Use of deep learning to develop continuous-risk models for adverse event prediction from electronic health records

Tomašev, Nenad; Harris, Natalie; Baur, Sebastien; Mottram, Anne; Glorot, Xavier; Rae, Jack W.; Zieliński, Michał; Askham, Harry; Saraiva, André; Magliulo, Valerio; Meyer, Clemens; Ravuri, Suman; Protsyuk, Ivan; Connell, Alistair; Hughes, Cían; Karthikesalingam, Alan; Cornebise, Julien; Montgomery, Hugh; Rees, Geraint; Laing, Chris; Baker, Clifton R.; Osborne, Thomas F.; Reeves, Ruth; Hassabis, Demis; King, Dominic; Suleyman, Mustafa; Back, Trevor; Nielson, Christopher; Seneviratne, Martin; Ledsam, Joseph R.; Mohamed, Shakir

doi:10.1038/s41596-021-00513-5

Cited by 66 publications

(59 citation statements)

References 94 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another area of potential future significance for machine learning-based DRP is the prediction of unwanted side effects or adverse drug reactions . Although methods for the prediction of more general adverse patient events currently appear to focus on electronic health records [ 165 ], increasing insight into the molecular basis of adverse drug reactions [ 166 ] and associated data may also permit the future integration with principles used today for DRP.…”

Section: Discussion and Outlookmentioning

confidence: 99%

An overview of machine learning methods for monotherapy drug response prediction

Firoozbakht

Yousefi

Schwikowski

2021

Briefings in Bioinformatics

View full text Add to dashboard Cite

For an increasing number of preclinical samples, both detailed molecular profiles and their responses to various drugs are becoming available. Efforts to understand, and predict, drug responses in a data-driven manner have led to a proliferation of machine learning (ML) methods, with the longer term ambition of predicting clinical drug responses. Here, we provide a uniquely wide and deep systematic review of the rapidly evolving literature on monotherapy drug response prediction, with a systematic characterization and classification that comprises more than 70 ML methods in 13 subclasses, their input and output data types, modes of evaluation, and code and software availability. ML experts are provided with a fundamental understanding of the biological problem, and how ML methods are configured for it. Biologists and biomedical researchers are introduced to the basic principles of applicable ML methods, and their application to the problem of drug response prediction. We also provide systematic overviews of commonly used data sources used for training and evaluation methods.

show abstract

Section: Discussion and Outlookmentioning

confidence: 99%

An overview of machine learning methods for monotherapy drug response prediction

Firoozbakht

Yousefi

Schwikowski

2021

Briefings in Bioinformatics

View full text Add to dashboard Cite

show abstract

“…By sharing our own experience as an example, we hope individuals trained in genetic counseling may continue to see opportunities for adaptation of their skill set into expanding roles. Although there is a genetic counseling workforce shortage that should not be overlooked, we must also acknowledge that there are a range of important reasons why genetic counselors may be compelled to (Gordon et al, 2018), wearable or ambient sensors used in the diagnosis and care of neurological disorders (Dorsey et al, 2020;Husebo et al, 2019), or application of artificial intelligence to electronic health record data for improving patient care (Jacoba et al, 2021;Lin et al, 2020;Tomašev et al, 2021).…”

Section: Discussionmentioning

confidence: 99%

Adapting skills from genetic counseling to wearables technology research during the COVID‐19 pandemic: Poised for the pivot

Higgs

Dagan‐Rosenfeld

Snyder

2021

Journal of Genetic Counseling

View full text Add to dashboard Cite

Genetic counselors have shown themselves to be adaptable in an evolving profession, with expansion into new sub-specialties, various non-clinical settings, and research roles. The COVID-19 pandemic caused a sudden and drastic shift in healthcare priorities. In an effort to contribute meaningfully to the COVID-19 crisis, and to adapt to a remote-and essential-only research environment, our workplace and thus our roles pivoted from genomics research to remote COVID-19 research using wearables technologies. With a deep understanding of genomic data, we were quickly able to apply similar concepts to wearables data including considering privacy implications, managing uncertain findings, and acknowledging the lack of ethnic diversity in many datasets. By sharing our own experience as an example, we hope individuals trained in genetic counseling may see opportunities for adaptation of their skills into expanding roles.

show abstract

“…Thus, they manipulated feature weights to see how doing so affects the predictive performance of the model. Causal knowledge was thus primarily used to identify spurious correlations and to improve model performance, as further described in more recent work by some of the same team (Tomašev et al, 2021 ). The paradigm apparent in the DeepMind approach is not “is there a causal association between these variables?”, but “does inclusion of this feature affect predictive performance?”.…”

Section: Epidemiological Applications Of MLmentioning

confidence: 99%

Can Robots Do Epidemiology? Machine Learning, Causal Inference, and Predicting the Outcomes of Public Health Interventions

Broadbent

Grote

2022

Philos. Technol.

View full text Add to dashboard Cite

This paper argues that machine learning (ML) and epidemiology are on collision course over causation. The discipline of epidemiology lays great emphasis on causation, while ML research does not. Some epidemiologists have proposed imposing what amounts to a causal constraint on ML in epidemiology, requiring it either to engage in causal inference or restrict itself to mere projection. We whittle down the issues to the question of whether causal knowledge is necessary for underwriting predictions about the outcomes of public health interventions. While there is great plausibility to the idea that it is, conviction that something is impossible does not by itself motivate a constraint to forbid trying. We disambiguate the possible motivations for such a constraint into definitional, metaphysical, epistemological, and pragmatic considerations and argue that “Proceed with caution” (rather than “Stop!”) is the outcome of each. We then argue that there are positive reasons to proceed, albeit cautiously. Causal inference enforces existing classification schema prior to the testing of associational claims (causal or otherwise), but associations and classification schema are more plausibly discovered (rather than tested or justified) in a back-and-forth process of gaining reflective equilibrium. ML instantiates this kind of process, we argue, and thus offers the welcome prospect of uncovering meaningful new concepts in epidemiology and public health—provided it is not causally constrained.

show abstract

Use of deep learning to develop continuous-risk models for adverse event prediction from electronic health records

Cited by 66 publications

References 94 publications

An overview of machine learning methods for monotherapy drug response prediction

An overview of machine learning methods for monotherapy drug response prediction

Adapting skills from genetic counseling to wearables technology research during the COVID‐19 pandemic: Poised for the pivot

Can Robots Do Epidemiology? Machine Learning, Causal Inference, and Predicting the Outcomes of Public Health Interventions

Contact Info

Product

Resources

About