Meaningless comparisons lead to false optimism in medical machine learning
Open accessAbstract
A new trend in medicine is the use of algorithms to analyze big datasets, e.g. using everything your phone measures about you for diagnostics or monitoring. However, these algorithms are commonly compared against weak baselines, which may contribute to excessive optimism. To assess how well an algorithm works, scientists typically ask how well its output correlates with medically assigned scores. Here we perform a meta-analysis to quantify how the literature evaluates their algorithms for monitoring mental wellbeing. We find that the bulk of the literature (∼77%) uses meaningless comparisons that ignore patient baseline state. For example, having an algorithm that uses phone data to diagnose mood disorders would be useful. However, it is possible to explain over 80% of the variance of some mood measures in the population by simply guessing that each patient has their own average mood-the patient-specific baseline. Thus, an algorithm that just predicts that our mood is like it usually is can explain the majority of variance, but is, obviously, entirely useless. Comparing to the wrong (population) baseline has a massive effect on the perceived quality of algorithms and produces baseless optimism in the field. To solve this problem we propose "user lift" that reduces these systematic errors in the evaluation of personalized medical monitoring.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
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?
Most previous work has not outperformed individual baselines in similar settings.
“Despite the simplicity of these estimates, most previous work has not outperformed individual baselines in similar settings (42).”
From Monitoring Changes in Depression Severity Using Wearable and Mobile Sensors (2020) · cites this paper as [42]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Breiman, L. (2001) · Machine Learning · doi:10.1023/a:1010933404324 · Cited by 133,538 · see citing works
Regression shrinkage and selection via the lasso
Tibshirani, R. (1996) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1996.tb02080.x · Cited by 51,482 · see citing works
Estimating the Dimension of a Model
Schwarz, G. (1978) · The Annals of Statistics · doi:10.1214/aos/1176344136 · Cited by 47,033 · see citing works
Data Mining: Practical Machine Learning Tools and Techniques
Witten, I. H., Frank, E., & Hall, M. A. (2011) · Elsevier eBooks · doi:10.1016/c2009-0-19715-5 · Cited by 25,493 · see citing works
The Elements of Statistical Learning
Hastie, T., Friedman, J., & Tibshirani, R. (2001) · Springer Series in Statistics · doi:10.1007/978-0-387-21606-5 · Cited by 24,162 · see citing works
An introduction to ROC analysis
Fawcett, T. (2006) · Pattern Recognition Letters · doi:10.1016/j.patrec.2005.10.010 · Cited by 21,879 · see citing works
Regularization and Variable Selection Via the Elastic Net
Zou, H., & Hastie, T. (2005) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.1467-9868.2005.00503.x · Cited by 20,957 · see citing works
He, H., & Garcia, E. A. (2009) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2008.239 · Cited by 10,359 · see citing works
(2007) · doi:10.5555/1283383.1283494 · Cited by 7,364 · see citing works
Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation
Powers, D. M. W. (2020) · arXiv · doi:10.48550/arxiv.2010.16061 · Cited by 5,440 · see citing works
A New Initiative on Precision Medicine
Collins, F. S., & Varmus, H. (2015) · New England Journal of Medicine · doi:10.1056/nejmp1500523 · Cited by 5,202 · see citing works
A Survey on Wearable Sensor-Based Systems for Health Monitoring and Prognosis
Pantelopoulos, A., & Bourbakis, N. G. (2010) · IEEE Transactions on Systems Man and Cybernetics Part C (Applications and Reviews) · doi:10.1109/tsmcc.2009.2032660 · Cited by 2,296 · see citing works
Mining with rarity: a unifying framework
Weiss, G. M. (2004) · ACM SIGKDD Explorations Newsletter · doi:10.1145/1007730.1007734 · Cited by 1,393 · see citing works
Wang, R., Chen, F., Chen, Z., Li, T., Harari, G., Tignor, S., Zhou, X., Ben-Zeev, D., & Campbell, A. T. (2014) · Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing · doi:10.1145/2632048.2632054 · Cited by 1,116 · see citing works
Saeb, S., Zhang, M., Karr, C. J., Schueller, S. M., Corden, M. E., Kording, K. P., & Mohr, D. C. (2015) · Journal of Medical Internet Research · doi:10.2196/jmir.4273 · Cited by 782 · see citing works
Self-citationA comparison of statistical significance tests for information retrieval evaluation
Smucker, M. D., Allan, J., & Carterette, B. (2007) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/1321440.1321528 · Cited by 728 · see citing works
Discovering Shifts to Suicidal Ideation from Mental Health Content in Social Media
Choudhury, M. D., Kıcıman, E., Dredze, M. H., Coppersmith, G., & Kumar, M. (2016) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/2858036.2858207 · Cited by 682 · see citing works
Lu, H., Frauendorfer, D., Rabbi, M., Mast, M. S., Chittaranjan, G. T., Campbell, A. T., Gática-Pérez, D., & Choudhury, T. (2012) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/2370216.2370270 · Cited by 511 · see citing works
Canzian, L., & Musolesi, M. (2015) · Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing · doi:10.1145/2750858.2805845 · Cited by 477 · see citing works
Social fMRI: Investigating and shaping social mechanisms in the real world
Aharony, N., Pan, W., Ip, C., Khayal, I. S., & Pentland, A. (2011) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2011.09.004 · Cited by 458 · see citing works
MoodScope: building a mood sensor from smartphone usage patterns
LiKamWa, R., Liu, Y., Lane, N. D., & Zhong, L. (2013) · Proceeding of the 11th annual international conference on Mobile systems, applications, and services · doi:10.1145/2462456.2464449 · Cited by 444 · see citing works
Machine Learning and the Profession of Medicine
Darcy, A. M., Louie, A. K., & Roberts, L. W. (2016) · JAMA · doi:10.1001/jama.2015.18421 · Cited by 436 · see citing works
Ben-Zeev, D., Scherer, E. A., Wang, R., Xie, H., & Campbell, A. T. (2015) · Psychiatric Rehabilitation Journal · doi:10.1037/prj0000130 · Cited by 429 · see citing works
Characterizing and predicting postpartum depression from shared facebook data
Choudhury, M. D., Counts, S., Horvitz, E., & Hoff, A. (2014) · ACM Conference on Computer Supported Cooperative Work (CSCW) · doi:10.1145/2531602.2531675 · Cited by 376 · see citing works
The need to approximate the use-case in clinical machine learning
Saeb, S., Lonini, L., Jayaraman, A., Mohr, D. C., & Kording, K. P. (2017) · GigaScience · doi:10.1093/gigascience/gix019 · Cited by 367 · see citing works
Self-citationProceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms
Gabow, H. N. (2007) · Cited by 344 · see citing works
Mobile mental health: a challenging research agenda
Olff, M. (2015) · European Journal of Psychotraumatology · doi:10.3402/ejpt.v6.27882 · Cited by 258 · see citing works
Daily Stress Recognition from Mobile Phone Data, Weather Conditions and Individual Traits
Bogomolov, A., Lepri, B., Ferron, M., Pianesi, F., & Pentland, A. (. (2014) · Proceedings of the 22nd ACM international conference on Multimedia · doi:10.1145/2647868.2654933 · Cited by 217 · see citing works
Call Center Stress Recognition with Person-Specific Models
Hernandez, J., Morris, R. R., & Picard, R. W. (2011) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-24600-5_16 · Cited by 210 · see citing works
The application of mHealth to mental health: opportunities and challenges
Marzano, L., Bardill, A., Fields, B., Herd, K., Veale, D., Grey, N., & Moran, P. (2015) · The Lancet Psychiatry · doi:10.1016/s2215-0366(15)00268-0 · Cited by 200 · see citing works
Asselbergs, J., Ruwaard, J., Ejdys, M., Schrader, N., Sijbrandij, M., & Riper, H. (2016) · Journal of Medical Internet Research · doi:10.2196/jmir.5505 · Cited by 192 · see citing works
Kessler, R. C., Stein, M. B., Petukhova, M. V., Bliese, P., Bossarte, R. M., Bromet, E. J., Fullerton, C. S., Gilman, S. E., Ivany, C., Lewandowski-Romps, L., et al. (2017) · Molecular Psychiatry · doi:10.1038/mp.2016.110 · Cited by 163 · see citing works
Valenza, G., Nardelli, M., Lanatá, A., Gentili, C., Bertschy, G., Paradiso, R., & Scilingo, E. (2014) · IEEE journal of biomedical and health informatics · doi:10.1109/jbhi.2013.2290382 · Cited by 158 · see citing works
Predicting students' happiness from physiology, phone, mobility, and behavioral data
Jaques, N., Taylor, S., Azaria, A., Ghandeharioun, A., Sano, A., & Picard, R. W. (2015) · International Conference on Affective Computing and Intelligent Interaction and workshops · doi:10.1109/acii.2015.7344575 · Cited by 140 · see citing works
Stress Detection Using Wearable Physiological Sensors
Săndulescu, V., Andrews, S., Ellis, D. A., Bellotto, N., & Mozos, Ó. M. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-18914-7_55 · Cited by 135 · see citing works
Ruling Out or Ruling In Disease with the Most sensitiue or Specific Diagnostic Test
Boyko, E. J. (1994) · Medical Decision Making · doi:10.1177/0272989x9401400210 · Cited by 124 · see citing works
The behaviour of some significance tests under experimental randomization
Kempthorne, O., & Doerfler, T. E. (1969) · Biometrika · doi:10.1093/biomet/56.2.231 · Cited by 115 · see citing works
Valenza, G., Gentili, C., Lanatà, A., & Scilingo, E. P. (2013) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2012.12.001 · Cited by 95 · see citing works
The Prediction of Suicide: Why Is It so Difficult?
Murphy, G. E. (1984) · American Journal of Psychotherapy · doi:10.1176/appi.psychotherapy.1984.38.3.341 · Cited by 77 · see citing works
Food and Mood: Just-in-Time Support for Emotional Eating
Carroll, E. A., Czerwinski, M., Roseway, A., Kapoor, A., Johns, P., Rowan, K., & Schraefel, M. C. (2013) · 2013 Humaine Association Conference on Affective Computing and Intelligent Interaction · doi:10.1109/acii.2013.48 · Cited by 64 · see citing works
Pervasive stress recognition for sustainable living
Bogomolov, A., Lepri, B., Ferron, M., Pianesi, F., & Pentland, A. (2014) · 2014 IEEE International Conference on Pervasive Computing and Communication Workshops (PERCOM WORKSHOPS) · doi:10.1109/percomw.2014.6815230 · Cited by 53 · see citing works
An integrated framework for suicide risk prediction
Tran, T., Phung, D., Luo, W., Harvey, R. J., Berk, M., & Venkatesh, S. (2013) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/2487575.2488196 · Cited by 47 · see citing works
Modeling perceived stress via HRV and accelerometer sensor streams
Wu, M., Cao, H., Nguyen, H.-L., Surmacz, K., & Hargrove, C. (2015) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2015.7318686 · Cited by 47 · see citing works
Calvo, R. A., Dinakar, K., Picard, R. W., & Maes, P. (2016) · CHI Conference Extended Abstracts on Human Factors in Computing Systems (CHI EA) · doi:10.1145/2851581.2856463 · Cited by 38 · see citing works
Sleep, mood and sociability in a healthy population
Moturu, S. T., Khayal, I., Aharony, N., Pan, W., & Pentland, A. (2011) · 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society · doi:10.1109/iembs.2011.6091303 · Cited by 35 · see citing works
Evaluating feature selection for stress identification
Deng, Y. J., Wu, Z., Chu, C., et al. (2012) · IEEE International Conference on Information Reuse and Integration (IRI) · doi:10.1109/iri.2012.6303062 · Cited by 33 · see citing works
The Behaviour of Some Significance Tests Under Experimental Randomization
Kempthorne, O., & Doerfler, T. E. (1969) · Biometrika · doi:10.2307/2334417 · Cited by 5 · see citing works
Flagged first, then by citations
Research works printed but not linked: 3
- [32]Saeb S, Lonini L, Jayaraman A, Mohr DC, Kording KP. Voodoo Machine Learning for Clinical Predictions. bioRxiv. 2016;
- [36]Jaques N, Taylor S, Sano A, Picard R. Multi-task, Multi-Kernel Learning for Estimating Individual Wellbeing. In: Proc. NIPS Workshop on Multimodal Machine Learning, Montreal, Quebec; 2015.
- [37]Lu H, Frauendorfer D, Rabbi M, Mast MS, Chittaranjan GT, Campbell AT, et al. Stresssense: Detecting stress in unconstrained acoustic environments using smartphones. In: Proceedings of the 2012 ACM Conference on Ubiquitous Computing. ACM; 2012. p. 351–360.
Cited by
Sorted by citations
Monitoring Changes in Depression Severity Using Wearable and Mobile Sensors
Pedrelli, P., Fedor, S., Ghandeharioun, A., Howe, E., Ionescu, D. F., Bhathena, D., Fisher, L. B., Cusin, C., Nyer, M., Yeung, A., et al. (2020) · Frontiers in Psychiatry · doi:10.3389/fpsyt.2020.584711 · Cited by 158 · see citing works
Pearson, R., Pisner, D., Meyer, B., Shumake, J., & Beevers, C. G. (2019) · Psychological Medicine · doi:10.1017/s003329171800315x · Cited by 81 · see citing works
Ebert, D. D., Harrer, M., Apolinário-Hagen, J., & Baumeister, H. (2019) · Advances in Experimental Medicine and Biology · doi:10.1007/978-981-32-9721-0_29 · Cited by 76 · see citing works
- Authors
- DeMasi, O., Kording, K., & Recht, B.
- Affiliations
- University of California, Berkeley (US)
- University of Pennsylvania (US)
- Corresponding author
- DeMasi, Orianna
- Venue
- PLoS ONE
- Publisher
- Public Library of Science (PLoS)
- Published
- 2017-09-26
- Volume
- 12
- Issue
- 9
- Pages
- e0184604
- ISSN
- 1932-6203 (ISSN-L)
- PMC
- PMC5614525
- OpenAlex
- W2739036537
- Cited by
- 63
- References
- 47
- Citations by year
- 2026: 1 · 2025: 2 · 2024: 7 · 2023: 7 · 2022: 5 · 2021: 9 · 2020: 18 · 2019: 5 · 2018: 6 · 2017: 3
- FWCI
- 12.01
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- PLoS ONE (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- arXiv (Cornell University)
- Funders
- National Science Foundation (1359814)
- National Institutes of Health
- Defense Advanced Research Projects Agency
- Office of Naval Research (17-1-2191, N00014-17-1-2191, N00014-14-1-0024, N00014-17-1, N00014, N00014-17)
- National Institute of Neurological Disorders and Stroke (R01NS063399)
- Topics
- Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
- Social Sciences › Psychology › Experimental and Cognitive Psychology › Mental Health Research Topics
- Social Sciences › Psychology › Experimental and Cognitive Psychology › Emotion and Mood Recognition
- Keywords
- algorithm evaluation, psychological well-being, mood disorders, smartphone data
- MeSH terms
- Machine Learning, Algorithms, Humans, Mood Disorders
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
