Federated learning of predictive models from federated Electronic Health Records

Brisimi, Theodora S.; Chen, Ruidi; Mela, Theofanie; Olshevsky, Alex; Paschalidis, Ioannis Ch.; Shi, Wei

doi:10.1016/j.ijmedinf.2018.01.007

Cited by 618 publications

(322 citation statements)

References 16 publications

Supporting

Mentioning

318

Contrasting

Order By: Relevance

“…Many models have been used to predict a patient admission to a hospital, mortality and other health care applications based on comorbidities. Some examples include: predicting morbidity of patients with chronic obstructive pulmonary disease [6], febrile neutropenia [7], as well as classifying the hospitalization of patients with preconditions on diabetes [8], heart disease [9,10], and hospital readmission for patients with mental or substance use disorders [11]. Recent advances in the machine learning literature have suggested that sparse classifiers, those that use few variables (e.g., l1-regularized Support Vector Machines), have stronger predictive power and generalize better on out-of-sample data points than very complex classifiers [12].…”

Section: Introductionmentioning

confidence: 99%

Personalized Predictive Models for Symptomatic COVID-19 Patients Using Basic Preconditions:Hospitalizations, Mortality, and the Need for an ICU or Ventilator

Wollenstein-Betech

Cassandras

Paschalidis

2020

Preprint

Self Cite

View full text Add to dashboard Cite

Background: The rapid global spread of the virus SARS-CoV-2 has provoked a spike in demand for hospital care. Hospital systems across the world have been over-extended, including in Northern Italy, Ecuador, and New York City, and many other systems face similar challenges. As a result, decisions on how to best allocate very limited medical resources have come to the forefront. Specifically, under consideration are decisions on who to test, who to admit into hospitals, who to treat in an Intensive Care Unit (ICU), and who to support with a ventilator. Given today's ability to gather, share, analyze and process data, personalized predictive models based on demographics and information regarding prior conditions can be used to (1) help decision-makers allocate limited resources, when needed, (2) advise individuals how to better protect themselves given their risk profile, (3) differentiate social distancing guidelines based on risk, and (4) prioritize vaccinations once a vaccine becomes available. Objective: To develop personalized models that predict the following events: (1) hospitalization, (2) mortality, (3) need for ICU, and (4) need for a ventilator. To predict hospitalization, it is assumed that one has access to a patient's basic preconditions, which can be easily gathered without the need to be at a hospital. For the remaining models, different versions developed include different sets of a patient's features, with some including information on how the disease is progressing (e.g., diagnosis of pneumonia). Materials and Methods: Data from a publicly available repository, updated daily, containing information from approximately 91,000 patients in Mexico were used. The data for each patient include demographics, prior medical conditions, SARS-CoV-2 test results, hospitalization, mortality and whether a patient has developed pneumonia or not. Several classification methods were applied, including robust versions of logistic regression, and support vector machines, as well as random forests and gradient boosted decision trees. Results: Interpretable methods (logistic regression and support vector machines) perform just as well as more complex models in terms of accuracy and detection rates, with the additional benefit of elucidating variables on which the predictions are based. Classification accuracies reached 61%, 76%, 83%, and 84% for predicting hospitalization, mortality, need for ICU and need for a ventilator, respectively. The analysis reveals the most important preconditions for making the predictions. For the four models derived, these are: (1) for hospitalization: age, gender, chronic renal insufficiency, diabetes, immunosuppression; (2) for mortality: age, SARS-CoV-2 test status, immunosuppression and pregnancy; (3) for ICU need: development of pneumonia (if available), cardiovascular disease, asthma, and SARS-CoV-2 test status; and (4) for ventilator need: ICU and pneumonia (if available), age, gender, cardiovascular disease, obesity, pregnancy, and SARS-CoV-2 test result.

show abstract

Section: Introductionmentioning

confidence: 99%

Personalized Predictive Models for Symptomatic COVID-19 Patients Using Basic Preconditions:Hospitalizations, Mortality, and the Need for an ICU or Ventilator

Wollenstein-Betech

Cassandras

Paschalidis

2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…We envision that different groups and institutions will have their own local version of the CKG, protecting the sensitive nature of healthcare data, but in a way that still enables cross-platform analyses. New approaches, such as differential privacy and federated learning (Bonawitz et al, 2019;Brisimi et al, 2018), would allow researchers to use the CKG to train models iteratively across institutions without direct access to the sensitive data ( Figure 7C). The CKG could also integrate with existing standardized health data warehouse solutions, such as i2b2 and OMOP (Boussadi and Zapletal, 2017).…”

Section: Discussionmentioning

confidence: 99%

Clinical Knowledge Graph Integrates Proteomics Data into Clinical Decision-Making

Santos

Colaço

Nielsen

et al. 2020

Preprint

View full text Add to dashboard Cite

The promise of precision medicine is to deliver personalized treatment based on the unique physiology of each patient. This concept was fueled by the genomic revolution, but it is now evident that integrating other types of omics data, like proteomics, into the clinical decisionmaking process will be essential to accomplish precision medicine goals. However, quantity and diversity of biomedical data, and the spread of clinically relevant knowledge across myriad biomedical databases and publications makes this exceptionally difficult. To address this, we developed the Clinical Knowledge Graph (CKG), an open source platform currently comprised of more than 16 million nodes and 220 million relationships to represent relevant experimental data, public databases and the literature. The CKG also incorporates the latest statistical and machine learning algorithms, drastically accelerating analysis and interpretation of typical proteomics workflows. We use several biomarker studies to illustrate how the CKG may support, enrich and accelerate clinical decision-making. Graphical Abstract

show abstract

“…Hence, the training model is brought to the data, rather than the data to the model. There are recent applications, sharing electronic health records, without revealing their sensitive content [12]. While federated learning addresses privacy concerns related to data sharing, it introduces threats to the transparency in the learning process, which is another key aspect of automated decision making.…”

Section: Background -Commercial Openness and Datamentioning

confidence: 99%

Open Collaborative Data - using OSS Principles to Share Data in SW Engineering

Runeson

2019

2019 IEEE/ACM 41st International Conference on Software Engineering: New Ideas and Emerging Results (ICSE-NIER)

View full text Add to dashboard Cite

Reliance on data for software systems engineering is increasing, e.g., to train machine learning applications. We foresee increasing costs for data collection and maintenance, leading to the risk of development budgets eaten up by commodity features, thus leaving little resources for differentiation and innovation. We therefore propose Open Collaborative Data (OCD) -a concept analogous to Open Source Software (OSS) -as a means to share data. In contrast to Open Data (OD), which e.g., governmental agencies provide to catalyze innovation, OCD is shared in open collaboration between commercial organizations, similar to OSS. To achieve this, there is a need for technical infrastructure (e.g., tools for version and access control), licence models, and governance models, all of which have to be tailored for data. However, as data may be sensitive for privacy, anonymization and obfuscation of data is also a research challenge. In this paper, we define the concept of Open Collaborative Data, demonstrate it by map data and image recognition examples, and outline a research agenda for OCD in software engineering as a basis for more efficient evolution of software systems.

show abstract

Federated learning of predictive models from federated Electronic Health Records

Cited by 618 publications

References 16 publications

Personalized Predictive Models for Symptomatic COVID-19 Patients Using Basic Preconditions:Hospitalizations, Mortality, and the Need for an ICU or Ventilator

Personalized Predictive Models for Symptomatic COVID-19 Patients Using Basic Preconditions:Hospitalizations, Mortality, and the Need for an ICU or Ventilator

Clinical Knowledge Graph Integrates Proteomics Data into Clinical Decision-Making

Open Collaborative Data - using OSS Principles to Share Data in SW Engineering

Contact Info

Product

Resources

About