue Gaston Veil, 44035 Nantes, France.
- 2014-03-04
Published in:
- Journal of clinical epidemiology. - 2014
Likelihood ratios
Predictive values
Prognostic factor
Sensitivity
Specificity
Survival analysis
Biomarkers
Chronic Disease
Humans
Kaplan-Meier Estimate
Likelihood Functions
Predictive Value of Tests
Prognosis
Sensitivity and Specificity
English
OBJECTIVES
Predicting chronic disease evolution from a prognostic marker is a key field of research in clinical epidemiology. However, the prognostic capacity of a marker is not systematically evaluated using the appropriate methodology. We proposed the use of simple equations to calculate time-dependent sensitivity and specificity based on published survival curves and other time-dependent indicators as predictive values, likelihood ratios, and posttest probability ratios to reappraise prognostic marker accuracy.
STUDY DESIGN AND SETTING
The methodology is illustrated by back calculating time-dependent indicators from published articles presenting a marker as highly correlated with the time to event, concluding on the high prognostic capacity of the marker, and presenting the Kaplan-Meier survival curves. The tools necessary to run these direct and simple computations are available online at http://www.divat.fr/en/online-calculators/evalbiom.
RESULTS
Our examples illustrate that published conclusions about prognostic marker accuracy may be overoptimistic, thus giving potential for major mistakes in therapeutic decisions.
CONCLUSION
Our approach should help readers better evaluate clinical articles reporting on prognostic markers. Time-dependent sensitivity and specificity inform on the inherent prognostic capacity of a marker for a defined prognostic time. Time-dependent predictive values, likelihood ratios, and posttest probability ratios may additionally contribute to interpret the marker's prognostic capacity.
Predicting chronic disease evolution from a prognostic marker is a key field of research in clinical epidemiology. However, the prognostic capacity of a marker is not systematically evaluated using the appropriate methodology. We proposed the use of simple equations to calculate time-dependent sensitivity and specificity based on published survival curves and other time-dependent indicators as predictive values, likelihood ratios, and posttest probability ratios to reappraise prognostic marker accuracy.
STUDY DESIGN AND SETTING
The methodology is illustrated by back calculating time-dependent indicators from published articles presenting a marker as highly correlated with the time to event, concluding on the high prognostic capacity of the marker, and presenting the Kaplan-Meier survival curves. The tools necessary to run these direct and simple computations are available online at http://www.divat.fr/en/online-calculators/evalbiom.
RESULTS
Our examples illustrate that published conclusions about prognostic marker accuracy may be overoptimistic, thus giving potential for major mistakes in therapeutic decisions.
CONCLUSION
Our approach should help readers better evaluate clinical articles reporting on prognostic markers. Time-dependent sensitivity and specificity inform on the inherent prognostic capacity of a marker for a defined prognostic time. Time-dependent predictive values, likelihood ratios, and posttest probability ratios may additionally contribute to interpret the marker's prognostic capacity.
- Language
-
- English
- Open access status
- closed
- Identifiers
-
- DOI 10.1016/j.jclinepi.2013.10.022
- PMID 24581297
- Persistent URL
- https://sonar.ch/global/documents/206022
Statistics
Document views: 24
File downloads: