Using natural frequencies to improve diagnostic inferences
Open accessAbstract
PURPOSE: To test whether physician's diagnostic inferences can be improved by communicating information using natural frequencies instead of probabilities. Whereas probabilities and relative frequencies are normalized with respect to disease base rates, natural frequencies are not normalized. METHOD: The authors asked 48 physicians in Munich and Düsseldorf to determine the positive predictive values (PPVs) of four diagnostic tests. Information presented in the four problems appeared either as probabilities (the traditional way) or as natural frequencies. RESULTS: When the information was presented as probabilities, the physicians correctly estimated the PPVs in only 10% of cases. When the same information was presented as natural frequencies, that percentage increased to 46%. CONCLUSION: Representing information in natural frequencies is a fast and effective way of facilitating diagnosis insight, which in turn helps physicians to better communicate risks to patients, and patients to better understand these risks.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The work in [178] shows that using probabilities to communicate results led users to estimate outcomes correctly in 10% of cases, versus 46% when using natural frequencies.
“the work in [178] shows that the usage of probabilities to communicate the results, implied that the users where able to estimate the outcomes correctly in 10% of the cases, as opposed to 46% of the cases when using natural frequencies.”
From Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI (2020) · cites this paper as [178]
Cited by
Sorted by citations
Barredo Arrieta, A., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., Garcia, S., Gil-Lopez, S., Molina, D., Benjamins, R., et al. (2020) · Information Fusion · doi:10.1016/j.inffus.2019.12.012 · Cited by 10,423 · see citing works
“For instance, the work in [178] shows that the usage of probabilities to communicate the results, implied that the users where able to estimate the outcomes correctly in 10% of the cases, as opposed to 46% of the cases when using natural frequencies.”
Helping Doctors and Patients Make Sense of Health Statistics
Gigerenzer, G., Gaissmaier, W., Kurz-Milcke, E., Schwartz, L. M., & Woloshin, S. (2007) · Psychological Science in the Public Interest · doi:10.1111/j.1539-6053.2008.00033.x · Cited by 1,336 · see citing works
Self-citationGRADE guidelines: 12. Preparing Summary of Findings tables—binary outcomes
Guyatt, G. H., Oxman, A. D., Santesso, N., Helfand, M., Vist, G., Kunz, R., Brozek, J., Norris, S., Meerpohl, J., Djulbegovic, B., et al. (2013) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2012.01.012 · Cited by 836 · see citing works
Hullman, J., Resnick, P., & Adar, E. (2015) · PLOS ONE · doi:10.1371/journal.pone.0142444 · Cited by 194 · see citing works
Hypothetical Outcome Plots Help Untrained Observers Judge Trends in Ambiguous Data
Kale, A., Nguyen, F., Kay, M., & Hullman, J. (2019) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2018.2864909 · Cited by 122 · see citing works
Survey on Explainable AI: Techniques, challenges and open issues
Abusitta, A., Li, M. Q., & Fung, B. C. M. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.124710 · Cited by 98 · see citing works
Procurement As Policy: Administrative Process for Machine Learning
Mulligan, D. K., & Bamberger, K. A. (2019) · SSRN Electronic Journal · doi:10.2139/ssrn.3464203 · Cited by 75 · see citing works
Hullman, J., Kay, M., Kim, Y.-S., & Shrestha, S. (2018) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2017.2743898 · Cited by 55 · see citing works
- Authors
- Hoffrage, U., & Gigerenzer, G.
- Affiliations
- Max Planck Society (DE)
- Venue
- Academic Medicine
- Publisher
- Oxford University Press (OUP)
- Published
- 1998-05
- Volume
- 73
- Issue
- 5
- Pages
- 538-40
- ISSN
- 1040-2446 (ISSN-L)
- OpenAlex
- W1973727966
- Cited by
- 501
- References
- 0
- Citations by year
- 2026: 11 · 2025: 15 · 2024: 18 · 2023: 18 · 2022: 27 · 2021: 20 · 2020: 25 · 2019: 63 · 2018: 20 · 2017: 15 · 2016: 15 · 2015: 26 · 2014: 14 · 2013: 19 · 2012: 35
- FWCI
- 8.56
- Type
- journal article
- Language
- en
- Open access
- green
- hdl.handle.net
- Licence
- other-oa
- Locations
- Academic Medicine (publishedVersion)
- PubMed (publishedVersion)
- Max Planck Institute for Plasma Physics (submittedVersion)
- Max Planck Digital Library (submittedVersion, other-oa)
- MPG.PuRe (Max Planck Society) (submittedVersion, other-oa)
- Topics
- Health Sciences › Medicine › Family Practice › Clinical Reasoning and Diagnostic Skills
- Health Sciences › Health Professions › General Health Professions › Patient-Provider Communication in Healthcare
- Social Sciences › Decision Sciences › General Decision Sciences › Decision-Making and Behavioral Economics
- Keywords
- natural frequencies, diagnostic reasoning, positive predictive value, diagnostic testing, base rates, probability communication, physician decision-making, risk communication, numeracy
- MeSH terms
- Bayes Theorem, Breast Neoplasms, Diagnosis, False Positive Reactions, Humans, Mammography, Predictive Value of Tests, Probability, Risk
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
