Predicting life expectancy with a long short-term memory recurrent neural network using electronic medical records View Full Text


Ontology type: schema:ScholarlyArticle      Open Access: True


Article Info

DATE

2019-12

AUTHORS

Merijn Beeksma, Suzan Verberne, Antal van den Bosch, Enny Das, Iris Hendrickx, Stef Groenewoud

ABSTRACT

BACKGROUND: Life expectancy is one of the most important factors in end-of-life decision making. Good prognostication for example helps to determine the course of treatment and helps to anticipate the procurement of health care services and facilities, or more broadly: facilitates Advance Care Planning. Advance Care Planning improves the quality of the final phase of life by stimulating doctors to explore the preferences for end-of-life care with their patients, and people close to the patients. Physicians, however, tend to overestimate life expectancy, and miss the window of opportunity to initiate Advance Care Planning. This research tests the potential of using machine learning and natural language processing techniques for predicting life expectancy from electronic medical records. METHODS: We approached the task of predicting life expectancy as a supervised machine learning task. We trained and tested a long short-term memory recurrent neural network on the medical records of deceased patients. We developed the model with a ten-fold cross-validation procedure, and evaluated its performance on a held-out set of test data. We compared the performance of a model which does not use text features (baseline model) to the performance of a model which uses features extracted from the free texts of the medical records (keyword model), and to doctors' performance on a similar task as described in scientific literature. RESULTS: Both doctors and the baseline model were correct in 20% of the cases, taking a margin of 33% around the actual life expectancy as the target. The keyword model, in comparison, attained an accuracy of 29% with its prognoses. While doctors overestimated life expectancy in 63% of the incorrect prognoses, which harms anticipation to appropriate end-of-life care, the keyword model overestimated life expectancy in only 31% of the incorrect prognoses. CONCLUSIONS: Prognostication of life expectancy is difficult for humans. Our research shows that machine learning and natural language processing techniques offer a feasible and promising approach to predicting life expectancy. The research has potential for real-life applications, such as supporting timely recognition of the right moment to start Advance Care Planning. More... »

PAGES

36

References to SciGraph publications

Identifiers

URI

http://scigraph.springernature.com/pub.10.1186/s12911-019-0775-2

DOI

http://dx.doi.org/10.1186/s12911-019-0775-2

DIMENSIONS

https://app.dimensions.ai/details/publication/pub.1112458869

PUBMED

https://www.ncbi.nlm.nih.gov/pubmed/30819172


Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
Incoming Citations Browse incoming citations for this publication using opencitations.net

JSON-LD is the canonical representation for SciGraph data.

TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

[
  {
    "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
    "about": [
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0801", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Artificial Intelligence and Image Processing", 
        "type": "DefinedTerm"
      }, 
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/08", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Information and Computing Sciences", 
        "type": "DefinedTerm"
      }
    ], 
    "author": [
      {
        "affiliation": {
          "alternateName": "Radboud University Nijmegen", 
          "id": "https://www.grid.ac/institutes/grid.5590.9", 
          "name": [
            "Centre for Language Studies, Radboud University, Erasmusplein 1, 6525, Nijmegen, HT, The Netherlands"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Beeksma", 
        "givenName": "Merijn", 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Leiden University", 
          "id": "https://www.grid.ac/institutes/grid.5132.5", 
          "name": [
            "Leiden Institute for Advanced Computer Sciences, Leiden University, Niels Bohrweg 1, 2333, Leiden, CA, The Netherlands"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Verberne", 
        "givenName": "Suzan", 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "name": [
            "KNAW Meertens Institute, Oudezijds Achterburgwal 185, 1012, Amsterdam, DK, The Netherlands"
          ], 
          "type": "Organization"
        }, 
        "familyName": "van den Bosch", 
        "givenName": "Antal", 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Radboud University Nijmegen", 
          "id": "https://www.grid.ac/institutes/grid.5590.9", 
          "name": [
            "Centre for Language Studies, Radboud University, Erasmusplein 1, 6525, Nijmegen, HT, The Netherlands"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Das", 
        "givenName": "Enny", 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Radboud University Nijmegen", 
          "id": "https://www.grid.ac/institutes/grid.5590.9", 
          "name": [
            "Centre for Language Studies, Radboud University, Erasmusplein 1, 6525, Nijmegen, HT, The Netherlands"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Hendrickx", 
        "givenName": "Iris", 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Radboud University Nijmegen Medical Centre", 
          "id": "https://www.grid.ac/institutes/grid.10417.33", 
          "name": [
            "IQ Healthcare, Radboudumc, Mailbox 9101, 6500, Nijmegen, HB, The Netherlands"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Groenewoud", 
        "givenName": "Stef", 
        "type": "Person"
      }
    ], 
    "citation": [
      {
        "id": "https://doi.org/10.1053/j.seminoncol.2010.04.001", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1001054578"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1038/89044", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1002292014", 
          "https://doi.org/10.1038/89044"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1038/89044", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1002292014", 
          "https://doi.org/10.1038/89044"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1177/0269216314526272", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1004165366"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1177/0269216314526272", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1004165366"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1016/j.neunet.2007.12.031", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1006555326"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1136/bmjspcare-2013-000488", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1007256947"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1001/jama.279.21.1709", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1008369019"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1001/jamainternmed.2013.14384", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1009257967"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1038/nrg3208", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1011498059", 
          "https://doi.org/10.1038/nrg3208"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.2215/cjn.00940208", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1012037189"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1016/j.ejso.2011.02.009", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1017204416"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1136/bmjspcare-2013-000527", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1018078814"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1016/s0933-3657(02)00049-0", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1018780126"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1136/amiajnl-2013-001612", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1019429101"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.3399/bjgp12x654597", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1020727877"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1186/1471-2296-14-42", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1022634826", 
          "https://doi.org/10.1186/1471-2296-14-42"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1214/aoms/1177729694", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1026070931"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1016/0002-9343(89)90436-1", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1026362064"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1191/026921698672034203", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1026526686"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1191/026921698672034203", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1026526686"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1136/bmj.320.7233.469", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1030526310"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1145/2110363.2110408", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1031090331"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1371/journal.pone.0161407", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1035773047"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1016/j.jbi.2014.05.013", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1036648877"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.2196/medinform.3172", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1036787255"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1155/2016/3483528", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1037208928"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1162/neco.1997.9.8.1735", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1038140272"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1038/mp.2012.138", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1044963448", 
          "https://doi.org/10.1038/mp.2012.138"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.2196/medinform.3445", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1048638200"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1007/3-540-70659-3_2", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1053125007", 
          "https://doi.org/10.1007/3-540-70659-3_2"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.2196/medinform.3022", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1053519814"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1089/jpm.2010.0018", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1059287312"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1109/4233.945291", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1061171333"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1109/titb.2005.847510", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1061656374"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1109/titb.2005.847510", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1061656374"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.2196/medinform.4192", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1069286814"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.18653/v1/n16-1056", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1079359497"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1016/j.jbi.2017.04.001", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1084740883"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1109/icseng.2011.80", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1094274138"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1109/icdm.2016.0036", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1094287897"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1109/iccoins.2014.6868414", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1094577345"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1109/eeei.2012.6377065", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1095398699"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1145/3007818.3007832", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1096143187"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1145/3007818.3007832", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1096143187"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.18653/v1/p16-1089", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1099113525"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1038/s41746-018-0029-1", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1103817511", 
          "https://doi.org/10.1038/s41746-018-0029-1"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "https://doi.org/10.1016/j.jbi.2018.08.005", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1106270771"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1186/s12911-018-0677-8", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1110542071", 
          "https://doi.org/10.1186/s12911-018-0677-8"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1186/s12911-018-0677-8", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1110542071", 
          "https://doi.org/10.1186/s12911-018-0677-8"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1186/s12911-018-0677-8", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1110542071", 
          "https://doi.org/10.1186/s12911-018-0677-8"
        ], 
        "type": "CreativeWork"
      }, 
      {
        "id": "sg:pub.10.1186/s12911-018-0677-8", 
        "sameAs": [
          "https://app.dimensions.ai/details/publication/pub.1110542071", 
          "https://doi.org/10.1186/s12911-018-0677-8"
        ], 
        "type": "CreativeWork"
      }
    ], 
    "datePublished": "2019-12", 
    "datePublishedReg": "2019-12-01", 
    "description": "BACKGROUND: Life expectancy is one of the most important factors in end-of-life decision making. Good prognostication for example helps to determine the course of treatment and helps to anticipate the procurement of health care services and facilities, or more broadly: facilitates Advance Care Planning. Advance Care Planning improves the quality of the final phase of life by stimulating doctors to explore the preferences for end-of-life care with their patients, and people close to the patients. Physicians, however, tend to overestimate life expectancy, and miss the window of opportunity to initiate Advance Care Planning. This research tests the potential of using machine learning and natural language processing techniques for predicting life expectancy from electronic medical records.\nMETHODS: We approached the task of predicting life expectancy as a supervised machine learning task. We trained and tested a long short-term memory recurrent neural network on the medical records of deceased patients. We developed the model with a ten-fold cross-validation procedure, and evaluated its performance on a held-out set of test data. We compared the performance of a model which does not use text features (baseline model) to the performance of a model which uses features extracted from the free texts of the medical records (keyword model), and to doctors' performance on a similar task as described in scientific literature.\nRESULTS: Both doctors and the baseline model were correct in 20% of the cases, taking a margin of 33% around the actual life expectancy as the target. The keyword model, in comparison, attained an accuracy of 29% with its prognoses. While doctors overestimated life expectancy in 63% of the incorrect prognoses, which harms anticipation to appropriate end-of-life care, the keyword model overestimated life expectancy in only 31% of the incorrect prognoses.\nCONCLUSIONS: Prognostication of life expectancy is difficult for humans. Our research shows that machine learning and natural language processing techniques offer a feasible and promising approach to predicting life expectancy. The research has potential for real-life applications, such as supporting timely recognition of the right moment to start Advance Care Planning.", 
    "genre": "research_article", 
    "id": "sg:pub.10.1186/s12911-019-0775-2", 
    "inLanguage": [
      "en"
    ], 
    "isAccessibleForFree": true, 
    "isPartOf": [
      {
        "id": "sg:journal.1028432", 
        "issn": [
          "1472-6947"
        ], 
        "name": "BMC Medical Informatics and Decision Making", 
        "type": "Periodical"
      }, 
      {
        "issueNumber": "1", 
        "type": "PublicationIssue"
      }, 
      {
        "type": "PublicationVolume", 
        "volumeNumber": "19"
      }
    ], 
    "name": "Predicting life expectancy with a long short-term memory recurrent neural network using electronic medical records", 
    "pagination": "36", 
    "productId": [
      {
        "name": "readcube_id", 
        "type": "PropertyValue", 
        "value": [
          "39ef6a619fedad168f849612194e6f6ac1a5d1670ab4be9cf99a79d3ee2f5426"
        ]
      }, 
      {
        "name": "pubmed_id", 
        "type": "PropertyValue", 
        "value": [
          "30819172"
        ]
      }, 
      {
        "name": "nlm_unique_id", 
        "type": "PropertyValue", 
        "value": [
          "101088682"
        ]
      }, 
      {
        "name": "doi", 
        "type": "PropertyValue", 
        "value": [
          "10.1186/s12911-019-0775-2"
        ]
      }, 
      {
        "name": "dimensions_id", 
        "type": "PropertyValue", 
        "value": [
          "pub.1112458869"
        ]
      }
    ], 
    "sameAs": [
      "https://doi.org/10.1186/s12911-019-0775-2", 
      "https://app.dimensions.ai/details/publication/pub.1112458869"
    ], 
    "sdDataset": "articles", 
    "sdDatePublished": "2019-04-11T13:51", 
    "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
    "sdPublisher": {
      "name": "Springer Nature - SN SciGraph project", 
      "type": "Organization"
    }, 
    "sdSource": "s3://com-uberresearch-data-dimensions-target-20181106-alternative/cleanup/v134/2549eaecd7973599484d7c17b260dba0a4ecb94b/merge/v9/a6c9fde33151104705d4d7ff012ea9563521a3ce/jats-lookup/v90/0000000371_0000000371/records_130797_00000006.jsonl", 
    "type": "ScholarlyArticle", 
    "url": "https://link.springer.com/10.1186%2Fs12911-019-0775-2"
  }
]
 

Download the RDF metadata as:  json-ld nt turtle xml License info

HOW TO GET THIS DATA PROGRAMMATICALLY:

JSON-LD is a popular format for linked data which is fully compatible with JSON.

curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1186/s12911-019-0775-2'

N-Triples is a line-based linked data format ideal for batch operations.

curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1186/s12911-019-0775-2'

Turtle is a human-readable linked data format.

curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1186/s12911-019-0775-2'

RDF/XML is a standard XML format for linked data.

curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1186/s12911-019-0775-2'


 

This table displays all metadata directly associated to this object as RDF triples.

244 TRIPLES      21 PREDICATES      73 URIs      21 LITERALS      9 BLANK NODES

Subject Predicate Object
1 sg:pub.10.1186/s12911-019-0775-2 schema:about anzsrc-for:08
2 anzsrc-for:0801
3 schema:author Nb1ee0ac2ac5841ffa063726e4586a23f
4 schema:citation sg:pub.10.1007/3-540-70659-3_2
5 sg:pub.10.1038/89044
6 sg:pub.10.1038/mp.2012.138
7 sg:pub.10.1038/nrg3208
8 sg:pub.10.1038/s41746-018-0029-1
9 sg:pub.10.1186/1471-2296-14-42
10 sg:pub.10.1186/s12911-018-0677-8
11 https://doi.org/10.1001/jama.279.21.1709
12 https://doi.org/10.1001/jamainternmed.2013.14384
13 https://doi.org/10.1016/0002-9343(89)90436-1
14 https://doi.org/10.1016/j.ejso.2011.02.009
15 https://doi.org/10.1016/j.jbi.2014.05.013
16 https://doi.org/10.1016/j.jbi.2017.04.001
17 https://doi.org/10.1016/j.jbi.2018.08.005
18 https://doi.org/10.1016/j.neunet.2007.12.031
19 https://doi.org/10.1016/s0933-3657(02)00049-0
20 https://doi.org/10.1053/j.seminoncol.2010.04.001
21 https://doi.org/10.1089/jpm.2010.0018
22 https://doi.org/10.1109/4233.945291
23 https://doi.org/10.1109/eeei.2012.6377065
24 https://doi.org/10.1109/iccoins.2014.6868414
25 https://doi.org/10.1109/icdm.2016.0036
26 https://doi.org/10.1109/icseng.2011.80
27 https://doi.org/10.1109/titb.2005.847510
28 https://doi.org/10.1136/amiajnl-2013-001612
29 https://doi.org/10.1136/bmj.320.7233.469
30 https://doi.org/10.1136/bmjspcare-2013-000488
31 https://doi.org/10.1136/bmjspcare-2013-000527
32 https://doi.org/10.1145/2110363.2110408
33 https://doi.org/10.1145/3007818.3007832
34 https://doi.org/10.1155/2016/3483528
35 https://doi.org/10.1162/neco.1997.9.8.1735
36 https://doi.org/10.1177/0269216314526272
37 https://doi.org/10.1191/026921698672034203
38 https://doi.org/10.1214/aoms/1177729694
39 https://doi.org/10.1371/journal.pone.0161407
40 https://doi.org/10.18653/v1/n16-1056
41 https://doi.org/10.18653/v1/p16-1089
42 https://doi.org/10.2196/medinform.3022
43 https://doi.org/10.2196/medinform.3172
44 https://doi.org/10.2196/medinform.3445
45 https://doi.org/10.2196/medinform.4192
46 https://doi.org/10.2215/cjn.00940208
47 https://doi.org/10.3399/bjgp12x654597
48 schema:datePublished 2019-12
49 schema:datePublishedReg 2019-12-01
50 schema:description BACKGROUND: Life expectancy is one of the most important factors in end-of-life decision making. Good prognostication for example helps to determine the course of treatment and helps to anticipate the procurement of health care services and facilities, or more broadly: facilitates Advance Care Planning. Advance Care Planning improves the quality of the final phase of life by stimulating doctors to explore the preferences for end-of-life care with their patients, and people close to the patients. Physicians, however, tend to overestimate life expectancy, and miss the window of opportunity to initiate Advance Care Planning. This research tests the potential of using machine learning and natural language processing techniques for predicting life expectancy from electronic medical records. METHODS: We approached the task of predicting life expectancy as a supervised machine learning task. We trained and tested a long short-term memory recurrent neural network on the medical records of deceased patients. We developed the model with a ten-fold cross-validation procedure, and evaluated its performance on a held-out set of test data. We compared the performance of a model which does not use text features (baseline model) to the performance of a model which uses features extracted from the free texts of the medical records (keyword model), and to doctors' performance on a similar task as described in scientific literature. RESULTS: Both doctors and the baseline model were correct in 20% of the cases, taking a margin of 33% around the actual life expectancy as the target. The keyword model, in comparison, attained an accuracy of 29% with its prognoses. While doctors overestimated life expectancy in 63% of the incorrect prognoses, which harms anticipation to appropriate end-of-life care, the keyword model overestimated life expectancy in only 31% of the incorrect prognoses. CONCLUSIONS: Prognostication of life expectancy is difficult for humans. Our research shows that machine learning and natural language processing techniques offer a feasible and promising approach to predicting life expectancy. The research has potential for real-life applications, such as supporting timely recognition of the right moment to start Advance Care Planning.
51 schema:genre research_article
52 schema:inLanguage en
53 schema:isAccessibleForFree true
54 schema:isPartOf N12049ecbe3444adc838d289177fe9bed
55 N9cf3980e137e4a6f9d3f59f0581e908b
56 sg:journal.1028432
57 schema:name Predicting life expectancy with a long short-term memory recurrent neural network using electronic medical records
58 schema:pagination 36
59 schema:productId N185ac48186d94daf9927e830697c75fb
60 N1b764e64cbee40eaae1435b41278b125
61 N7160421c651b44b9a6d686509f30febc
62 N870a8486da7f45bfbc97173a8250b5c1
63 Nf9b86d5c0c4b4853b663b5ba3c4936f7
64 schema:sameAs https://app.dimensions.ai/details/publication/pub.1112458869
65 https://doi.org/10.1186/s12911-019-0775-2
66 schema:sdDatePublished 2019-04-11T13:51
67 schema:sdLicense https://scigraph.springernature.com/explorer/license/
68 schema:sdPublisher N78c24d4282854606a35937764fc5a5f7
69 schema:url https://link.springer.com/10.1186%2Fs12911-019-0775-2
70 sgo:license sg:explorer/license/
71 sgo:sdDataset articles
72 rdf:type schema:ScholarlyArticle
73 N0e6bf22a75054fd1a24d8f3d26d63fd8 rdf:first N665de51d7b05431283ce4695181345d8
74 rdf:rest Na17a17413eb14222a0ae2b6ceca1dffd
75 N0fa97a7463e14cf2ab05b9f25e368baf schema:affiliation Neae8d63160f14e1eb0ec496fa102141e
76 schema:familyName van den Bosch
77 schema:givenName Antal
78 rdf:type schema:Person
79 N10f1e216b88b485b98305fee8d62714c rdf:first N5eb3d1292d6b43c8bba3d17e73b9e28e
80 rdf:rest N4320c326833f42219c1f32f9dd754b7f
81 N12049ecbe3444adc838d289177fe9bed schema:issueNumber 1
82 rdf:type schema:PublicationIssue
83 N185ac48186d94daf9927e830697c75fb schema:name nlm_unique_id
84 schema:value 101088682
85 rdf:type schema:PropertyValue
86 N1b764e64cbee40eaae1435b41278b125 schema:name readcube_id
87 schema:value 39ef6a619fedad168f849612194e6f6ac1a5d1670ab4be9cf99a79d3ee2f5426
88 rdf:type schema:PropertyValue
89 N4320c326833f42219c1f32f9dd754b7f rdf:first N0fa97a7463e14cf2ab05b9f25e368baf
90 rdf:rest N9447b343e11f4c558d6ae842840cf8df
91 N5eb3d1292d6b43c8bba3d17e73b9e28e schema:affiliation https://www.grid.ac/institutes/grid.5132.5
92 schema:familyName Verberne
93 schema:givenName Suzan
94 rdf:type schema:Person
95 N62bade29c09f44d2a2f97ff0014e6c97 schema:affiliation https://www.grid.ac/institutes/grid.5590.9
96 schema:familyName Beeksma
97 schema:givenName Merijn
98 rdf:type schema:Person
99 N640a75930ad640e5ac6e7fd2cff32d14 schema:affiliation https://www.grid.ac/institutes/grid.5590.9
100 schema:familyName Das
101 schema:givenName Enny
102 rdf:type schema:Person
103 N665de51d7b05431283ce4695181345d8 schema:affiliation https://www.grid.ac/institutes/grid.5590.9
104 schema:familyName Hendrickx
105 schema:givenName Iris
106 rdf:type schema:Person
107 N7160421c651b44b9a6d686509f30febc schema:name dimensions_id
108 schema:value pub.1112458869
109 rdf:type schema:PropertyValue
110 N78c24d4282854606a35937764fc5a5f7 schema:name Springer Nature - SN SciGraph project
111 rdf:type schema:Organization
112 N870a8486da7f45bfbc97173a8250b5c1 schema:name doi
113 schema:value 10.1186/s12911-019-0775-2
114 rdf:type schema:PropertyValue
115 N87db27bb8bc24484b1f088cb4742f7e4 schema:affiliation https://www.grid.ac/institutes/grid.10417.33
116 schema:familyName Groenewoud
117 schema:givenName Stef
118 rdf:type schema:Person
119 N9447b343e11f4c558d6ae842840cf8df rdf:first N640a75930ad640e5ac6e7fd2cff32d14
120 rdf:rest N0e6bf22a75054fd1a24d8f3d26d63fd8
121 N9cf3980e137e4a6f9d3f59f0581e908b schema:volumeNumber 19
122 rdf:type schema:PublicationVolume
123 Na17a17413eb14222a0ae2b6ceca1dffd rdf:first N87db27bb8bc24484b1f088cb4742f7e4
124 rdf:rest rdf:nil
125 Nb1ee0ac2ac5841ffa063726e4586a23f rdf:first N62bade29c09f44d2a2f97ff0014e6c97
126 rdf:rest N10f1e216b88b485b98305fee8d62714c
127 Neae8d63160f14e1eb0ec496fa102141e schema:name KNAW Meertens Institute, Oudezijds Achterburgwal 185, 1012, Amsterdam, DK, The Netherlands
128 rdf:type schema:Organization
129 Nf9b86d5c0c4b4853b663b5ba3c4936f7 schema:name pubmed_id
130 schema:value 30819172
131 rdf:type schema:PropertyValue
132 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
133 schema:name Information and Computing Sciences
134 rdf:type schema:DefinedTerm
135 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
136 schema:name Artificial Intelligence and Image Processing
137 rdf:type schema:DefinedTerm
138 sg:journal.1028432 schema:issn 1472-6947
139 schema:name BMC Medical Informatics and Decision Making
140 rdf:type schema:Periodical
141 sg:pub.10.1007/3-540-70659-3_2 schema:sameAs https://app.dimensions.ai/details/publication/pub.1053125007
142 https://doi.org/10.1007/3-540-70659-3_2
143 rdf:type schema:CreativeWork
144 sg:pub.10.1038/89044 schema:sameAs https://app.dimensions.ai/details/publication/pub.1002292014
145 https://doi.org/10.1038/89044
146 rdf:type schema:CreativeWork
147 sg:pub.10.1038/mp.2012.138 schema:sameAs https://app.dimensions.ai/details/publication/pub.1044963448
148 https://doi.org/10.1038/mp.2012.138
149 rdf:type schema:CreativeWork
150 sg:pub.10.1038/nrg3208 schema:sameAs https://app.dimensions.ai/details/publication/pub.1011498059
151 https://doi.org/10.1038/nrg3208
152 rdf:type schema:CreativeWork
153 sg:pub.10.1038/s41746-018-0029-1 schema:sameAs https://app.dimensions.ai/details/publication/pub.1103817511
154 https://doi.org/10.1038/s41746-018-0029-1
155 rdf:type schema:CreativeWork
156 sg:pub.10.1186/1471-2296-14-42 schema:sameAs https://app.dimensions.ai/details/publication/pub.1022634826
157 https://doi.org/10.1186/1471-2296-14-42
158 rdf:type schema:CreativeWork
159 sg:pub.10.1186/s12911-018-0677-8 schema:sameAs https://app.dimensions.ai/details/publication/pub.1110542071
160 https://doi.org/10.1186/s12911-018-0677-8
161 rdf:type schema:CreativeWork
162 https://doi.org/10.1001/jama.279.21.1709 schema:sameAs https://app.dimensions.ai/details/publication/pub.1008369019
163 rdf:type schema:CreativeWork
164 https://doi.org/10.1001/jamainternmed.2013.14384 schema:sameAs https://app.dimensions.ai/details/publication/pub.1009257967
165 rdf:type schema:CreativeWork
166 https://doi.org/10.1016/0002-9343(89)90436-1 schema:sameAs https://app.dimensions.ai/details/publication/pub.1026362064
167 rdf:type schema:CreativeWork
168 https://doi.org/10.1016/j.ejso.2011.02.009 schema:sameAs https://app.dimensions.ai/details/publication/pub.1017204416
169 rdf:type schema:CreativeWork
170 https://doi.org/10.1016/j.jbi.2014.05.013 schema:sameAs https://app.dimensions.ai/details/publication/pub.1036648877
171 rdf:type schema:CreativeWork
172 https://doi.org/10.1016/j.jbi.2017.04.001 schema:sameAs https://app.dimensions.ai/details/publication/pub.1084740883
173 rdf:type schema:CreativeWork
174 https://doi.org/10.1016/j.jbi.2018.08.005 schema:sameAs https://app.dimensions.ai/details/publication/pub.1106270771
175 rdf:type schema:CreativeWork
176 https://doi.org/10.1016/j.neunet.2007.12.031 schema:sameAs https://app.dimensions.ai/details/publication/pub.1006555326
177 rdf:type schema:CreativeWork
178 https://doi.org/10.1016/s0933-3657(02)00049-0 schema:sameAs https://app.dimensions.ai/details/publication/pub.1018780126
179 rdf:type schema:CreativeWork
180 https://doi.org/10.1053/j.seminoncol.2010.04.001 schema:sameAs https://app.dimensions.ai/details/publication/pub.1001054578
181 rdf:type schema:CreativeWork
182 https://doi.org/10.1089/jpm.2010.0018 schema:sameAs https://app.dimensions.ai/details/publication/pub.1059287312
183 rdf:type schema:CreativeWork
184 https://doi.org/10.1109/4233.945291 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061171333
185 rdf:type schema:CreativeWork
186 https://doi.org/10.1109/eeei.2012.6377065 schema:sameAs https://app.dimensions.ai/details/publication/pub.1095398699
187 rdf:type schema:CreativeWork
188 https://doi.org/10.1109/iccoins.2014.6868414 schema:sameAs https://app.dimensions.ai/details/publication/pub.1094577345
189 rdf:type schema:CreativeWork
190 https://doi.org/10.1109/icdm.2016.0036 schema:sameAs https://app.dimensions.ai/details/publication/pub.1094287897
191 rdf:type schema:CreativeWork
192 https://doi.org/10.1109/icseng.2011.80 schema:sameAs https://app.dimensions.ai/details/publication/pub.1094274138
193 rdf:type schema:CreativeWork
194 https://doi.org/10.1109/titb.2005.847510 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061656374
195 rdf:type schema:CreativeWork
196 https://doi.org/10.1136/amiajnl-2013-001612 schema:sameAs https://app.dimensions.ai/details/publication/pub.1019429101
197 rdf:type schema:CreativeWork
198 https://doi.org/10.1136/bmj.320.7233.469 schema:sameAs https://app.dimensions.ai/details/publication/pub.1030526310
199 rdf:type schema:CreativeWork
200 https://doi.org/10.1136/bmjspcare-2013-000488 schema:sameAs https://app.dimensions.ai/details/publication/pub.1007256947
201 rdf:type schema:CreativeWork
202 https://doi.org/10.1136/bmjspcare-2013-000527 schema:sameAs https://app.dimensions.ai/details/publication/pub.1018078814
203 rdf:type schema:CreativeWork
204 https://doi.org/10.1145/2110363.2110408 schema:sameAs https://app.dimensions.ai/details/publication/pub.1031090331
205 rdf:type schema:CreativeWork
206 https://doi.org/10.1145/3007818.3007832 schema:sameAs https://app.dimensions.ai/details/publication/pub.1096143187
207 rdf:type schema:CreativeWork
208 https://doi.org/10.1155/2016/3483528 schema:sameAs https://app.dimensions.ai/details/publication/pub.1037208928
209 rdf:type schema:CreativeWork
210 https://doi.org/10.1162/neco.1997.9.8.1735 schema:sameAs https://app.dimensions.ai/details/publication/pub.1038140272
211 rdf:type schema:CreativeWork
212 https://doi.org/10.1177/0269216314526272 schema:sameAs https://app.dimensions.ai/details/publication/pub.1004165366
213 rdf:type schema:CreativeWork
214 https://doi.org/10.1191/026921698672034203 schema:sameAs https://app.dimensions.ai/details/publication/pub.1026526686
215 rdf:type schema:CreativeWork
216 https://doi.org/10.1214/aoms/1177729694 schema:sameAs https://app.dimensions.ai/details/publication/pub.1026070931
217 rdf:type schema:CreativeWork
218 https://doi.org/10.1371/journal.pone.0161407 schema:sameAs https://app.dimensions.ai/details/publication/pub.1035773047
219 rdf:type schema:CreativeWork
220 https://doi.org/10.18653/v1/n16-1056 schema:sameAs https://app.dimensions.ai/details/publication/pub.1079359497
221 rdf:type schema:CreativeWork
222 https://doi.org/10.18653/v1/p16-1089 schema:sameAs https://app.dimensions.ai/details/publication/pub.1099113525
223 rdf:type schema:CreativeWork
224 https://doi.org/10.2196/medinform.3022 schema:sameAs https://app.dimensions.ai/details/publication/pub.1053519814
225 rdf:type schema:CreativeWork
226 https://doi.org/10.2196/medinform.3172 schema:sameAs https://app.dimensions.ai/details/publication/pub.1036787255
227 rdf:type schema:CreativeWork
228 https://doi.org/10.2196/medinform.3445 schema:sameAs https://app.dimensions.ai/details/publication/pub.1048638200
229 rdf:type schema:CreativeWork
230 https://doi.org/10.2196/medinform.4192 schema:sameAs https://app.dimensions.ai/details/publication/pub.1069286814
231 rdf:type schema:CreativeWork
232 https://doi.org/10.2215/cjn.00940208 schema:sameAs https://app.dimensions.ai/details/publication/pub.1012037189
233 rdf:type schema:CreativeWork
234 https://doi.org/10.3399/bjgp12x654597 schema:sameAs https://app.dimensions.ai/details/publication/pub.1020727877
235 rdf:type schema:CreativeWork
236 https://www.grid.ac/institutes/grid.10417.33 schema:alternateName Radboud University Nijmegen Medical Centre
237 schema:name IQ Healthcare, Radboudumc, Mailbox 9101, 6500, Nijmegen, HB, The Netherlands
238 rdf:type schema:Organization
239 https://www.grid.ac/institutes/grid.5132.5 schema:alternateName Leiden University
240 schema:name Leiden Institute for Advanced Computer Sciences, Leiden University, Niels Bohrweg 1, 2333, Leiden, CA, The Netherlands
241 rdf:type schema:Organization
242 https://www.grid.ac/institutes/grid.5590.9 schema:alternateName Radboud University Nijmegen
243 schema:name Centre for Language Studies, Radboud University, Erasmusplein 1, 6525, Nijmegen, HT, The Netherlands
244 rdf:type schema:Organization
 




Preview window. Press ESC to close (or click here)


...