PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 10845402

A contaminated binormal model for ROC data: Part II. A formal model.

D D Dorfman·K S Berbaum
View original source ↗Export metadata ↓

Abstract

RATIONALE AND OBJECTIVES: A contaminated binormal receiver operating characteristic (ROC) model is proposed to account for ROC data that have very few false-positive reports even though many healthy subjects are sampled. The model assumes that no signal information is captured for a proportion of abnormalities, and that these abnormalities have the same distribution as noise along the latent decision axis. MATERIALS AND METHODS: The authors developed a formal psychophysical model, presented here in detail. They have specified the psychophysical assumptions of the theory, and have provided proofs that include all essential details, from assumptions to implications. With the technical details that are provided, this theory can be implemented with computer programs to fit data. RESULTS: The new model can fit ROC data in which some or all of the ROC points have false-positive fractions of 0 and true-positive fractions of less than 1, without implying that performance is perfect. The resulting ROC curves are always proper, never exhibiting inappropriate chance line crossings. The model predicts that, under certain conditions, a bimodal categorical rating histogram will be observed for the signal distribution. The model predicts a relationship between the mean and standard deviation of the signal distribution and holds that, for expert decision makers, there are situations in which the prevalence and utility matrix preclude operating points in some ROC regions. The model has a straightforward extension to the joint detection and localization ROC curve. CONCLUSION: The contaminated binormal model accounts for ROC data with few or no false-positive reports.

Explore related subjects

Decision TheoryDiagnostic ImagingHumansModels, StatisticalPsychophysicsROC CurveComparative StudyJournal ArticleResearch Support, U.S. Gov't, Non-P.H.S.Research Support, U.S. Gov't, P.H.S.

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

D D Dorfman, K S Berbaum. 2000. A contaminated binormal model for ROC data: Part II. A formal model.. https://doi.org/10.1016/s1076-6332(00)80383-9

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2006

What does the precautionary principle mean for evidence-based dentistry?

The precautionary principle calls for preventive actions in the face of uncertain information about risks. It serves as a compass to better guide more health-protective decisions in the face of complex risks. Applying precaution requires thinking more broadly about risks, taking an interdisciplinary approach to science and policy, and considering a wide range of alternatives to potentially harmful activities. While often criticized as antiscientific, the precautionary principle represents a challenge to scientists and public health professionals to develop newer and more effective tools for characterizing and preventing complex risks, in addition to being more explicit about uncertainties. This article examines the role and application of precaution in the context of dental practice, where activities that may convey risks also have public health benefits, and risk trade offs are a possibility. We conclude that the precautionary principle is not at odds with, but rather complements evidence-based practice in situations of scientific uncertainty and complex risks.

Decision Theory↗
PubMed2006

Science-policy in environmental and health risk assessment: if we cannot do without, can we do better?

How can empirical evidence of adverse effects from exposure to noxious agents, which is often incomplete and uncertain, be used most appropriately to protect human health? We examine several important questions on the best uses of empirical evidence in regulatory risk management decision-making raised by the US Environmental Protection Agency (EPA)'s science-policy concerning uncertainty and variability in human health risk assessment. In our view, the US EPA (and other agencies that have adopted similar views of risk management) can often improve decision-making by decreasing reliance on default values and assumptions, particularly when causation is uncertain. This can be achieved by more fully exploiting decision-theoretic methods and criteria that explicitly account for uncertain, possibly conflicting scientific beliefs and that can be fully studied by advocates and adversaries of a policy choice, in administrative decision-making involving risk assessment. The substitution of decision-theoretic frameworks for default assumption-driven policies also allows stakeholder attitudes toward risk to be incorporated into policy debates, so that the public and risk managers can more explicitly identify the roles of risk-aversion or other attitudes toward risk and uncertainty in policy recommendations. Decision theory provides a sound scientific way explicitly to account for new knowledge and its effects on eventual policy choices. Although these improvements can complicate regulatory analyses, simplifying default assumptions can create substantial costs to society and can prematurely cut off consideration of new scientific insights (e.g., possible beneficial health effects from exposure to sufficiently low 'hormetic' doses of some agents). In many cases, the administrative burden of applying decision-analytic methods is likely to be more than offset by improved effectiveness of regulations in achieving desired goals. Because many foreign jurisdictions adopt US EPA reasoning and methods of risk analysis, it may be especially valuable to incorporate decision-theoretic principles that transcend local differences among jurisdictions.

Decision Theory↗