<?xml version="1.0" encoding="UTF-8"?><!DOCTYPE article PUBLIC "-//NLM//DTD Journal Publishing DTD v2.0 20040830//EN" "journalpublishing.dtd"><article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" dtd-version="2.0" xml:lang="en" article-type="review-article"><front><journal-meta><journal-id journal-id-type="nlm-ta">Interact J Med Res</journal-id><journal-id journal-id-type="publisher-id">i-jmr</journal-id><journal-id journal-id-type="index">3</journal-id><journal-title>Interactive Journal of Medical Research</journal-title><abbrev-journal-title>Interact J Med Res</abbrev-journal-title><issn pub-type="epub">1929-073X</issn><publisher><publisher-name>JMIR Publications</publisher-name><publisher-loc>Toronto, Canada</publisher-loc></publisher></journal-meta><article-meta><article-id pub-id-type="publisher-id">v15i1e96779</article-id><article-id pub-id-type="doi">10.2196/96779</article-id><article-categories><subj-group subj-group-type="heading"><subject>Review</subject></subj-group></article-categories><title-group><article-title>Application of Machine Learning Algorithms to Health Systems Planning and Public Policy Using Real-World Data: Scoping Review</article-title></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><name name-style="western"><surname>Rocha</surname><given-names>Deusa Maria Arruda</given-names></name><degrees>MPH</degrees><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="aff" rid="aff2">2</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Barbosa</surname><given-names>Gentil Veloso</given-names></name><degrees>PhD</degrees><xref ref-type="aff" rid="aff2">2</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>De Souza Bermejo</surname><given-names>Paulo Henrique</given-names></name><degrees>PhD</degrees><xref ref-type="aff" rid="aff3">3</xref></contrib></contrib-group><aff id="aff1"><institution>Department of Regulation, Evaluation, and Control (DRAC), Secretariat of Specialized Healthcare (SAES), Minist&#x00E9;rio da Sa&#x00FA;de</institution><addr-line>Esplanada dos Minist&#x00E9;rios, Bloco G, Edif&#x00ED;cio Anexo</addr-line><addr-line>Bras&#x00ED;lia</addr-line><addr-line>Federal District</addr-line><country>Brazil</country></aff><aff id="aff2"><institution>Professional Postgraduate Program in Governance and Digital Transformation, Federal University of Tocantins</institution><addr-line>Palmas</addr-line><addr-line>Tocantins</addr-line><country>Brazil</country></aff><aff id="aff3"><institution>Sinerji Institute for Technological and Business Research and Development</institution><addr-line>Bras&#x00ED;lia</addr-line><addr-line>Federal District</addr-line><country>Brazil</country></aff><contrib-group><contrib contrib-type="editor"><name name-style="western"><surname>Balcarras</surname><given-names>Matthew</given-names></name></contrib></contrib-group><contrib-group><contrib contrib-type="reviewer"><name name-style="western"><surname>Herrera</surname><given-names>Claire Nierva</given-names></name></contrib><contrib contrib-type="reviewer"><name name-style="western"><surname>Causio</surname><given-names>Francesco Andrea</given-names></name></contrib><contrib contrib-type="reviewer"><name name-style="western"><surname>Boucher</surname><given-names>Jean-Christophe</given-names></name></contrib><contrib contrib-type="reviewer"><name name-style="western"><surname>Odeniya</surname><given-names>Joshua</given-names></name></contrib></contrib-group><author-notes><corresp>Correspondence to Deusa Maria Arruda Rocha, MPH, Department of Regulation, Evaluation, and Control (DRAC), Secretariat of Specialized Healthcare (SAES), Minist&#x00E9;rio da Sa&#x00FA;de, Esplanada dos Minist&#x00E9;rios, Bloco G, Edif&#x00ED;cio Anexo, Bras&#x00ED;lia, Federal District, 70058-900, Brazil, +55 61 98295-2606; <email>deusa.rocha@saude.gov.br</email></corresp></author-notes><pub-date pub-type="collection"><year>2026</year></pub-date><pub-date pub-type="epub"><day>16</day><month>11</month><year>2026</year></pub-date><volume>15</volume><elocation-id>e96779</elocation-id><history><date date-type="received"><day>20</day><month>04</month><year>2026</year></date><date date-type="rev-recd"><day>01</day><month>09</month><year>2026</year></date><date date-type="accepted"><day>14</day><month>09</month><year>2026</year></date></history><copyright-statement>&#x00A9; Deusa Maria Arruda Rocha, Gentil Veloso Barbosa, Paulo Henrique De Souza Bermejo. Originally published in the Interactive Journal of Medical Research (<ext-link ext-link-type="uri" xlink:href="https://www.i-jmr.org/">https://www.i-jmr.org/</ext-link>), 16.11.2026. </copyright-statement><copyright-year>2026</copyright-year><license license-type="open-access" xlink:href="https://creativecommons.org/licenses/by/4.0/"><p>This is an open-access article distributed under the terms of the Creative Commons Attribution License (<ext-link ext-link-type="uri" xlink:href="https://creativecommons.org/licenses/by/4.0/">https://creativecommons.org/licenses/by/4.0/</ext-link>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Interactive Journal of Medical Research, is properly cited. The complete bibliographic information, a link to the original publication on <ext-link ext-link-type="uri" xlink:href="https://www.i-jmr.org/">https://www.i-jmr.org/</ext-link>, as well as this copyright and license information must be included.</p></license><self-uri xlink:type="simple" xlink:href="https://www.i-jmr.org/2026/1/e96779"/><abstract><sec><title>Background</title><p>The rapid advancement of information technology and machine learning (ML) has garnered significant interest from health care managers and policymakers. However, incorporating reliable and transparent predictions into health systems remains challenging, necessitating an investigation into gaps in digital health interventions within real-world settings characterized by resource constraints and reliability issues.</p></sec><sec><title>Objective</title><p>This study aimed to map and synthesize the available evidence on the application of ML algorithms to real-world data in health systems and identify their operational targets and clinical implications for health planning and public policy formulation.</p></sec><sec sec-type="methods"><title>Methods</title><p>Following the Joanna Briggs Institute (JBI) methodology for scoping reviews and PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines, we searched the PubMed, Embase, Scopus, and LILACS databases for empirical studies published between 2016 and 2025 that evaluated the application of ML to health management, surgical queues, and public policy using real-world data. The study protocol was retrospectively registered with the OSF.</p></sec><sec sec-type="results"><title>Results</title><p>Of the studies identified through the database searches (2016-2025), only 6 empirical studies met all inclusion criteria, revealing that the real-world deployment of ML in health management and public policy remains at an early stage. The included models addressed micro-operational tasks, such as surgical blood transfusion forecasting, early warning system mortality prediction, postsurgical discharge planning, laboratory error detection, patient-reported outcome measure (PROM) prediction, and emergency care classification. Although high discriminatory performance was reported (area under the receiver operating characteristic curve [AUROC] 0.76-0.99), the studies were limited by single-center retrospective designs, a lack of independent external validation, and inadequate reporting of model calibration.</p></sec><sec sec-type="conclusions"><title>Conclusions</title><p>Current evidence on ML for macrolevel health planning and public policy formulation is limited and micro-operationally focused. Translating these algorithms into public health governance (eg, waiting list regulation in universal health systems such as the Sistema &#x00DA;nico de Sa&#x00FA;de [SUS]) requires moving beyond &#x201C;black-box&#x201D; predictions toward explainable AI (XAI), natural language processing (NLP) to analyze unstructured data, transparent model calibration, and rigorous evaluations of health equity.</p></sec></abstract><kwd-group><kwd>artificial intelligence</kwd><kwd>AI</kwd><kwd>machine learning</kwd><kwd>digital health</kwd><kwd>health management</kwd><kwd>waiting lists</kwd><kwd>elective surgical procedures</kwd><kwd>public health systems</kwd><kwd>Brazil&#x2019;s unified health system</kwd><kwd>Sistema &#x00DA;nico de Sa&#x00FA;de</kwd><kwd>SUS</kwd></kwd-group></article-meta></front><body><sec id="s1"><title>Background</title><p>The emergence of AI as a tool for enhancing health care offers unprecedented opportunities to optimize clinical outcomes, reduce costs, and positively impact population health [<xref ref-type="bibr" rid="ref1">1</xref>]. Among its subfields, machine learning (ML) stands out for its use of algorithmic structures and statistical modeling to identify patterns in real-world data (RWD), encompassing clinical, behavioral, and demographic variables. This facilitates the establishment of functional relationships that significantly improve predictive accuracy [<xref ref-type="bibr" rid="ref2">2</xref>-<xref ref-type="bibr" rid="ref4">4</xref>]. AI-driven interventions provide robust decision support for medical professionals and create new opportunities for therapeutic monitoring, particularly in specialized care [<xref ref-type="bibr" rid="ref5">5</xref>,<xref ref-type="bibr" rid="ref6">6</xref>].</p><p>This technological evolution is part of a data-driven digital revolution, marking the transition from Industry 4.0 to Industry 5.0. In the health care sector, this phase&#x2014;known as &#x201C;Health 5.0&#x201D; (H5.0)&#x2014;explores technological potential to reshape service delivery, promoting personalized medicine and, most importantly, patient-centered care [<xref ref-type="bibr" rid="ref2">2</xref>,<xref ref-type="bibr" rid="ref7">7</xref>-<xref ref-type="bibr" rid="ref10">10</xref>]. Interdisciplinary interest in these technologies grows alongside the global need to optimize health care spending, which accounted for approximately 11% of global gross domestic product (GDP) in 2023. This requires IT infrastructures and algorithms capable of generating clinically reliable and actionable predictions [<xref ref-type="bibr" rid="ref11">11</xref>,<xref ref-type="bibr" rid="ref12">12</xref>].</p><p>However, despite the exponential growth of ML research, a gap persists in both the scientific literature and government agendas regarding the effective integration of these techniques into service delivery and public policy [<xref ref-type="bibr" rid="ref2">2</xref>,<xref ref-type="bibr" rid="ref6">6</xref>]. This lack of standardized guidelines leaves many nations in the early stages of technological adoption [<xref ref-type="bibr" rid="ref13">13</xref>].</p><p>A critical example is the management of waiting lists for elective surgeries, which represents a growing challenge because of population aging and the increased demand for complex procedures [<xref ref-type="bibr" rid="ref3">3</xref>]. In this scenario, it is imperative for managers to adopt ML models to streamline workflows, ensure patient safety, and expedite care delivery. As noted by Lungu et al [<xref ref-type="bibr" rid="ref14">14</xref>], waiting times are a central public concern and a priority for health policymakers. Yet, planning effective interventions to address waiting times remains challenging because of demographic shifts and increasing public confidence in surgical treatments.</p><p>Consequently, there is an urgent need for research addressing the intervention gap regarding the application of these tools in real-world settings. A critical limitation of the current literature is the predominance of idealized, laboratory-based simulations that fail to reflect the operational friction, missing data, and resource constraints inherent to day-to-day health system management. Demonstrating the true utility of ML in clinical practice and health governance requires evaluating models deployed within these complex real-world environments [<xref ref-type="bibr" rid="ref2">2</xref>,<xref ref-type="bibr" rid="ref15">15</xref>].</p><p>To address these gaps, this scoping review was guided by the following research questions:</p><list list-type="order"><list-item><p>How are ML algorithms currently being applied in real-world health care settings to support health systems planning and public policy formulation?</p></list-item><list-item><p>What are the main methodological characteristics, predictive performance measures, and clinical or operational targets of these deployed ML models?</p></list-item><list-item><p>What are the key gaps, methodological limitations, and barriers (eg, model calibration, external validation, and explainability) that prevent the translation of micro-operational ML models into macrolevel public health policies?</p></list-item></list></sec><sec id="s2"><title>Objective</title><p>Given the need for evidence that transcends theoretical models and addresses the complexity of health management, this scoping review aims to map and synthesize the application of ML algorithms to RWD. Specifically, we aim to identify how these tools are used in health systems planning, elective surgical flows, and public policy formulation, while characterizing the methodological landscape and identifying the limitations of the current literature.</p></sec><sec id="s3" sec-type="methods"><title>Methods</title><sec id="s3-1"><title>Overview</title><p>This scoping review was conducted in accordance with the Joanna Briggs Institute (JBI) methodology for scoping reviews [<xref ref-type="bibr" rid="ref16">16</xref>] and reported in accordance with the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines [<xref ref-type="bibr" rid="ref17">17</xref>]. The completed PRISMA-ScR checklist is provided in <xref ref-type="supplementary-material" rid="app3">Checklist 1</xref>. The study focused on pragmatic, nonpharmacological health interventions incorporating ML tools and algorithms to address gaps in clinical and public health contexts.</p></sec><sec id="s3-2"><title>Protocol Registration and Open Science Transparency</title><p>The protocol for this scoping review was registered retrospectively on OSF (ucefv) [<xref ref-type="bibr" rid="ref18">18</xref>]. Search syntaxes and documentation for both the initial search and the 2025 update were deposited and are publicly accessible via searchRxiv [<xref ref-type="bibr" rid="ref19">19</xref>-<xref ref-type="bibr" rid="ref22">22</xref>] and OSF repositories, ensuring full auditability and adherence to open science practices and PRISMA-ScR standards.</p></sec><sec id="s3-3"><title>Search Strategy</title><sec id="s3-3-1"><title>Overview</title><p>The search syntax and definition of the search time frame were built and expanded upon the methodological benchmark framework established by Triantafyllidis and Tsanas [<xref ref-type="bibr" rid="ref2">2</xref>] for ML in digital health. To capture the recent decade&#x2019;s acceleration in the development of health management algorithms while maintaining methodological continuity, the search time frame was defined from 2016 through 2025. The search was also expanded to include LILACS to cover regional public health systems.</p><p>The systematic search was conducted in 2 distinct phases: an initial search in August 2020 and a comprehensive update in June 2025.</p></sec><sec id="s3-3-2"><title>2020 Search Phase</title><p>The initial search strategy was developed by an experienced health librarian and structured as follows: (1) search string: &#x201C;machine learning&#x201D; AND &#x201C;patient monitoring;&#x201D; (2) period: 2016 to 2020; and (3) databases: MEDLINE (via PubMed), Embase, and Scopus. Specific details and results for this phase (n=333) are available in <xref ref-type="table" rid="table1">Table 1</xref> and <xref ref-type="supplementary-material" rid="app1">Multimedia Appendix 1</xref>.</p><table-wrap id="t1" position="float"><label>Table 1.</label><caption><p>Quality assessment of included studies using the Effective Public Health Practice Project (EPHPP) quality assessment tool<sup><xref ref-type="table-fn" rid="table1fn1">a</xref></sup>.</p></caption><table id="table1" frame="hsides" rules="groups"><thead><tr><td align="left" valign="top">Studies</td><td align="left" valign="top">Selection bias</td><td align="left" valign="top">Study design</td><td align="left" valign="top">Confusion factors</td><td align="left" valign="top">Blinding</td><td align="left" valign="top">Data collection methods</td><td align="left" valign="top">Withdrawals and dropouts</td><td align="left" valign="top">Global assessment</td></tr></thead><tbody><tr><td align="left" valign="top">Campagner et al [<xref ref-type="bibr" rid="ref23">23</xref>], 2020</td><td align="left" valign="top">Moderate</td><td align="left" valign="top">Moderate</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Moderate</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td></tr><tr><td align="left" valign="top">Ye et al [<xref ref-type="bibr" rid="ref24">24</xref>], 2019</td><td align="left" valign="top">Moderate</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Moderate</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Moderate</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td></tr><tr><td align="left" valign="top">Fontana et al [<xref ref-type="bibr" rid="ref25">25</xref>], 2019</td><td align="left" valign="top">Moderate</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Moderate</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Strong</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Moderate</td></tr><tr><td align="left" valign="top">Karhade et al [<xref ref-type="bibr" rid="ref26">26</xref>], 2018</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Strong</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td></tr><tr><td align="left" valign="top">Hayn et al [<xref ref-type="bibr" rid="ref27">27</xref>], 2017</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Moderate</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Strong</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td></tr><tr><td align="left" valign="top">Kr&#x00E4;mer et al [<xref ref-type="bibr" rid="ref28">28</xref>], 2017</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Strong</td><td align="left" valign="top">Weak</td><td align="left" valign="top">Weak</td></tr></tbody></table><table-wrap-foot><fn id="table1fn1"><p><sup>a</sup>This table was adapted from the study by Triantafyllidis and Tsanas [<xref ref-type="bibr" rid="ref2">2</xref>]. Ratings are based on the EPHPP tool, where &#x201C;strong&#x201D; indicates no weak ratings, &#x201C;moderate&#x201D; indicates 1 weak rating, and &#x201C;weak&#x201D; indicates 2 or more weak ratings in the specific domains.</p></fn></table-wrap-foot></table-wrap></sec><sec id="s3-3-3"><title>2025 Update Phase</title><p>The 2025 update expanded the scope to capture the latest advancements. The search strategy, developed by a health librarian, used the following 2 main conceptual blocks:</p><list list-type="bullet"><list-item><p>Block 1 (AI and ML) included the MeSH term &#x201C;Machine Learning&#x201D; and keywords such as &#x201C;artificial intelligence,&#x201D; &#x201C;deep learning,&#x201D; &#x201C;neural network,&#x201D; and &#x201C;predictive analytics.&#x201D;</p></list-item><list-item><p>Block 2 (health management) focused on organizational aspects and included MeSH terms such as &#x201C;Health Planning&#x201D; and &#x201C;Organizational Policy,&#x201D; combined with keywords such as &#x201C;integrated care&#x201D; and &#x201C;health systems.&#x201D;</p></list-item></list><p>This phase was conducted in June 2025 across 4 databases: MEDLINE (via PubMed), Embase, Scopus, and LILACS. The expanded strings are available via searchRxiv [<xref ref-type="bibr" rid="ref19">19</xref>-<xref ref-type="bibr" rid="ref22">22</xref>] and are presented in <xref ref-type="table" rid="table2">Table 2</xref> and <xref ref-type="supplementary-material" rid="app2">Multimedia Appendix 2</xref>.</p><table-wrap id="t2" position="float"><label>Table 2.</label><caption><p>Characteristics and predictive performance of the included studies on machine learning applications in health care management<sup><xref ref-type="table-fn" rid="table2fn1">a</xref></sup>.</p></caption><table id="table2" frame="hsides" rules="groups"><thead><tr><td align="left" valign="bottom">Items</td><td align="left" valign="bottom">Campagner et al [<xref ref-type="bibr" rid="ref23">23</xref>], 2020 (Italy)</td><td align="left" valign="bottom">Ye et al [<xref ref-type="bibr" rid="ref24">24</xref>], 2019 (United States)</td><td align="left" valign="bottom">Fontana et al [<xref ref-type="bibr" rid="ref25">25</xref>], 2019 (United States)</td><td align="left" valign="bottom">Karhade et al [<xref ref-type="bibr" rid="ref26">26</xref>], 2018 (United States)</td><td align="left" valign="bottom">Hayn et al [<xref ref-type="bibr" rid="ref27">27</xref>], 2017 (Austria)</td><td align="left" valign="bottom">Kr&#x00E4;mer et al [<xref ref-type="bibr" rid="ref28">28</xref>], 2019 (Germany)</td></tr></thead><tbody><tr><td align="left" valign="top">Application (objective)</td><td align="left" valign="top">Predict surgical invasiveness and recommend optimal procedures</td><td align="left" valign="top">Validate an ML<sup><xref ref-type="table-fn" rid="table2fn2">b</xref></sup>-based early warning system for in-hospital mortality</td><td align="left" valign="top">Predict the risk of not achieving MCID<sup><xref ref-type="table-fn" rid="table2fn3">c</xref></sup> in PROMs<sup><xref ref-type="table-fn" rid="table2fn4">d</xref></sup> after total knee arthroplasty</td><td align="left" valign="top">Predict nonroutine discharge after lumbar surgery to reduce hospital stay duration</td><td align="left" valign="top">Predict red blood cell transfusion volume for PBM<sup><xref ref-type="table-fn" rid="table2fn5">e</xref></sup></td><td align="left" valign="top">Predict emergency department vs elective care urgency</td></tr><tr><td align="left" valign="top">ML algorithms</td><td align="left" valign="top">Random forest (best-performing model), ordinal regression, SVM<sup><xref ref-type="table-fn" rid="table2fn6">f</xref></sup>, LASSO<sup><xref ref-type="table-fn" rid="table2fn7">g</xref></sup>, KNN<sup><xref ref-type="table-fn" rid="table2fn8">h</xref></sup>, and decision trees</td><td align="left" valign="top">Random forest (best-performing model), XGBoost<sup><xref ref-type="table-fn" rid="table2fn9">i</xref></sup>, gradient boosting, SVM, LASSO, and KNN</td><td align="left" valign="top">Logistic LASSO, random forest, and linear SVM</td><td align="left" valign="top">ANN<sup><xref ref-type="table-fn" rid="table2fn10">j</xref></sup> (best-performing model), SVM, BPM<sup><xref ref-type="table-fn" rid="table2fn11">k</xref></sup>, and decision tree; random forest used for variable selection</td><td align="left" valign="top">Random forest, multivariate logistic regression, linear models, and the Mercuriali algorithm</td><td align="left" valign="top">Random forest and KNN (for data imputation)</td></tr><tr><td align="left" valign="top">Key results (performance)</td><td align="left" valign="top">AUC<sup><xref ref-type="table-fn" rid="table2fn12">l</xref></sup> of 0.87 for the invasiveness score and 0.76 for procedure invasiveness</td><td align="left" valign="top">Strong discrimination (AUC 0.884); mean lead time of 40.8 hours</td><td align="left" valign="top">Fair-to-good ability to predict postsurgical MCID achievement</td><td align="left" valign="top">ANN outperformed the other models (PPV<sup><xref ref-type="table-fn" rid="table2fn13">m</xref></sup> 33%; sensitivity 54%)</td><td align="left" valign="top">Best-performing algorithm showed a moderate correlation (correlation coefficient=0.61)</td><td align="left" valign="top">96% accuracy and excellent discrimination (AUROC<sup><xref ref-type="table-fn" rid="table2fn14">n</xref></sup>&#x003E;0.99)</td></tr><tr><td align="left" valign="top">Statistical significance and evaluation metrics</td><td align="left" valign="top">Statistically significant (<italic>P</italic>=.04)</td><td align="left" valign="top">High AUC, indicating robust model validity</td><td align="left" valign="top">AUROC used as the primary evaluation metric</td><td align="left" valign="top">Superior AUROC, calibration, and Brier score</td><td align="left" valign="top">High statistical significance (<italic>P</italic>&#x003C;.001)</td><td align="left" valign="top">High precision and AUROC demonstrate validity</td></tr><tr><td align="left" valign="top">Clinical implications</td><td align="left" valign="top">High physician acceptance and consistent predictive variables</td><td align="left" valign="top">Improves patient safety through early detection of clinical deterioration</td><td align="left" valign="top">Focus on MCID strengthens practical clinical applicability</td><td align="left" valign="top">Open-access tool available to support replicability and validation</td><td align="left" valign="top">Optimize PBM; applicable to various medical fields</td><td align="left" valign="top">Improves patient flow management in emergency services</td></tr></tbody></table><table-wrap-foot><fn id="table2fn1"><p><sup>a</sup>This table was adapted from the study by Triantafyllidis and Tsanas [<xref ref-type="bibr" rid="ref2">2</xref>].</p></fn><fn id="table2fn2"><p><sup>b</sup>ML: machine learning.</p></fn><fn id="table2fn3"><p><sup>c</sup>MCID: minimal clinically important difference.</p></fn><fn id="table2fn4"><p><sup>d</sup>PROM: patient-reported outcome measure.</p></fn><fn id="table2fn5"><p><sup>e</sup>PBM: patient blood management.</p></fn><fn id="table2fn6"><p><sup>f</sup>SVM: support vector machine.</p></fn><fn id="table2fn7"><p><sup>g</sup>LASSO: least absolute shrinkage and selection operator.</p></fn><fn id="table2fn8"><p><sup>h</sup>KNN: k-nearest neighbors.</p></fn><fn id="table2fn9"><p><sup>i</sup>XGBoost: extreme gradient boosting.</p></fn><fn id="table2fn10"><p><sup>j</sup>ANN: artificial neural network.</p></fn><fn id="table2fn11"><p><sup>k</sup>BPM: Bayes point machine.</p></fn><fn id="table2fn12"><p><sup>l</sup>AUC: area under the curve.</p></fn><fn id="table2fn13"><p><sup>m</sup>PPV: positive predictive value.</p></fn><fn id="table2fn14"><p><sup>n</sup>AUROC: area under the receiver operating characteristic curve.</p></fn></table-wrap-foot></table-wrap></sec></sec><sec id="s3-4"><title>Eligibility Criteria</title><p>To capture the full spectrum of relevant literature, our search strategy encompassed 4 major databases (PubMed, Embase, Scopus, and LILACS) and covered the period from 2016 through 2025. The inclusion criteria required empirical studies that applied ML algorithms specifically to health services management, care regulation, surgical queue prioritization, or public policy formulation using RWD.</p><p>The inclusion and exclusion criteria were operationalized to exclude broad, nonapplied research. Studies were excluded if they (1) evaluated ML exclusively in silico or using synthetic or benchmark datasets, without validation against real-world health system records; (2) focused exclusively on clinical diagnostic accuracy (eg, radiological image classification or genomic sequencing) without linking algorithmic outputs to operational planning, bed allocation, or administrative decision-making; (3) consisted of conceptual frameworks, reviews, editorials, or conference abstracts without full empirical implementation data; or (4) examined closed commercial software without transparent algorithmic reporting.</p></sec><sec id="s3-5"><title>Study Selection and Screening</title><p>The selection process was conducted as follows: for 2020 records, a total of 333 were identified and imported into EndNote Basic (Clarivate Analytics) [<xref ref-type="bibr" rid="ref29">29</xref>], of which 4 duplicate records were removed. For 2025 records, a total of 611 were identified and imported into the Rayyan (Rayyan Systems Inc) web application (software as a service) [<xref ref-type="bibr" rid="ref30">30</xref>], of which 14 duplicate records were removed.</p></sec><sec id="s3-6"><title>Consolidation and Final Selection</title><p>A total of 944 records were identified across both phases. After removing all duplicate records (n=18), 926 records underwent independent title and abstract screening. Of these, 891 were excluded. For the remaining 35 reports, full-text retrieval was successfully achieved (100%). During the eligibility assessment, 29 reports were excluded (eg, because they were reviews, lacked technical detail, or inappropriate publication date). Ultimately, 6 studies met all the criteria and were included. The complete process is illustrated in the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram.</p></sec><sec id="s3-7"><title>Data Extraction and Methodological Quality Assessment</title><p>Data extraction was performed using a standardized charting form inspired by the TRIPOD (Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis) statement [<xref ref-type="bibr" rid="ref31">31</xref>], focusing on key predictive performance metrics such as area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity. Methodological quality was independently assessed using the Effective Public Health Practice Project (EPHPP) quality assessment tool [<xref ref-type="bibr" rid="ref32">32</xref>,<xref ref-type="bibr" rid="ref33">33</xref>] by a multidisciplinary team of specialists in computer science, dentistry, psychology, and statistics.</p></sec><sec id="s3-8"><title>Data Synthesis and Mapping</title><p>The data synthesis and mapping process followed a structured approach. Detailed information was extracted regarding the study objectives, methodological designs, sample sizes, and follow-up durations. Additionally, primary outcomes, reported levels of statistical significance, specific ML algorithms used, and their practical implications for clinical practice and public policy formulation were systematically mapped.</p></sec><sec id="s3-9"><title>Ethical Considerations</title><p>As this study is a scoping review of previously published literature and publicly available data, it did not involve direct interaction with human participants or the collection of identifiable private information. Consequently, formal institutional review board approval was not required. The research was conducted in accordance with the PRISMA-ScR guidelines to ensure transparency, reproducibility, and ethical scientific conduct (<xref ref-type="supplementary-material" rid="app3">Checklist 1</xref>).</p></sec></sec><sec id="s4" sec-type="results"><title>Results</title><sec id="s4-1"><title>Literature Search</title><p>The systematic search, conducted in 2 phases (2020 and 2025), yielded a total of 944 records across the MEDLINE (via PubMed), Embase, Scopus, and LILACS databases. After removing 18 duplicate records, 926 records underwent an initial title and abstract screening. Of these, 35 articles were selected for full-text review, with a 100% retrieval success rate.</p><p>Following the eligibility assessment, 29 reports were excluded for specific reasons (eg, reviews, case reports, or lack of technical detail regarding ML implementation). Ultimately<bold>,</bold> 6 studies met all the criteria and were included in the review. These studies originated from the United States (n=3), Austria (n=1), Germany (n=1), and Italy (n=1). The complete selection process is illustrated in the PRISMA flow diagram (<xref ref-type="fig" rid="figure1">Figure 1</xref>).</p><fig position="float" id="figure1"><label>Figure 1.</label><caption><p>PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 flow diagram for updated systematic reviews, which included searches of databases. Adapted from the study by Page et al [<xref ref-type="bibr" rid="ref34">34</xref>], reflecting the structure of the PRISMA 2020 flow diagram adapted for this study&#x2019;s database search.</p></caption><graphic alt-version="no" mimetype="image" position="float" xlink:type="simple" xlink:href="i-jmr_v15i1e96779_fig01.png"/></fig></sec><sec id="s4-2"><title>Quality Assessment</title><p>Study quality was assessed using the EPHPP tool. It was observed that 83% (5/6) [<xref ref-type="bibr" rid="ref23">23</xref>,<xref ref-type="bibr" rid="ref24">24</xref>,<xref ref-type="bibr" rid="ref26">26</xref>-<xref ref-type="bibr" rid="ref28">28</xref>] of the studies presented &#x201C;weak&#x201D; methodological quality, while only 17% (n=1) [<xref ref-type="bibr" rid="ref25">25</xref>] achieved a &#x201C;moderate&#x201D; rating. The primary weaknesses identified were selection bias and a lack of detailed information regarding participant recruitment. Regarding study design, observational cohorts prevailed (5/6, 83%) [<xref ref-type="bibr" rid="ref30">30</xref>-<xref ref-type="bibr" rid="ref34">34</xref>], with only 1 study characterized as a nonrandomized experimental design [<xref ref-type="bibr" rid="ref23">23</xref>]. These findings are detailed in <xref ref-type="table" rid="table1">Table 1</xref>.</p></sec><sec id="s4-3"><title>Intervention Type and Target Population</title><p>The samples varied significantly, using sources ranging from retrospective electronic health records to administrative insurance billing data. Interventions focused on patients undergoing elective surgeries (spine, hip, and knee), emergency care, red blood cell transfusion volume management, and urgency or emergency triage classification. The average follow-up duration across the studies was approximately 5 years.</p></sec><sec id="s4-4"><title>ML Applications and Clinical Outcomes</title><p>The identified ML applications were predominantly directed toward the development of predictive models and clinical decision support systems (CDSS). As synthesized in <xref ref-type="table" rid="table2">Table 2</xref>, key application objectives encompassed risk prediction, such as modeling in-hospital mortality and nonroutine discharge following lumbar surgeries, and surgical planning, which focused on evaluating invasiveness levels and estimating required red blood cell transfusion volumes. Furthermore, technologies were implemented in triage and diagnosis, including emergency department urgency classification and predicting minimal clinically important differences (MCID) following joint arthroplasties. Algorithms used across studies included artificial neural networks (ANNs), random forests (RF), support vector machines (SVMs), and other predictive prognostic models. Although 33% of the studies reported high statistical significance (<italic>P</italic>&#x003C;.001), a notable lack of metric standardization was evident. Regarding performance, the models demonstrated robust results, with area under the curve (AUC) values ranging from 0.76 to 0.99, indicating excellent discriminatory performance. These clinical implications highlight the expanding role of AI tools as modalities complementary to medical judgment, fostering data-driven and evidence-based practice.</p></sec></sec><sec id="s5" sec-type="discussion"><title>Discussion</title><sec id="s5-1"><title>Principal Findings</title><p>A critical reflection on the yield of this scoping review reveals that identifying only 6 empirical studies meeting all the strict inclusion criteria is highly informative. Rather than indicating a limitation of the search strategy, this restricted yield directly reflects an emerging evidence base in digital health. Although the theoretical literature on ML in health care is vast, true real-world implementation of ML for macrolevel health planning and public policy remains in its infancy. Most existing algorithms remain confined to laboratory simulations or isolated micro-operational tasks without integrating administrative, regulatory, or policy end points. The scarcity of published empirical deployments highlights a current structural gap between technical algorithmic development and its actual integration into public health governance, such as national waiting list regulation in universal health systems such as Brazil&#x2019;s Sistema &#x00DA;nico de Sa&#x00FA;de (SUS). Recognizing this field as an emerging area of research underscores the imperative to shift future research frameworks toward implementation science, model generalizability, and transparent calibration [<xref ref-type="bibr" rid="ref6">6</xref>,<xref ref-type="bibr" rid="ref35">35</xref>,<xref ref-type="bibr" rid="ref36">36</xref>].</p><p>The empirical evidence reveals that ML use in health planning is predominantly anchored in micro-operational, clinical, and process management decision support. In elective surgical settings, models were used to optimize resource allocation and anticipate operational bottlenecks. Hayn et al [<xref ref-type="bibr" rid="ref27">27</xref>] demonstrated that ML significantly outperformed routine practice in predicting surgical blood transfusion volumes for inventory management, while Karhade et al [<xref ref-type="bibr" rid="ref26">26</xref>] showed that preoperative identification of nonroutine discharges mitigates financial waste and bed turnover. For long-term outcomes, Fontana et al [<xref ref-type="bibr" rid="ref25">25</xref>] applied logistic least absolute shrinkage and selection operator (LASSO) regression (AUROC 0.60-0.89) to predict the preoperative achievement of the MCID in patient-reported outcome measures (PROMs), supporting shared decision-making. In diagnostic and systemic regulation, Campagner et al [<xref ref-type="bibr" rid="ref23">23</xref>] applied RF models (AUROC 0.76-0.87) to administrative and laboratory data to identify sample identification errors, potentially reducing unnecessary retesting and operational costs. At the macropolicy level, Kr&#x00E4;mer et al [<xref ref-type="bibr" rid="ref28">28</xref>] used an RF (accuracy=96%; AUROC&#x003E;0.99) with data from Germany&#x2019;s national inpatient database (which covered 99.84% of inpatient cases) to classify emergency and elective admissions. They revealed that emergency department growth was concentrated in low-urgency cases driven by hospital reimbursement incentives (diagnosis-related groups), proving ML&#x2019;s utility for regulatory auditing and financial incentive design. Additionally, Ye et al [<xref ref-type="bibr" rid="ref24">24</xref>] investigated in-hospital mortality prediction to support early warning systems for acute patient flow management.</p><p>These micro-operational ML outputs aggregate directly into 3 primary macrogovernance mechanisms: surgical queue regulation, by aggregating patient-level discharge predictions to optimize bed turnover; supply chain planning, by forecasting regional blood component demand (Hayn et al [<xref ref-type="bibr" rid="ref27">27</xref>]); and emergency care streamlining, by stabilizing patient flow management through automated error detection and risk triage (Kr&#x00E4;mer et al [<xref ref-type="bibr" rid="ref28">28</xref>]). To align with evidence-informed policy frameworks, ML deployment in universal health systems must address social determinants of health (SDOH) and health equity [<xref ref-type="bibr" rid="ref37">37</xref>]. As shown by Fontana et al [<xref ref-type="bibr" rid="ref25">25</xref>], socioeconomic variables influence predictive outcomes, posing risks that algorithms might unfairly prioritize patients with higher baseline probabilities of success over socially vulnerable individuals. Furthermore, data integrity at the source, including within systems such as e-SUS primary health care (PHC) [<xref ref-type="bibr" rid="ref38">38</xref>], must be prioritized to avoid automation bias (Campagner et al [<xref ref-type="bibr" rid="ref23">23</xref>]). Without continuous bias auditing, regulatory impact assessments, and robust data governance, including compliance with the General Data Protection Regulation (GDPR) [<xref ref-type="bibr" rid="ref39">39</xref>] or Lei Geral de Prote&#x00E7;&#x00E3;o de Dados Pessoais (LGPD) [<xref ref-type="bibr" rid="ref40">40</xref>], predictive algorithms risk amplifying historical disparities in access to care.</p></sec><sec id="s5-2"><title>Limitations</title><sec id="s5-2-1"><title>Methodological Vulnerabilities and Performance Limits</title><p>Interpreting these findings in the context of public policy requires caution. A TRIPOD-aligned appraisal revealed high discriminatory performance (AUROC 0.76-0.99) but a pervasive omission of formal model calibration measures, such as calibration slopes, intercepts [<xref ref-type="bibr" rid="ref35">35</xref>], or decision curve analysis. The upper bound (AUROC 0.99 in Kr&#x00E4;mer et al [<xref ref-type="bibr" rid="ref28">28</xref>]) reflects a narrow binary triage task, whereas complex health management tasks exhibited more modest discrimination (AUROC 0.76-0.84 in Karhade et al [<xref ref-type="bibr" rid="ref26">26</xref>]). In public health systems such as SUS, deploying miscalibrated models introduces severe risks of inappropriate queue prioritization or resource misallocation. Furthermore, 83% (5/6) of the studies exhibited overall weak methodological quality under the EPHPP tool [<xref ref-type="bibr" rid="ref32">32</xref>,<xref ref-type="bibr" rid="ref33">33</xref>], primarily due to high selection bias in single-center retrospective samples and unadjusted confounding. Only Campagner et al [<xref ref-type="bibr" rid="ref23">23</xref>] achieved a moderate rating through internal validation, although still limited to a single institution. Current evidence thus represents preliminary proof-of-concept evidence of feasibility rather than evidence supporting generalizable decision-making. Scalability requires prospective multicenter validation, adherence to TRIPOD-AI [<xref ref-type="bibr" rid="ref41">41</xref>] and PROBAST (Prediction Model Risk of Bias Assessment Tool)&#x2013;AI guidelines [<xref ref-type="bibr" rid="ref42">42</xref>], and compliance with FAIR (findable, accessible, interoperable, and reusable) data-sharing principles [<xref ref-type="bibr" rid="ref43">43</xref>].</p></sec><sec id="s5-2-2"><title>Implementation Barriers Across the Nonadoption, Abandonment, and Scale-up, Spread, and Sustainability Framework</title><p>Analyzing implementation barriers through the nonadoption, abandonment, and scale-up, spread, and sustainability (NASSS) framework [<xref ref-type="bibr" rid="ref6">6</xref>] shows that ML operates primarily as a decision support technology rather than an autonomous system. Regarding the technology and value proposition [<xref ref-type="bibr" rid="ref44">44</xref>,<xref ref-type="bibr" rid="ref45">45</xref>], the lack of external validation and calibration transparency [<xref ref-type="bibr" rid="ref35">35</xref>] restricts technical trust. Small sample sizes relative to high feature dimensionality and single-institution coding biases further compromise transferability. Concerning adopters, the &#x201C;black-box&#x201D; nature [<xref ref-type="bibr" rid="ref46">46</xref>] of complex models&#x2014;such as neural networks in Karhade et al [<xref ref-type="bibr" rid="ref26">26</xref>] or RF in Ye et al [<xref ref-type="bibr" rid="ref24">24</xref>], Campagner et al [<xref ref-type="bibr" rid="ref23">23</xref>], and Kr&#x00E4;mer et al [<xref ref-type="bibr" rid="ref28">28</xref>]&#x2014;fosters clinical skepticism. Multicollinearity further obscures coefficient interpretation (Fontana et al [<xref ref-type="bibr" rid="ref25">25</xref>]), underscoring the need for explainable AI (XAI) tools [<xref ref-type="bibr" rid="ref47">47</xref>,<xref ref-type="bibr" rid="ref48">48</xref>] such as SHAP [<xref ref-type="bibr" rid="ref49">49</xref>] (Shapley additive explanations) or LIME (local interpretable model-agnostic explanations), at the point of care. At the organizational level, data fragmentation across electronic health records and administrative systems hinders seamless integration. Furthermore, continuous real-time monitoring faces an &#x201C;intervention paradox,&#x201D; as described by Ye et al [<xref ref-type="bibr" rid="ref24">24</xref>] and Campagner et al [<xref ref-type="bibr" rid="ref23">23</xref>], wherein successful clinical interventions avert adverse outcomes, thereby dynamically altering the model&#x2019;s performance metrics into statistical false positives.</p></sec></sec><sec id="s5-3"><title>Future Directions</title><p>Bridging the current implementation gap requires 2 key future research priorities. First, adopting XAI [<xref ref-type="bibr" rid="ref47">47</xref>,<xref ref-type="bibr" rid="ref48">48</xref>] is critical to move beyond global discrimination metrics such as AUROC toward interpretable feature attributions, thereby ensuring ethical accountability and clinical trust. Second, integrating natural language processing (NLP) [<xref ref-type="bibr" rid="ref50">50</xref>] is necessary to overcome the limitation of current studies, which relied exclusively on structured tabular data while ignoring rich clinical narratives. Deploying domain-adapted NLP architectures, such as ClinicalBERT, enables the automated parsing of unstructured free-text records, including referral letters, emergency intake notes, and cancellation logs in regulation systems such as SisCNRAC. Structuring unstructured text converts administrative logs into actionable intelligence, optimizing waitlist regulation, syndromic surveillance, and resource allocation across public health networks.</p></sec><sec id="s5-4"><title>Conclusions</title><p>This scoping review synthesizes the empirical baseline of ML applications deployed for health system management and public policy execution. The evidence demonstrates that real-world applications remain in an incipient phase, operating predominantly as micro-operational prediction tools for surgical queue optimization, emergency triage, postdischarge destination planning, and blood component supply management. Although these operational algorithms offer preliminary feasibility for local resource optimization, the predominance of single-center retrospective cohorts and overall weak methodological quality&#x2014;as appraised via EPHPP&#x2014;prevent immediate, uncritical scalability to macrolevel health governance. Bridging the gap between micro-operational predictions and systemic health policy requires transitioning to prospective multicenter validations, adopting transparent reporting guidelines (TRIPOD-AI and PROBAST-AI), enforcing strict ethical data governance, and integrating domain-specific NLP to structure real-world health records.</p></sec></sec></body><back><ack><p>The authors express their gratitude to the Secretariat of Specialized Healthcare (SAES) and the Department of Healthcare Regulation, Evaluation and Control (DRAC) for their institutional support, with special thanks to Juliana F Lujan and D&#x00E9;bora Spalding Verdi for providing the necessary resources and support for this research. We are grateful to Departamento de Inform&#x00E1;tica do Sistema &#x00DA;nico de Sa&#x00FA;de, Secretaria de Informa&#x00E7;&#x00E3;o e Sa&#x00FA;de Digital, and Minist&#x00E9;rio da Sa&#x00FA;de for the development and support of the National High Complexity Regulation Center Information System (SisCNRAC). We also thank the Professional Doctorate Program in Governance and Digital Transformation at the Federal University of Tocantins (Universidade Federal do Tocantins), in partnership with the University of Bras&#x00ED;lia. Finally, the authors thank the Library of the Ministry of Health (Biblioteca Virtual em Sa&#x00FA;de do Minist&#x00E9;rio da Sa&#x00FA;de) and the National School of Public Health (Escola Nacional de Sa&#x00FA;de P&#x00FA;blica/Fiocruz) for their expert technical assistance with literature searching and bibliographic normalization.</p></ack><notes><sec><title>Generative AI Use</title><p>During the preparation of this work, the authors used a generative AI model, Gemini (Google) solely for technical language translation, syntactic refinement, and vocabulary optimization when translating the manuscript from Portuguese into academic English. The AI tool was not used to generate scientific concepts, synthesize data, conduct statistical analyses, or formulate research conclusions. Following translation and refinement, the authors thoroughly reviewed, edited, and validated all content to ensure scientific accuracy, and they take full responsibility for the final content of the manuscript.</p></sec><sec><title>Funding</title><p>The Brazilian Ministry of Health supported this study through the professional doctorate program. The funder had no role in the study design, data collection, analysis, interpretation of data, or writing of the manuscript. The authors had full access to all the data in the study and take responsibility for data integrity and the accuracy of the data analysis.</p></sec></notes><fn-group><fn fn-type="con"><p>DMAR contributed to the study&#x2019;s conceptualization, methodology, and formal analysis of the results. She led the literature screening and data extraction and drafted the original manuscript. GVB acted as the primary supervisor, contributing to the study&#x2019;s conceptualization and methodological oversight and providing critical revisions to the intellectual content. PHDSB contributed to the validation of the findings and the critical review of the final manuscript, offering strategic insights into governance and digital transformation policies. All authors have read and approved the final version of the manuscript and are accountable for all aspects of the work.</p></fn><fn fn-type="conflict"><p>None declared.</p></fn></fn-group><glossary><title>Abbreviations</title><def-list><def-item><term id="abb1">ANN</term><def><p>artificial neural network</p></def></def-item><def-item><term id="abb2">AUC</term><def><p>area under the curve</p></def></def-item><def-item><term id="abb3">AUROC</term><def><p>area under the receiver operating characteristic curve</p></def></def-item><def-item><term id="abb4">CDSS</term><def><p>clinical decision support system</p></def></def-item><def-item><term id="abb5">EPHPP</term><def><p>Effective Public Health Practice Project</p></def></def-item><def-item><term id="abb6">FAIR</term><def><p>findable, accessible, interoperable, and reusable</p></def></def-item><def-item><term id="abb7">GDP</term><def><p>gross domestic product</p></def></def-item><def-item><term id="abb8">GDPR</term><def><p>General Data Protection Regulation</p></def></def-item><def-item><term id="abb9">H5.0</term><def><p>Health 5.0</p></def></def-item><def-item><term id="abb10">JBI</term><def><p>Joanna Briggs Institute</p></def></def-item><def-item><term id="abb11">LASSO</term><def><p>least absolute shrinkage and selection operator</p></def></def-item><def-item><term id="abb12">LGPD</term><def><p>Lei Geral de Prote&#x00E7;&#x00E3;o de Dados Pessoais</p></def></def-item><def-item><term id="abb13">LIME</term><def><p>local interpretable model-agnostic explanations</p></def></def-item><def-item><term id="abb14">MCID</term><def><p>minimal clinically important difference</p></def></def-item><def-item><term id="abb15">ML</term><def><p>machine learning</p></def></def-item><def-item><term id="abb16">NASSS</term><def><p>nonadoption, abandonment, and scale-up, spread, and sustainability</p></def></def-item><def-item><term id="abb17">NLP</term><def><p>natural language processing</p></def></def-item><def-item><term id="abb18">PHC</term><def><p>primary health care</p></def></def-item><def-item><term id="abb19">PRISMA</term><def><p>Preferred Reporting Items for Systematic Reviews and Meta-Analyses</p></def></def-item><def-item><term id="abb20">PRISMA-ScR</term><def><p>Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews</p></def></def-item><def-item><term id="abb21">PROBAST</term><def><p>Prediction Model Risk of Bias Assessment Tool</p></def></def-item><def-item><term id="abb22">PROM</term><def><p>patient-reported outcome measure</p></def></def-item><def-item><term id="abb23">RF</term><def><p>random forest</p></def></def-item><def-item><term id="abb24">RQ</term><def><p>research question</p></def></def-item><def-item><term id="abb25">RWD</term><def><p>real-world data</p></def></def-item><def-item><term id="abb26">SDOH</term><def><p>social determinants of health</p></def></def-item><def-item><term id="abb27">SHAP</term><def><p>Shapley additive explanations</p></def></def-item><def-item><term id="abb28">SUS</term><def><p>Sistema &#x00DA;nico de Sa&#x00FA;de</p></def></def-item><def-item><term id="abb29">SVM</term><def><p>support vector machine</p></def></def-item><def-item><term id="abb30">TRIPOD</term><def><p>Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis</p></def></def-item><def-item><term id="abb31">XAI</term><def><p>explainable AI</p></def></def-item></def-list></glossary><ref-list><title>References</title><ref id="ref1"><label>1</label><nlm-citation citation-type="book"><person-group person-group-type="author"><collab>National Academy of Medicine</collab></person-group><person-group person-group-type="editor"><name name-style="western"><surname>Whicher</surname><given-names>D</given-names> </name><name name-style="western"><surname>Ahmed</surname><given-names>M</given-names> </name><name name-style="western"><surname>Israni</surname><given-names>ST</given-names> </name><name name-style="western"><surname>Matheny</surname><given-names>M</given-names> </name></person-group><source>Artificial Intelligence in Health Care: The Hope, the Hype, the Promise, the Peril</source><year>2023</year><publisher-name>National Academies Press (US)</publisher-name><pub-id pub-id-type="doi">10.17226/27111</pub-id></nlm-citation></ref><ref id="ref2"><label>2</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Triantafyllidis</surname><given-names>AK</given-names> </name><name name-style="western"><surname>Tsanas</surname><given-names>A</given-names> </name></person-group><article-title>Applications of machine learning in real-life digital health interventions: review of the literature</article-title><source>J Med Internet Res</source><year>2019</year><month>04</month><day>5</day><volume>21</volume><issue>4</issue><fpage>e12286</fpage><pub-id pub-id-type="doi">10.2196/12286</pub-id><pub-id pub-id-type="medline">30950797</pub-id></nlm-citation></ref><ref id="ref3"><label>3</label><nlm-citation citation-type="book"><person-group person-group-type="author"><name name-style="western"><surname>Russell</surname><given-names>SJ</given-names> </name><name name-style="western"><surname>Norvig</surname><given-names>P</given-names> </name></person-group><source>Artificial Intelligence: A Modern Approach</source><year>1995</year><publisher-name>Prentice Hall</publisher-name><pub-id pub-id-type="other">9780136042594</pub-id></nlm-citation></ref><ref id="ref4"><label>4</label><nlm-citation citation-type="book"><person-group person-group-type="author"><name name-style="western"><surname>Witten</surname><given-names>IH</given-names> </name><name name-style="western"><surname>Frank</surname><given-names>E</given-names> </name><name name-style="western"><surname>Hall</surname><given-names>MA</given-names> </name><name name-style="western"><surname>Pal</surname><given-names>CJ</given-names> </name></person-group><source>Data Mining: Practical Machine Learning Tools and Techniques</source><year>2017</year><publisher-name>Morgan Kaufmann</publisher-name><pub-id pub-id-type="doi">10.1016/C2015-0-02071-8</pub-id></nlm-citation></ref><ref id="ref5"><label>5</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Schwalbe</surname><given-names>N</given-names> </name><name name-style="western"><surname>Wahl</surname><given-names>B</given-names> </name></person-group><article-title>Artificial intelligence and the future of global health</article-title><source>Lancet</source><year>2020</year><month>05</month><day>16</day><volume>395</volume><issue>10236</issue><fpage>1579</fpage><lpage>1586</lpage><pub-id pub-id-type="doi">10.1016/S0140-6736(20)30226-9</pub-id><pub-id pub-id-type="medline">32416782</pub-id></nlm-citation></ref><ref id="ref6"><label>6</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Shaw</surname><given-names>J</given-names> </name><name name-style="western"><surname>Rudzicz</surname><given-names>F</given-names> </name><name name-style="western"><surname>Jamieson</surname><given-names>T</given-names> </name><name name-style="western"><surname>Goldfarb</surname><given-names>A</given-names> </name></person-group><article-title>Artificial intelligence and the implementation challenge</article-title><source>J Med Internet Res</source><year>2019</year><month>07</month><day>10</day><volume>21</volume><issue>7</issue><fpage>e13659</fpage><pub-id pub-id-type="doi">10.2196/13659</pub-id><pub-id pub-id-type="medline">31293245</pub-id></nlm-citation></ref><ref id="ref7"><label>7</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Vishnoi</surname><given-names>S</given-names> </name><name name-style="western"><surname>Yadav</surname><given-names>SL</given-names> </name><name name-style="western"><surname>Saini</surname><given-names>AK</given-names> </name><name name-style="western"><surname>Vats</surname><given-names>P</given-names> </name></person-group><article-title>Transforming healthcare delivery: assessing the role of e-governance strategies in IOMT and AIML-based medical system administration during the pandemic in the Indian subcontinent</article-title><source>Biomed Pharmacol J</source><year>2025</year><volume>18</volume><issue>December Spl Edition</issue><fpage>331</fpage><lpage>346</lpage><pub-id pub-id-type="doi">10.13005/bpj/3091</pub-id></nlm-citation></ref><ref id="ref8"><label>8</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Kaladharan</surname><given-names>S</given-names> </name><name name-style="western"><surname>Manayath</surname><given-names>D</given-names> </name><name name-style="western"><surname>Gopalakrishnan</surname><given-names>R</given-names> </name></person-group><article-title>Regulatory challenges in AI/ML-enabled medical devices: a scoping review and conceptual framework</article-title><source>J Med Device</source><year>2024</year><month>12</month><volume>18</volume><issue>4</issue><fpage>040801</fpage><pub-id pub-id-type="doi">10.1115/1.4066280</pub-id></nlm-citation></ref><ref id="ref9"><label>9</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Basulo-Ribeiro</surname><given-names>J</given-names> </name><name name-style="western"><surname>Teixeira</surname><given-names>L</given-names> </name></person-group><article-title>The future of healthcare with industry 5.0: preliminary interview-based qualitative analysis</article-title><source>Future Internet</source><year>2024</year><volume>16</volume><issue>3</issue><fpage>68</fpage><pub-id pub-id-type="doi">10.3390/fi16030068</pub-id></nlm-citation></ref><ref id="ref10"><label>10</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Reddy</surname><given-names>S</given-names> </name><name name-style="western"><surname>Fox</surname><given-names>J</given-names> </name><name name-style="western"><surname>Purohit</surname><given-names>MP</given-names> </name></person-group><article-title>Artificial intelligence-enabled healthcare delivery</article-title><source>J R Soc Med</source><year>2019</year><month>01</month><volume>112</volume><issue>1</issue><fpage>22</fpage><lpage>28</lpage><pub-id pub-id-type="doi">10.1177/0141076818815510</pub-id><pub-id pub-id-type="medline">30507284</pub-id></nlm-citation></ref><ref id="ref11"><label>11</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Saha</surname><given-names>E</given-names> </name><name name-style="western"><surname>Rathore</surname><given-names>P</given-names> </name></person-group><article-title>The impact of healthcare 4.0 technologies on healthcare supply chain performance: extending the organizational information processing theory</article-title><source>Technol Forecast Soc Change</source><year>2024</year><month>04</month><volume>201</volume><fpage>123256</fpage><pub-id pub-id-type="doi">10.1016/j.techfore.2024.123256</pub-id></nlm-citation></ref><ref id="ref12"><label>12</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Collins</surname><given-names>GS</given-names> </name><name name-style="western"><surname>Moons</surname><given-names>KG</given-names> </name></person-group><article-title>Reporting of artificial intelligence prediction models</article-title><source>Lancet</source><year>2019</year><month>04</month><day>20</day><volume>393</volume><issue>10181</issue><fpage>1577</fpage><lpage>1579</lpage><pub-id pub-id-type="doi">10.1016/S0140-6736(19)30037-6</pub-id><pub-id pub-id-type="medline">31007185</pub-id></nlm-citation></ref><ref id="ref13"><label>13</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Hazarika</surname><given-names>I</given-names> </name></person-group><article-title>Artificial intelligence: opportunities and implications for the health workforce</article-title><source>Int Health</source><year>2020</year><month>07</month><day>1</day><volume>12</volume><issue>4</issue><fpage>241</fpage><lpage>245</lpage><pub-id pub-id-type="doi">10.1093/inthealth/ihaa007</pub-id><pub-id pub-id-type="medline">32300794</pub-id></nlm-citation></ref><ref id="ref14"><label>14</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Lungu</surname><given-names>DA</given-names> </name><name name-style="western"><surname>Grillo Ruggieri</surname><given-names>T</given-names> </name><name name-style="western"><surname>Nuti</surname><given-names>S</given-names> </name></person-group><article-title>Decision making tools for managing waiting times and treatment rates in elective surgery</article-title><source>BMC Health Serv Res</source><year>2019</year><month>06</month><day>11</day><volume>19</volume><issue>1</issue><fpage>369</fpage><pub-id pub-id-type="doi">10.1186/s12913-019-4199-6</pub-id><pub-id pub-id-type="medline">31185989</pub-id></nlm-citation></ref><ref id="ref15"><label>15</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Agboola</surname><given-names>S</given-names> </name><name name-style="western"><surname>Hale</surname><given-names>TM</given-names> </name><name name-style="western"><surname>Masters</surname><given-names>C</given-names> </name><name name-style="western"><surname>Kvedar</surname><given-names>J</given-names> </name><name name-style="western"><surname>Jethwani</surname><given-names>K</given-names> </name></person-group><article-title>&#x201C;Real-world&#x201D; practical evaluation strategies: a review of telehealth evaluation</article-title><source>JMIR Res Protoc</source><year>2014</year><month>12</month><day>17</day><volume>3</volume><issue>4</issue><fpage>e75</fpage><pub-id pub-id-type="doi">10.2196/resprot.3459</pub-id><pub-id pub-id-type="medline">25524892</pub-id></nlm-citation></ref><ref id="ref16"><label>16</label><nlm-citation citation-type="book"><person-group person-group-type="author"><name name-style="western"><surname>Peters</surname><given-names>MD</given-names> </name><name name-style="western"><surname>Godfrey</surname><given-names>C</given-names> </name><name name-style="western"><surname>McInerney</surname><given-names>P</given-names> </name><name name-style="western"><surname>Munn</surname><given-names>Z</given-names> </name><name name-style="western"><surname>Tricco</surname><given-names>AC</given-names> </name><name name-style="western"><surname>Khalil</surname><given-names>H</given-names> </name></person-group><person-group person-group-type="editor"><name name-style="western"><surname>Aromataris</surname><given-names>E</given-names> </name><name name-style="western"><surname>Lockwood</surname><given-names>C</given-names> </name><name name-style="western"><surname>Porritt</surname><given-names>K</given-names> </name><name name-style="western"><surname>Pilla</surname><given-names>B</given-names> </name><name name-style="western"><surname>Jordan</surname><given-names>Z</given-names> </name></person-group><article-title>Scoping reviews</article-title><source>JBI Manual for Evidence Synthesis</source><year>2024</year><publisher-name>JBI</publisher-name><pub-id pub-id-type="doi">10.46658/JBIMES-24-09</pub-id></nlm-citation></ref><ref id="ref17"><label>17</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Tricco</surname><given-names>AC</given-names> </name><name name-style="western"><surname>Lillie</surname><given-names>E</given-names> </name><name name-style="western"><surname>Zarin</surname><given-names>W</given-names> </name><etal/></person-group><article-title>PRISMA Extension for Scoping Reviews (PRISMA-ScR): checklist and explanation</article-title><source>Ann Intern Med</source><year>2018</year><month>10</month><day>2</day><volume>169</volume><issue>7</issue><fpage>467</fpage><lpage>473</lpage><pub-id pub-id-type="doi">10.7326/M18-0850</pub-id><pub-id pub-id-type="medline">30178033</pub-id></nlm-citation></ref><ref id="ref18"><label>18</label><nlm-citation citation-type="other"><person-group person-group-type="author"><name name-style="western"><surname>Rocha</surname><given-names>DM</given-names> </name><name name-style="western"><surname>Barbosa</surname><given-names>GV</given-names> </name><name name-style="western"><surname>De Souza Bermejo</surname><given-names>PH</given-names> </name></person-group><article-title>Application of machine learning algorithms in health systems planning and public policy: a scoping review</article-title><source>Open Science Framework</source><comment>Preprint posted online on  Jul 10, 2026</comment><pub-id pub-id-type="doi">10.17605/OSF.IO/VWBHX</pub-id></nlm-citation></ref><ref id="ref19"><label>19</label><nlm-citation citation-type="other"><person-group person-group-type="author"><name name-style="western"><surname>Rocha</surname><given-names>DMA</given-names> </name></person-group><article-title>Search strategy about applications of machine learning algorithms in the context of real-life interventions and resulting clinical implications to guide planning and public policy formulation actions for the health system (MEDLINE)</article-title><source>searchRxiv</source><comment>Preprint posted online on  Sep 18, 2025</comment><pub-id pub-id-type="doi">10.1079/searchRxiv.2025.01103</pub-id></nlm-citation></ref><ref id="ref20"><label>20</label><nlm-citation citation-type="other"><person-group person-group-type="author"><name name-style="western"><surname>Rocha</surname><given-names>DM</given-names> </name></person-group><article-title>Search strategy about applications of machine learning algorithms in the context of real-life interventions and resulting clinical implications to guide planning and public policy formulation actions for the health system (EMBASE)</article-title><source>searchRxiv</source><comment>Preprint posted online on  Sep 18, 2025</comment><pub-id pub-id-type="doi">10.1079/searchRxiv.2025.01104</pub-id></nlm-citation></ref><ref id="ref21"><label>21</label><nlm-citation citation-type="other"><person-group person-group-type="author"><name name-style="western"><surname>Rocha</surname><given-names>DM</given-names> </name></person-group><article-title>Search strategy about applications of machine learning algorithms in the context of real-life interventions and resulting clinical implications to guide planning and public policy formulation actions for the health system (Scopus)</article-title><source>searchRxiv</source><comment>Preprint posted online on  Sep 18, 2025</comment><pub-id pub-id-type="doi">10.1079/searchRxiv.2025.01105</pub-id></nlm-citation></ref><ref id="ref22"><label>22</label><nlm-citation citation-type="other"><person-group person-group-type="author"><name name-style="western"><surname>Rocha</surname><given-names>DM</given-names> </name></person-group><article-title>Search strategy about applications of machine learning algorithms in the context of real-life interventions and resulting clinical implications to guide planning and public policy formulation actions for the health system (LILACS)</article-title><source>searchRxiv</source><comment>Preprint posted online on  Sep 18, 2025</comment><pub-id pub-id-type="doi">10.1079/searchRxiv.2025.01106</pub-id></nlm-citation></ref><ref id="ref23"><label>23</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Campagner</surname><given-names>A</given-names> </name><name name-style="western"><surname>Berjano</surname><given-names>P</given-names> </name><name name-style="western"><surname>Lamartina</surname><given-names>C</given-names> </name><name name-style="western"><surname>Langella</surname><given-names>F</given-names> </name><name name-style="western"><surname>Lombardi</surname><given-names>G</given-names> </name><name name-style="western"><surname>Cabitza</surname><given-names>F</given-names> </name></person-group><article-title>Assessment and prediction of spine surgery invasiveness with machine learning techniques</article-title><source>Comput Biol Med</source><year>2020</year><month>06</month><volume>121</volume><fpage>103796</fpage><pub-id pub-id-type="doi">10.1016/j.compbiomed.2020.103796</pub-id><pub-id pub-id-type="medline">32568677</pub-id></nlm-citation></ref><ref id="ref24"><label>24</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Ye</surname><given-names>C</given-names> </name><name name-style="western"><surname>Wang</surname><given-names>O</given-names> </name><name name-style="western"><surname>Liu</surname><given-names>M</given-names> </name><etal/></person-group><article-title>A real-time early warning system for monitoring inpatient mortality risk: prospective study using electronic medical record data</article-title><source>J Med Internet Res</source><year>2019</year><month>07</month><day>5</day><volume>21</volume><issue>7</issue><fpage>e13719</fpage><pub-id pub-id-type="doi">10.2196/13719</pub-id><pub-id pub-id-type="medline">31278734</pub-id></nlm-citation></ref><ref id="ref25"><label>25</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Fontana</surname><given-names>MA</given-names> </name><name name-style="western"><surname>Lyman</surname><given-names>S</given-names> </name><name name-style="western"><surname>Sarker</surname><given-names>GK</given-names> </name><name name-style="western"><surname>Padgett</surname><given-names>DE</given-names> </name><name name-style="western"><surname>MacLean</surname><given-names>CH</given-names> </name></person-group><article-title>Can machine learning algorithms predict which patients will achieve minimally clinically important differences from total joint arthroplasty?</article-title><source>Clin Orthop Relat Res</source><year>2019</year><month>06</month><volume>477</volume><issue>6</issue><fpage>1267</fpage><lpage>1279</lpage><pub-id pub-id-type="doi">10.1097/CORR.0000000000000687</pub-id><pub-id pub-id-type="medline">31094833</pub-id></nlm-citation></ref><ref id="ref26"><label>26</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Karhade</surname><given-names>AV</given-names> </name><name name-style="western"><surname>Ogink</surname><given-names>P</given-names> </name><name name-style="western"><surname>Thio</surname><given-names>Q</given-names> </name><etal/></person-group><article-title>Development of machine learning algorithms for prediction of discharge disposition after elective inpatient surgery for lumbar degenerative disc disorders</article-title><source>Neurosurg Focus</source><year>2018</year><month>11</month><day>1</day><volume>45</volume><issue>5</issue><fpage>E6</fpage><pub-id pub-id-type="doi">10.3171/2018.8.FOCUS18340</pub-id><pub-id pub-id-type="medline">30453463</pub-id></nlm-citation></ref><ref id="ref27"><label>27</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Hayn</surname><given-names>D</given-names> </name><name name-style="western"><surname>Kreiner</surname><given-names>K</given-names> </name><name name-style="western"><surname>Ebner</surname><given-names>H</given-names> </name><etal/></person-group><article-title>Development of multivariable models to predict and benchmark transfusion in elective surgery supporting patient blood management</article-title><source>Appl Clin Inform</source><year>2017</year><month>06</month><day>14</day><volume>8</volume><issue>2</issue><fpage>617</fpage><lpage>631</lpage><pub-id pub-id-type="doi">10.4338/ACI-2016-11-RA-0195</pub-id><pub-id pub-id-type="medline">28850152</pub-id></nlm-citation></ref><ref id="ref28"><label>28</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Kr&#x00E4;mer</surname><given-names>J</given-names> </name><name name-style="western"><surname>Schrey&#x00F6;gg</surname><given-names>J</given-names> </name><name name-style="western"><surname>Busse</surname><given-names>R</given-names> </name></person-group><article-title>Classification of hospital admissions into emergency and elective care: a machine learning approach</article-title><source>Health Care Manag Sci</source><year>2019</year><month>03</month><volume>22</volume><issue>1</issue><fpage>85</fpage><lpage>105</lpage><pub-id pub-id-type="doi">10.1007/s10729-017-9423-5</pub-id><pub-id pub-id-type="medline">29177993</pub-id></nlm-citation></ref><ref id="ref29"><label>29</label><nlm-citation citation-type="web"><article-title>EndNote</article-title><source>Clarivate</source><access-date>2026-08-23</access-date><comment><ext-link ext-link-type="uri" xlink:href="https://clarivate.com/accessibility/endnote/">https://clarivate.com/accessibility/endnote/</ext-link></comment></nlm-citation></ref><ref id="ref30"><label>30</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Ouzzani</surname><given-names>M</given-names> </name><name name-style="western"><surname>Hammady</surname><given-names>H</given-names> </name><name name-style="western"><surname>Fedorowicz</surname><given-names>Z</given-names> </name><name name-style="western"><surname>Elmagarmid</surname><given-names>A</given-names> </name></person-group><article-title>Rayyan-a web and mobile app for systematic reviews</article-title><source>Syst Rev</source><year>2016</year><month>12</month><day>5</day><volume>5</volume><issue>1</issue><fpage>210</fpage><pub-id pub-id-type="doi">10.1186/s13643-016-0384-4</pub-id><pub-id pub-id-type="medline">27919275</pub-id></nlm-citation></ref><ref id="ref31"><label>31</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Collins</surname><given-names>GS</given-names> </name><name name-style="western"><surname>Reitsma</surname><given-names>JB</given-names> </name><name name-style="western"><surname>Altman</surname><given-names>DG</given-names> </name><name name-style="western"><surname>Moons</surname><given-names>KG</given-names> </name></person-group><article-title>Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): the TRIPOD statement</article-title><source>Ann Intern Med</source><year>2015</year><month>01</month><day>6</day><volume>162</volume><issue>1</issue><fpage>55</fpage><lpage>63</lpage><pub-id pub-id-type="doi">10.7326/M14-0697</pub-id><pub-id pub-id-type="medline">25560714</pub-id></nlm-citation></ref><ref id="ref32"><label>32</label><nlm-citation citation-type="web"><article-title>EPHPP quality assessment tool for quantitative studies 2026</article-title><source>Elite Providers Hub for Progressive Play</source><access-date>2026-08-23</access-date><comment><ext-link ext-link-type="uri" xlink:href="https://www.ephpp.ca/quality-assessment-tool-for-quantitative-studies/">https://www.ephpp.ca/quality-assessment-tool-for-quantitative-studies/</ext-link></comment></nlm-citation></ref><ref id="ref33"><label>33</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Armijo-Olivo</surname><given-names>S</given-names> </name><name name-style="western"><surname>Stiles</surname><given-names>CR</given-names> </name><name name-style="western"><surname>Hagen</surname><given-names>NA</given-names> </name><name name-style="western"><surname>Biondo</surname><given-names>PD</given-names> </name><name name-style="western"><surname>Cummings</surname><given-names>GG</given-names> </name></person-group><article-title>Assessment of study quality for systematic reviews: a comparison of the Cochrane Collaboration Risk of Bias Tool and the Effective Public Health Practice Project Quality Assessment Tool: methodological research</article-title><source>J Eval Clin Pract</source><year>2012</year><month>02</month><volume>18</volume><issue>1</issue><fpage>12</fpage><lpage>18</lpage><pub-id pub-id-type="doi">10.1111/j.1365-2753.2010.01516.x</pub-id><pub-id pub-id-type="medline">20698919</pub-id></nlm-citation></ref><ref id="ref34"><label>34</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Page</surname><given-names>MJ</given-names> </name><name name-style="western"><surname>McKenzie</surname><given-names>JE</given-names> </name><name name-style="western"><surname>Bossuyt</surname><given-names>PM</given-names> </name><etal/></person-group><article-title>The PRISMA 2020 statement: an updated guideline for reporting systematic reviews</article-title><source>BMJ</source><year>2021</year><month>03</month><day>29</day><volume>372</volume><fpage>n71</fpage><pub-id pub-id-type="doi">10.1136/bmj.n71</pub-id><pub-id pub-id-type="medline">33782057</pub-id></nlm-citation></ref><ref id="ref35"><label>35</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Stevens</surname><given-names>RJ</given-names> </name><name name-style="western"><surname>Poppe</surname><given-names>KK</given-names> </name></person-group><article-title>Validation of clinical prediction models: what does the &#x201C;calibration slope&#x201D; really measure?</article-title><source>J Clin Epidemiol</source><year>2020</year><month>02</month><volume>118</volume><fpage>93</fpage><lpage>99</lpage><pub-id pub-id-type="doi">10.1016/j.jclinepi.2019.09.016</pub-id><pub-id pub-id-type="medline">31605731</pub-id></nlm-citation></ref><ref id="ref36"><label>36</label><nlm-citation citation-type="book"><person-group person-group-type="author"><name name-style="western"><surname>Toma</surname><given-names>TS</given-names> </name><name name-style="western"><surname>Pereira</surname><given-names>TV</given-names> </name><name name-style="western"><surname>Vanni</surname><given-names>T</given-names> </name><name name-style="western"><surname>Barreto</surname><given-names>JO</given-names> </name></person-group><source>Avaliac&#x0327;a&#x0303;o de tecnologias de sau&#x0301;de &#x0026; poli&#x0301;ticas informadas por evide&#x0302;ncias [Book in Portuguese]</source><year>2017</year><publisher-name>Instituto de Sa&#x00FA;de</publisher-name><pub-id pub-id-type="other">9788588169319</pub-id></nlm-citation></ref><ref id="ref37"><label>37</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Ramezani</surname><given-names>M</given-names> </name><name name-style="western"><surname>Mobinizadeh</surname><given-names>M</given-names> </name><name name-style="western"><surname>Bakhtiari</surname><given-names>A</given-names> </name><etal/></person-group><article-title>Agenda setting for health equity assessment through the lenses of social determinants of health using machine learning approach: a framework and preliminary pilot study</article-title><source>BioData Min</source><year>2025</year><month>02</month><day>10</day><volume>18</volume><issue>1</issue><fpage>14</fpage><pub-id pub-id-type="doi">10.1186/s13040-025-00428-x</pub-id><pub-id pub-id-type="medline">39930525</pub-id></nlm-citation></ref><ref id="ref38"><label>38</label><nlm-citation citation-type="web"><article-title>E-SUS primary health care strategy (e-SUS APS)</article-title><source>Ministry of Health, Brasil</source><year>2021</year><access-date>2026-10-29</access-date><comment><ext-link ext-link-type="uri" xlink:href="https://www.gov.br/saude/pt-br/assuntos/esua-aps">https://www.gov.br/saude/pt-br/assuntos/esua-aps</ext-link></comment></nlm-citation></ref><ref id="ref39"><label>39</label><nlm-citation citation-type="web"><article-title>Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation) (Text with EEA relevance)</article-title><source>European Union</source><year>2016</year><access-date>2026-10-29</access-date><comment><ext-link ext-link-type="uri" xlink:href="https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng">https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng</ext-link></comment></nlm-citation></ref><ref id="ref40"><label>40</label><nlm-citation citation-type="web"><article-title>General Personal Data Protection Law (LGPD) - law no. 13.709/2018</article-title><source>Federal Government of Brazil</source><year>2018</year><access-date>2026-10-29</access-date><comment><ext-link ext-link-type="uri" xlink:href="https://www2.senado.leg.br/bdsf/handle/id/545155">https://www2.senado.leg.br/bdsf/handle/id/545155</ext-link></comment></nlm-citation></ref><ref id="ref41"><label>41</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Collins</surname><given-names>GS</given-names> </name><name name-style="western"><surname>Moons</surname><given-names>KG</given-names> </name><name name-style="western"><surname>Dhiman</surname><given-names>P</given-names> </name><etal/></person-group><article-title>TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods</article-title><source>BMJ</source><year>2024</year><month>04</month><day>16</day><volume>385</volume><fpage>e078378</fpage><pub-id pub-id-type="doi">10.1136/bmj-2023-078378</pub-id><pub-id pub-id-type="medline">38626948</pub-id></nlm-citation></ref><ref id="ref42"><label>42</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Moons</surname><given-names>KG</given-names> </name><name name-style="western"><surname>Damen</surname><given-names>JA</given-names> </name><name name-style="western"><surname>Kaul</surname><given-names>T</given-names> </name><etal/></person-group><article-title>PROBAST+AI: an updated quality, risk of bias, and applicability assessment tool for prediction models using regression or artificial intelligence methods</article-title><source>BMJ</source><year>2025</year><month>03</month><day>24</day><volume>388</volume><fpage>e082505</fpage><pub-id pub-id-type="doi">10.1136/bmj-2024-082505</pub-id><pub-id pub-id-type="medline">40127903</pub-id></nlm-citation></ref><ref id="ref43"><label>43</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Wilkinson</surname><given-names>MD</given-names> </name><name name-style="western"><surname>Dumontier</surname><given-names>M</given-names> </name><name name-style="western"><surname>Aalbersberg</surname><given-names>IJ</given-names> </name><etal/></person-group><article-title>The FAIR guiding principles for scientific data management and stewardship</article-title><source>Sci Data</source><year>2016</year><month>03</month><day>15</day><volume>3</volume><fpage>160018</fpage><pub-id pub-id-type="doi">10.1038/sdata.2016.18</pub-id><pub-id pub-id-type="medline">26978244</pub-id></nlm-citation></ref><ref id="ref44"><label>44</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Porter</surname><given-names>ME</given-names> </name></person-group><article-title>A strategy for health care reform--toward a value-based system</article-title><source>N Engl J Med</source><year>2009</year><month>07</month><day>9</day><volume>361</volume><issue>2</issue><fpage>109</fpage><lpage>112</lpage><pub-id pub-id-type="doi">10.1056/NEJMp0904131</pub-id><pub-id pub-id-type="medline">19494209</pub-id></nlm-citation></ref><ref id="ref45"><label>45</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Porter</surname><given-names>ME</given-names> </name></person-group><article-title>What is value in health care?</article-title><source>N Engl J Med</source><year>2010</year><month>12</month><day>23</day><volume>363</volume><issue>26</issue><fpage>2477</fpage><lpage>2481</lpage><pub-id pub-id-type="doi">10.1056/NEJMp1011024</pub-id><pub-id pub-id-type="medline">21142528</pub-id></nlm-citation></ref><ref id="ref46"><label>46</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Qinggui</surname><given-names>H</given-names> </name><name name-style="western"><surname>Xiuqiong</surname><given-names>T</given-names> </name><name name-style="western"><surname>Hui</surname><given-names>C</given-names> </name><name name-style="western"><surname>Jinsong</surname><given-names>W</given-names> </name></person-group><article-title>The ethical risks and solutions of the algorithm black box in medical artificial intelligence</article-title><source>Acta Bioeth</source><year>2025</year><volume>31</volume><issue>1</issue><fpage>24</fpage><lpage>34</lpage><pub-id pub-id-type="doi">10.4067/S1726-569X2025000100024</pub-id></nlm-citation></ref><ref id="ref47"><label>47</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Huo</surname><given-names>Z</given-names> </name><name name-style="western"><surname>Booth</surname><given-names>J</given-names> </name><name name-style="western"><surname>Monks</surname><given-names>T</given-names> </name><etal/></person-group><article-title>Dynamic mortality prediction in critically ill children during interhospital transports to PICUs using explainable AI</article-title><source>NPJ Digit Med</source><year>2025</year><month>02</month><day>17</day><volume>8</volume><issue>1</issue><fpage>108</fpage><pub-id pub-id-type="doi">10.1038/s41746-025-01465-w</pub-id><pub-id pub-id-type="medline">39962177</pub-id></nlm-citation></ref><ref id="ref48"><label>48</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Gu</surname><given-names>F</given-names> </name><name name-style="western"><surname>Meyer</surname><given-names>AJ</given-names> </name><name name-style="western"><surname>Je&#x017E;ek</surname><given-names>F</given-names> </name><etal/></person-group><article-title>Identification of digital twins to guide interpretable AI for diagnosis and prognosis in heart failure</article-title><source>NPJ Digit Med</source><year>2025</year><month>02</month><day>18</day><volume>8</volume><issue>1</issue><fpage>110</fpage><pub-id pub-id-type="doi">10.1038/s41746-025-01501-9</pub-id><pub-id pub-id-type="medline">39966509</pub-id></nlm-citation></ref><ref id="ref49"><label>49</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Cammarata</surname><given-names>R</given-names> </name><name name-style="western"><surname>Ruffini</surname><given-names>F</given-names> </name><name name-style="western"><surname>Catamer&#x00F2;</surname><given-names>A</given-names> </name><etal/></person-group><article-title>Machine learning for predicting the low risk of postoperative pancreatic fistula after pancreaticoduodenectomy: toward a dynamic and personalized postoperative management strategy</article-title><source>Cancers (Basel)</source><year>2025</year><month>05</month><day>31</day><volume>17</volume><issue>11</issue><fpage>1846</fpage><pub-id pub-id-type="doi">10.3390/cancers17111846</pub-id><pub-id pub-id-type="medline">40507327</pub-id></nlm-citation></ref><ref id="ref50"><label>50</label><nlm-citation citation-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Singh</surname><given-names>NP</given-names> </name><name name-style="western"><surname>Mujawar</surname><given-names>MA</given-names> </name><name name-style="western"><surname>Golani</surname><given-names>A</given-names> </name></person-group><article-title>Role of artificial intelligence in enhancing mechanical ventilation - a peek into the future</article-title><source>Indian J Anaesth</source><year>2025</year><month>07</month><volume>69</volume><issue>7</issue><fpage>722</fpage><lpage>728</lpage><pub-id pub-id-type="doi">10.4103/ija.ija_995_24</pub-id><pub-id pub-id-type="medline">40657139</pub-id></nlm-citation></ref></ref-list><app-group><supplementary-material id="app1"><label>Multimedia Appendix 1</label><p>Initial search strategy and search strings.</p><media xlink:href="i-jmr_v15i1e96779_app1.docx" xlink:title="DOCX File, 114 KB"/></supplementary-material><supplementary-material id="app2"><label>Multimedia Appendix 2</label><p>Search strategies deposited in the searchRxiv repository [<xref ref-type="bibr" rid="ref19">19</xref>-<xref ref-type="bibr" rid="ref22">22</xref>]; this appendix also contains the final MEDLINE search strategy.</p><media xlink:href="i-jmr_v15i1e96779_app2.docx" xlink:title="DOCX File, 52 KB"/></supplementary-material><supplementary-material id="app3"><label>Checklist 1</label><p>PRISMA-ScR checklist.</p><media xlink:href="i-jmr_v15i1e96779_app3.docx" xlink:title="DOCX File, 109 KB"/></supplementary-material></app-group></back></article>