CoChat
Sign Up
Cited by 269Machine Learning in HealthcareNursing Diagnosis and DocumentationElectronic Health Records Systems

The DOI matches its Crossref record; 34 of the 35 references it cites were found in a registry or an index.

Combining structured and unstructured data for predictive models: a deep learning approach

Open access
Zhang, Dongdong, Yin, Changchang, Zeng, Jucheng, Yuan, Xiaohui, Zhang, Ping·2020·BMC Medical Informatics and Decision Making
DOI 10.1186/s12911-020-01297-6PMC PMC7596962

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

BACKGROUND: The broad adoption of electronic health records (EHRs) provides great opportunities to conduct health care research and solve various clinical problems in medicine. With recent advances and success, methods based on machine learning and deep learning have become increasingly popular in medical informatics. However, while many research studies utilize temporal structured data on predictive modeling, they typically neglect potentially valuable information in unstructured clinical notes. Integrating heterogeneous data types across EHRs through deep learning techniques may help improve the performance of prediction models. METHODS: In this research, we proposed 2 general-purpose multi-modal neural network architectures to enhance patient representation learning by combining sequential unstructured notes with structured data. The proposed fusion models leverage document embeddings for the representation of long clinical note documents and either convolutional neural network or long short-term memory networks to model the sequential clinical notes and temporal signals, and one-hot encoding for static information representation. The concatenated representation is the final patient representation which is used to make predictions. RESULTS: We evaluate the performance of proposed models on 3 risk prediction tasks (i.e. in-hospital mortality, 30-day hospital readmission, and long length of stay prediction) using derived data from the publicly available Medical Information Mart for Intensive Care III dataset. Our results show that by combining unstructured clinical notes with structured data, the proposed models outperform other models that utilize either unstructured notes or structured data only. CONCLUSIONS: The proposed fusion models learn better patient representation by combining structured and unstructured data. Integrating heterogeneous data types across EHRs helps improve the performance of prediction models and reduce errors.

Overview

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.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1186/s12911-020-01297-6

Cites

34 of 34 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Architecture of CNN-based fusion-CNN. Fusion-CNN uses document embeddings, 2-layer CNN and max-pooling to model sequential clinical notes. Similarly, 2-layer CNN and max-pooling are used to model temporal signals. The final patient representation is the concatenation of the latent representation of sequential clinical notes, temporal signals, and the static information vector. Then the final patient representation is passed to output layers to make predictions
Fig. 1 Architecture of CNN-based fusion-CNN. Fusion-CNN uses document embeddings, 2-layer CNN and max-pooling to model sequential clinical notes. Similarly, 2-layer CNN and max-pooling are used to model temporal signals. The final patient representation is the concatenation of the latent representation of sequential clinical notes, temporal signals, and the static information vector. Then the final patient representation is passed to output layers to make predictionsCC BY · PubMed Central
Fig. 2 Architecture of LSTM-based Fusion-LSTM. Fusion-LSTM uses document embeddings, a BiLSTM layer, and a max-pooling layer to model sequential clinical notes. 2-layer LSTMs are used to model temporal signals. The concatenated patient representation is passed to output layers to make predictions
Fig. 2 Architecture of LSTM-based Fusion-LSTM. Fusion-LSTM uses document embeddings, a BiLSTM layer, and a max-pooling layer to model sequential clinical notes. 2-layer LSTMs are used to model temporal signals. The concatenated patient representation is passed to output layers to make predictionsCC BY · PubMed Central
Fig. 3 Comparison of model running time with different inputs
Fig. 3 Comparison of model running time with different inputsCC BY · PubMed Central

Verification

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.

A few free checks a day · papers already verified open instantly · results are public

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?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 35 references

Research works printed but not linked: 3

  1. Henry J, Pylypchuk Y, Searcy T, Patel V. Adoption of electronic health record systems among US non-federal acute care hospitals: 2008–2015. ONC Data Brief. 2016;35:1–9.
  2. Ghassemi M, Naumann T, Joshi R, Rumshisky A. Topic models for mortality modeling in intensive care units. In: ICML machine learning for clinical data analysis workshop; 2012. p. 1–4.
  3. Hackbarth G, Reischauer R, Miller M. Report to the congress: promoting greater efficiency in medicare. Washington, DC: MedPAC; 2007.

Cited by

Sorted by citations

Items 201–268 of 268

Related

See all

Metadata

Authors
Zhang, D., Yin, C., Zeng, J., Yuan, X., & Zhang, P.
Affiliations
  • Wuhan University of Technology (CN)
  • The Ohio State University (US)
Corresponding author
Zhang, Ping
ORCID
Venue
BMC Medical Informatics and Decision Making
Publisher
Springer Science and Business Media LLC
Published
2020-10-29
Volume
20
Issue
1
Pages
280-280
ISSN
1472-6947 (ISSN-L)
DOI
10.1186/s12911-020-01297-6
PMC
PMC7596962
OpenAlex
W3095791468
Cited by
269
References
35
Citations by year
2026: 41 · 2025: 72 · 2024: 52 · 2023: 52 · 2022: 34 · 2021: 18
FWCI
10.13
Type
journal article
Language
en
Open access
Licence
Locations
  • BMC Medical Informatics and Decision Making (publishedVersion, cc-by)
  • medRxiv (acceptedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
  • Health Sciences › Nursing › Research and Theory › Nursing Diagnosis and Documentation
  • Health Sciences › Health Professions › Health Information Management › Electronic Health Records Systems
Keywords
electronic health records, unstructured clinical notes, multimodal neural networks, patient representation learning, in-hospital mortality prediction, hospital readmission prediction, length of stay prediction, MIMIC-III, clinical risk prediction
MeSH terms
Machine Learning, Deep Learning, Humans, Patient Readmission, Neural Networks, Computer, Electronic Health Records

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