David Chan
University of California-Berkeley, CA·Social Science
Doctors who misread a test can miss a disease or diagnose one that is not there. Yet health care has few ways to measure how good individual clinicians are at diagnosis, unlike the many measures that exist for treatment quality.
- Fields
Social Science
Medicine & Clinical Research
Mathematics & Statistics
- Topics
diagnostic quality
diagnostic errors
health care quality
- Grants
- 1 known grant, through Jun 2027
What they work on
- Validate data-driven measures of pneumonia diagnoses and outcomes among 5.5 million chest X-ray visits in Veterans Health Administration emergency departments. Use quasi-random assignment of patients to radiologists to infer differences in false-positive and false-negative rates.
- Interpret each radiologist's error rates in a receiver operating characteristic (ROC) framework, to tell apart errors from choosing the wrong diagnostic threshold and errors from poor diagnostic accuracy.
- Explore what drives diagnostic quality by linking radiologist quality measures to the characteristics and actions of thousands of radiologists.
- Assess the consequences of diagnostic quality by comparing health outcomes of patients quasi-experimentally assigned to radiologists of differing quality.
More about the research
Without such measures, it is hard to spot patterns in diagnostic errors or to test ways of preventing them. This project aims to build and validate a data-driven way to measure diagnostic quality across clinicians.
Why it matters
Better ways to measure diagnostic quality could help health systems and policymakers find where diagnostic errors arise and design and evaluate fixes, such as adjusting how readily clinicians diagnose or improving their accuracy. Patients, starting with emergency department patients evaluated for pneumonia, could benefit from fewer missed and unnecessary diagnoses.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Yulei Jiang
University of Chicago, IL·Mathematics & Statistics
Derive weighted ROC estimation algorithms that assign a weight to each case, covering the non-parametric, conventional binormal and proper binormal models, using maximum-likelihood derivation and validation against Monte Carlo simulations.
Shipra Arya
Stanford University, CA·Medicine & Clinical Research
Test whether greater market competition among surgeons is associated with higher rates of guideline-discordant small abdominal aortic aneurysm repair and tibial interventions for claudication, using linked Vascular Quality Initiative–Medicare and Veterans Health Administration data.
Justin Silverman
Pennsylvania State University, PA·Mathematics & Statistics
Build a theoretical framework for objectively evaluating existing analysis methods for sequence count data in light of scale limitations and measurement bias, the two problems that can dominate study conclusions.
The school
University of California-Berkeley: 46 labs on PI Pages, $1.1B a year on research, 825 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.