Developing Deep Learning Continuous Risk Models for Early Adverse Event Prediction in Electronic Health Records: an AKI Case Study
preprint
OA: closed
CC-BY-4.0
Abstract
Abstract Early detection of patient deterioration is key to unlocking the potential for targeted preventative care and improving patient outcomes. This protocol describes a workflow for developing deep learning continuous risk models for early prediction of future acute adverse events from electronic health records (EHR), taking the prediction of the risk of future acute kidney injury (AKI) as an exemplar. The protocol consists of 34 steps grouped into the following stages: formal problem definition, data processing, model architecture selection, risk calibration and uncertainty, and evaluating model generalisability. For the protocol to be applicable to modelling the future risk of a particular condition, the problem formulation should be clinically and physiologically plausible and there needs to be sufficient associated predictive signal in routinely collected EHR data. Prospective validation is key in evaluating whether retrospective models developed by following the proposed protocol are clinically applicable and useful.
Full text
5,726 characters
· extracted from
preprint-html
· click to expand
:root { --font-inter: 'Inter', sans-serif, Symbol; --font-monospace: monospace; --font-sans-serif: sans-serif; } html, body, div, span, h1, h2, h3, h4, h5, h6, p, ol, ul, li, table, caption, tbody, tfoot, thead, tr, th, td {margin: 0;padding: 0;border: 0;font-size: 100%;font: inherit;vertical-align: baseline;} html {font-size: 14px;line-height: 1.5;} body {font-size: 14px;line-height: 1.5;font-family: var(--font-Inter);} ol, ul {list-style: none;} blockquote, q {quotes: none;} blockquote:before, blockquote:after, q:before, q:after {content: '';content: none;} img {max-width: 100%;} table {border-collapse: collapse;border-spacing: 0;} b {font-weight: bold;} a {color: #2c71d6;text-decoration: none;} p {margin: 10px 0 10px 0;} .no-select {-o-user-select: none;-moz-user-select: none;-khtml-user-select: none;-webkit-user-select: none;user-select: none;} * {box-sizing: border-box;} @keyframes blink {0% {opacity: .1;}20% {opacity: 1;}100% {opacity: .1;}} .main-preloader{position:fixed;display: flex;top:0;bottom:0;left:0;right:0;align-items: center;justify-content: center;z-index:200000000;text-align:center;background:#F2F3F6;} .main-preloader.dark{background:#363636;} .main-preloader.hidden {display: none;} .main-preloader .mp-logo{height:100px;width:100px;} Developing Deep Learning Continuous Risk Models for Early Adverse Event Prediction in Electronic Health Rec... {"@context":"https://schema.org","@id":"https://www.protocols.io/#org","@type":"Organization","logo":{"@type":"ImageObject","height":"1550","url":"https://www.protocols.io/img/protocols-og.jpg","width":"1920"},"name":"protocols.io","url":"https://www.protocols.io/"} {"@context":"https://schema.org","@id":"https://www.protocols.io/#website","@type":"WebSite","name":"protocols.io","publisher":{"@id":"https://www.protocols.io/#org"},"url":"https://www.protocols.io/"} {"@context":"https://schema.org","@id":"https://www.protocols.io/view/developing-deep-learning-continuous-risk-models-fo-eq2ly6y3egx9/v1#webpage","@type":"WebPage","about":{"@id":"https://www.protocols.io/#org"},"description":"Developing Deep Learning Continuous Risk Models for Early Adverse Event Prediction in Electronic Health Records: an AKI Case Study\n. . Read full protocol, st...","isPartOf":{"@id":"https://www.protocols.io/#website"},"name":"Developing Deep Learning Continuous Risk Models for Early Adverse Event Prediction in Electronic Health Rec...","url":"https://www.protocols.io/view/developing-deep-learning-continuous-risk-models-fo-eq2ly6y3egx9/v1"} {"@context":"https://schema.org","@id":"https://www.protocols.io/view/developing-deep-learning-continuous-risk-models-fo-d3i38kgn#article","@type":"Article","author":{"@type":"Person","affiliation":"Dev","name":"Protocols Importer"},"dateModified":"2019-07-31T16:36:20Z","datePublished":"2019-07-31T16:36:20Z","description":"Developing Deep Learning Continuous Risk Models for Early Adverse Event Prediction in Electronic Health Records: an AKI Case Study\n. . Read full protocol, st...","headline":"Developing Deep Learning Continuous Risk Models for Early Adverse Event Prediction in Electronic Health Rec...","image":{"@type":"ImageObject","url":"https://files.protocols.io/webapp/images/q2ckeg/meta/protocolsio.png"},"mainEntityOfPage":{"@id":"https://www.protocols.io/view/developing-deep-learning-continuous-risk-models-fo-d3i38kgn#webpage"},"publisher":{"@id":"https://www.protocols.io/#org"}} var WebappBundlePath = "https:\/\/files.protocols.io\/webapp\/bundles\/production-12becbb\/"; var AMPLITUDE_API_KEY="c250ed1d862cb9c01587e1f779a32c9";var FFMPEG_HOST="https://ffmpeg.protocols.io";var FILES_CDN_URL="https://files.protocols.io";var FILES_S3_BUCKET="protocols-io-files";var FILES_S3_PREFIX="external";var GA_MEASUREMENT_ID="G-33ZGV5CT0B";var GOOGLE_CLIENT_ID="182746271041-rdtjk0b1h9706r0fgpq1o2h55lgbo8un.apps.googleusercontent.com";var HAS_SSR= false ;var HOST_ROLE= 1 ;var HOST_ROLES={"HOST_DEV":2,"HOST_LOCAL":3,"HOST_PROD":1};var IMAGE_BASE_PATH="https://files.protocols.io/webapp/images/q2ckeg/";var NATS_SERVER="wss://ws.protocols.io";var ORCID_CLIENT_ID="APP-7405TP4958GFSZXV";var ORCID_URL="https://orcid.org/oauth/authorize";var RECAPTCHA_SITE_ID="6LeMF6YUAAAAAOYzOqfc9Ml3PdTgIzO1oDdIwCYc";var SITE_URL="https://www.protocols.io";var SPRINGERNATURE_CLIENT_ID="protocolsio";var SPRINGERNATURE_REDIRECT="https://www.protocols.io/api/v1/auth/springernature/authcallback";var SPRINGERNATURE_URL="https://idp.springernature.com";var STRIPE_PUB_KEY="pk_live_51PfMxdRshFFmaQRVGeDsSX7NFFLMkMuqr17SxP6JcJuieglmQYHPgM7cz0XdO91IEVP9k4E4MC7PZB2ATyuuaOy800ivYpRURi";var UIPE="ccW9Kn73Dw3C4V1lovN9dhdEvizyoN5hpBOzwM24c7Y=";var VPC_ALLOW_REGENTS_LINKING= false ;var VPC_DISABLED_INVITE_COLLEAGUES= false ;var VPC_DISABLED_PUBLISHING= false ;var VPC_DISABLED_SIGN_UP= false ;var VPC_DISABLED_WORKSPACE_CREATION= false ;var VPC_EXTERNAL_STORAGES= null ;var VPC_FORWARDED_VIEW_URL_PREFIX="";var VPC_MAIN_CLOUD_URL="https://www.protocols.io";var VPC_REQUIRE_SIGN_IN= false ;var VPC_USE_MAIN_GLOBAL_SEARCH= false ;var WEBAPP_VERSION="production-12becbb";var YJS_WEBSOCKET_URL="wss://www.protocols.io";var __stripeGlobal= null ;var cookie_session_guid="A4F7C6D287C111F183C60A58A9FEAC02";var gosessid="A4F7C5DB87C111F183C60A58A9FEAC02";var is_premium= false ;var loadedAmplitude= false ;var loadedGA= false ;var logged_in= false ;var session_guid="";var show_loader= true ;var user_theme="default"; var is_grecaptcha_loaded = false; if (!Element.prototype.matches) Element.prototype.matches = Element.prototype.msMatchesSelector; if (!Element.prototype.closest) Element.prototype.closest = function (selector) {var el = this; while (el) {if (el.matches(selector)) return el; el = el.parentElement;}};
Text is read by the "Ask this paper" AI Q&A widget below. Extraction quality varies by source — PMC NXML preserves structure cleanly, OA-HTML may include some navigation residue, and OA-PDF can have broken hyphenation. The publisher copy (via DOI) is the canonical version.
My notes (saved in your browser only)
Ask this paper
Answers must be backed by verbatim quotes from this paper's full text. Hallucinated quotes are dropped automatically; if no verbatim passage answers the question, we say so. How this works
Citation neighborhood (no data yet)
We don't have any in-corpus citations linked to this paper yet. The paper's references may be in our DB but unresolved to ``paper_id`` (resolution happens at ingest when the cited DOI matches a row we already have). Run the cross-source citation reconcile pass to retry.
Source provenance
- europepmc
- last seen: 2026-05-19T01:45:01.086888+00:00
- unpaywall
- last seen: 2026-06-04T02:00:05.705006+00:00
License: CC-BY-4.0