Do no harm: a roadmap for responsible machine learning for health care

Wiens, Jenna; Saria, Suchi; Sendak, Mark; Ghassemi, Marzyeh; Vincent, Jean‐Louis; Doshi‐Velez, Finale; Jung, Kenneth; Heller, Katherine; Kale, David C.; Saeed, Mohammed; Ossorio, Pilar N.; Thadaney-Israni, Sonoo; Goldenberg, Anna

doi:10.1038/s41591-019-0548-6

Cited by 625 publications

(510 citation statements)

References 14 publications

Supporting

Mentioning

442

Contrasting

Unclassified

Order By: Relevance

“…Our framework provides a reproducible structure to investigators wanting to train, evaluate, and interpret their own ML models to generate hypotheses regarding which OTUs might be biologically relevant. However, deploying microbiome-based models to make clinical diagnoses or predictions is a significantly more challenging and distinct undertaking (41). For example, we currently lack standardized methods to collect patient samples, generate sequence data, and report clinical data.…”

Section: Discussionmentioning

confidence: 99%

A framework for effective application of machine learning to microbiome-based classification problems

Topçuoğlu

Lesniak

Wiens

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

Section: Discussionmentioning

confidence: 99%

A framework for effective application of machine learning to microbiome-based classification problems

Topçuoğlu

Lesniak

Wiens

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…89 Finally, before being integrated into clinical care, a machine learning product needs to be evaluated and validated in a 'silent' mode, "in which predictions are made in real-time and exposed to a group of clinical experts." 14 This period is crucial for finalising workflows and product configurations as well as serving as a temporal validation (Step 2b). An example of a silent mode evaluation is an eCart feasibility study.…”

Section: Evaluate and Validatementioning

confidence: 99%

“…Machine learning technologies that analysed images were excluded. This review also advances prior work to propose best practices for teams building machine learning models within a healthcare setting 14 and for teams conducting quality improvement work following the learning health system framework. 15 However, there is not a unifying translational path to inform teams beyond success within a single setting to diffuse and scale across healthcare.…”

Section: Introductionmentioning

confidence: 98%

A Path for Translation of Machine Learning Products into Healthcare Delivery

Sendak¹,

D’Arcy²,

Kashyap³

et al. 2020

EMJ Innov

View full text Add to dashboard Cite

Despite enormous enthusiasm, machine learning models are rarely translated into clinical care and there is minimal evidence of clinical or economic impact. New conference venues and academic journals have emerged to promote the proliferating research; however, the translational path remains unclear. This review undertakes the first in-depth study to identify how machine learning models that ingest structured electronic health record data can be applied to clinical decision support tasks and translated into clinical practice. The authors complement their own work with the experience of 21 machine learning products that address problems across clinical domains and across geographic populations. Four phases of translation emerge: design and develop, evaluate and validate, diffuse and scale, and continuing monitoring and maintenance. The review highlights the varying approaches taken across each phase by teams building machine learning products and presents a discussion of challenges and opportunities. The translational path and associated findings are instructive to researchers and developers building machine learning products, policy makers regulating machine learning products, and health system leaders who are considering adopting a machine learning product.

show abstract

“…An interesting observation is the fact that by personalizing the ANNs there is no apparent need to worry about generalization and algorithm fairness, i.e. known and unknown biases related to subject age, weight, and other relevant attributes [22].…”

Section: Discussionmentioning

confidence: 99%

Detection of transient neurotransmitter response using personalized neural networks

Klyuzhin

Bevington

Cheng

2019

Preprint

View full text Add to dashboard Cite

Measurement of transient neurotransmitter response (TNR), such as stimulus-related dopamine release, from dynamic positron emission tomography (PET) images typically suffers from limited detection sensitivity and/or high false positive rates. In this work, we perform voxel-level TNR detection using artificial neural networks (ANN), and compare their performance to previously used standard statistical tests. We use acquired and simulated [ 11 C]raclopride (a D2 receptor antagonist) human data to train and test different ANN architectures for the task of dopamine release detection. A distinguishing feature of our approach is the use of "personalized" ANNs that are designed to operate on the dynamic image from a specific subject and scan. Training of personalized ANNs was performed using simulated data that have been matched with an acquired image in terms of the signal and noise. In our tests of TNR detection performance, the F-test of the linear parametric neurotransmitter PET (lp-ntPET) model fit residuals was used as the reference method. For a moderate TNR magnitude, the areas under the receiver operating characteristic curves were 0.77-0.79 for the best ANNs and 0.64 for the F-test. At a fixed false positive rate of 0.01, the true positive rates were 0.8-0.9 for the ANNs and 0.6 for the F-test. When applied to a real image, the ANNs identified an additional TNR cluster compared to the F-test. These results demonstrate that personalized ANNs may offer a greater detection sensitivity of dopamine release and other types of TNR compared to previously used methods.

show abstract

Do no harm: a roadmap for responsible machine learning for health care

Cited by 625 publications

References 14 publications

A framework for effective application of machine learning to microbiome-based classification problems

A framework for effective application of machine learning to microbiome-based classification problems

A Path for Translation of Machine Learning Products into Healthcare Delivery

Detection of transient neurotransmitter response using personalized neural networks

Contact Info

Product

Resources

About