<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.1 20151215//EN" "https://jats.nlm.nih.gov/publishing/1.1/JATS-journalpublishing1.dtd">
<article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" article-type="research-article" dtd-version="1.1" specific-use="sps-1.9" xml:lang="pt">
  <front>
    <journal-meta>
      <journal-id journal-id-type="nlm-ta">Arq Bras Cardiol</journal-id>
      <journal-id journal-id-type="publisher-id">abc</journal-id>
      <journal-title-group>
        <journal-title>Arquivos Brasileiros de Cardiologia</journal-title>
        <abbrev-journal-title abbrev-type="publisher">Arq. Bras. Cardiol.</abbrev-journal-title>
      </journal-title-group>
      <issn pub-type="ppub">0066-782X</issn>
      <issn pub-type="epub">1678-4170</issn>
      <publisher>
        <publisher-name>Sociedade Brasileira de Cardiologia - SBC</publisher-name>
      </publisher>
    </journal-meta>
    <article-meta>
      <article-id specific-use="scielo-v3" pub-id-type="publisher-id">XVyRRmkytVYdx8hM6nd7qwM</article-id>
      <article-id specific-use="scielo-v2" pub-id-type="publisher-id">S0066-782X2025001100310</article-id>
      <article-id pub-id-type="other">00310</article-id>
      <article-id pub-id-type="doi">10.36660/abc.20250136</article-id>
      <article-categories>
        <subj-group subj-group-type="heading">
          <subject>Artigo Original</subject>
        </subj-group>
      </article-categories>
      <title-group>
        <article-title>Um Novo Escore de Risco Baseado em Aprendizado de M&#225;quina (Machine Learning) em Pacientes com Insufici&#234;ncia Card&#237;aca Aguda: O Escore ML-HF</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0009-0003-4069-8781</contrib-id>
          <name>
            <surname>Ferreira</surname>
            <given-names>Matheus Bissa Duarte</given-names>
          </name>
          <role>Concep&#231;&#227;o e desenho da pesquisa</role>
          <role>Obten&#231;&#227;o de dados</role>
          <role>An&#225;lise e interpreta&#231;&#227;o dos dados</role>
          <role>An&#225;lise estat&#237;stica</role>
          <role>Reda&#231;&#227;o do manuscrito</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0002-2385-0361</contrib-id>
          <name>
            <surname>Daikubara</surname>
            <given-names>Jorge Tadashi</given-names>
            <suffix>Neto</suffix>
          </name>
          <role>Concep&#231;&#227;o e desenho da pesquisa</role>
          <role>Obten&#231;&#227;o de dados</role>
          <role>An&#225;lise e interpreta&#231;&#227;o dos dados</role>
          <role>An&#225;lise estat&#237;stica</role>
          <role>Reda&#231;&#227;o do manuscrito</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Cunha</surname>
            <given-names>Gustavo S. Pereira da</given-names>
          </name>
          <role>Obten&#231;&#227;o de dados</role>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Moretti</surname>
            <given-names>Rafael</given-names>
          </name>
          <role>Obten&#231;&#227;o de dados</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Reichert</surname>
            <given-names>Jessica Tamires</given-names>
          </name>
          <role>Obten&#231;&#227;o de dados</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Prado</surname>
            <given-names>Lucas M&#252;ller</given-names>
          </name>
          <role>Obten&#231;&#227;o de dados</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Cirino</surname>
            <given-names>Raphael Henrique D&#233;a</given-names>
          </name>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Smith</surname>
            <given-names>Sidney C</given-names>
            <suffix>Jr</suffix>
          </name>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff2">
            <sup>2</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0003-0674-5997</contrib-id>
          <name>
            <surname>Taniguchi</surname>
            <given-names>F&#225;bio Papa</given-names>
          </name>
          <role>Concep&#231;&#227;o e desenho da pesquisa</role>
          <role>Obten&#231;&#227;o de financiamento</role>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff3">
            <sup>3</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Sposito</surname>
            <given-names>Andrei C.</given-names>
          </name>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff4">
            <sup>4</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Silvestre</surname>
            <given-names>Odilson M.</given-names>
          </name>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff5">
            <sup>5</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Nadruz</surname>
            <given-names>Wilson</given-names>
          </name>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff5">
            <sup>5</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0002-0095-9049</contrib-id>
          <name>
            <surname>Fernandes-Silva</surname>
            <given-names>Miguel Morita</given-names>
          </name>
          <role>Concep&#231;&#227;o e desenho da pesquisa</role>
          <role>An&#225;lise e interpreta&#231;&#227;o dos dados</role>
          <role>Reda&#231;&#227;o do manuscrito</role>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
      </contrib-group>
      <aff id="aff1">
        <label>1</label>
        <institution content-type="orgname">Universidade Federal do Paran&#225;</institution>
        <addr-line>
          <named-content content-type="city">Curitiba</named-content>
          <named-content content-type="state">PR </named-content>
        </addr-line>
        <country country="BR">Brasil</country>
        <institution content-type="original">Universidade Federal do Paran&#225;, Curitiba, PR &#8211; Brasil</institution>
      </aff>
      <aff id="aff2">
        <label>2</label>
        <institution content-type="orgname">University of North Carolina At Chapel Hill</institution>
        <addr-line>
          <named-content content-type="city">Chapel Hill</named-content>
          <named-content content-type="state">North Carolina</named-content>
        </addr-line>
        <country country="US">EUA </country>
        <institution content-type="original">University of North Carolina At Chapel Hill, Chapel Hill, North Carolina &#8211; EUA </institution>
      </aff>
      <aff id="aff3">
        <label>3</label>
        <institution content-type="orgname">Hospital do Cora&#231;&#227;o</institution>
        <addr-line>
          <named-content content-type="city">S&#227;o Paulo</named-content>
          <named-content content-type="state">SP</named-content>
        </addr-line>
        <country country="BR">Brasil</country>
        <institution content-type="original">Hospital do Cora&#231;&#227;o (HCor), S&#227;o Paulo, SP &#8211; Brasil</institution>
      </aff>
      <aff id="aff4">
        <label>4</label>
        <institution content-type="orgname">Universidade Estadual de Campinas</institution>
        <addr-line>
          <named-content content-type="city">Campinas</named-content>
          <named-content content-type="state">SP</named-content>
        </addr-line>
        <country country="BR">Brasil</country>
        <institution content-type="original">Universidade Estadual de Campinas, Campinas, SP &#8211; Brasil</institution>
      </aff>
      <aff id="aff5">
        <label>5</label>
        <institution content-type="orgname">Universidade Federal do Acre</institution>
        <addr-line>
          <named-content content-type="city">Rio Branco</named-content>
          <named-content content-type="state">AC</named-content>
        </addr-line>
        <country country="BR">Brasil </country>
        <institution content-type="original">Universidade Federal do Acre, Rio Branco, AC &#8211; Brasil </institution>
      </aff>
      <author-notes>
        <corresp id="c01"><label>Correspond&#234;ncia:</label> Miguel Morita Fernandes-Silva Universidade Federal do Paran&#225; &#8211; Rua General Carneiro, 181, 10&#186; andar. CEP 80060-900, Curitiba, PR &#8211; Brasil E-mail: <email>miguelmorita@ufpr.br</email> </corresp>
        <fn fn-type="edited-by">
          <label>Editor respons&#225;vel pela revis&#227;o:</label>
          <p>Gl&#225;ucia Maria Moraes de Oliveira</p>
        </fn>
        <fn fn-type="coi-statement">
          <label>Potencial conflito de interesse:</label>
          <p>N&#227;o h&#225; conflito com o presente artigo.</p>
        </fn>
      </author-notes>
      <pub-date date-type="pub" publication-format="electronic">
        <day>16</day>
        <month>01</month>
        <year>2026</year>
      </pub-date>
      <pub-date date-type="collection" publication-format="electronic">
        <month>11</month>
        <year>2025</year>
      </pub-date>
      <volume>122</volume>
      <issue>11</issue>
      <elocation-id>e20250136</elocation-id>
      <history>
        <date date-type="received">
          <day>24</day>
          <month>02</month>
          <year>2025</year>
        </date>
        <date date-type="rev-recd">
          <day>18</day>
          <month>06</month>
          <year>2025</year>
        </date>
        <date date-type="accepted">
          <day>28</day>
          <month>07</month>
          <year>2025</year>
        </date>
      </history>
      <permissions>
        <license license-type="open-access" xlink:href="https://creativecommons.org/licenses/by/4.0/" xml:lang="en">
          <license-p>Este &#233; um artigo de acesso aberto distribu&#237;do sob os termos da licen&#231;a de atribui&#231;&#227;o pelo Creative Commons</license-p>
        </license>
      </permissions>
      <abstract>
        <title>Resumo</title>
        <sec>
          <title>Fundamento</title>
          <p> Os escores de avalia&#231;&#227;o progn&#243;stica convencionais muitas vezes n&#227;o t&#234;m desempenho suficiente para prever a mortalidade em pacientes com insufici&#234;ncia card&#237;aca aguda (ICA).</p>
        </sec>
        <sec>
          <title>Objetivo</title>
          <p> Desenvolver e validar um escore progn&#243;stico baseado em aprendizado de m&#225;quina (<italic>Machine Learning - escore ML-HF)</italic> para prever morte hospitalar em pacientes com ICA e comparar seu desempenho com os principais escores tradicionais.</p>
        </sec>
        <sec>
          <title>M&#233;todos</title>
          <p> Pacientes admitidos por ICA em hospitais brasileiros do &#8220;Programa de Boas Pr&#225;ticas em Cardiologia&#8221; de 2016 a 2022 foram inclu&#237;dos. Dados cl&#237;nicos, resultados laboratoriais e o question&#225;rio de Qualidade de Vida da Organiza&#231;&#227;o Mundial da Sa&#250;de (<italic>World Health Organization Quality of Life)</italic> WHOQOL-Bref foram coletados na admiss&#227;o hospitalar. O desfecho foi &#243;bito hospitalar. O modelo foi treinado usando 70% das admiss&#245;es (conjunto de treinamento) e validado com os 30% restantes (conjunto de teste). A &#225;rea sob a curva ROC (AUC) do escore ML-HF foi comparada com a AUC dos escores tradicionais <italic>Acute Decompensated Heart Failure National Registry</italic> (ADHERE) e <italic>Get With the Guidelines&#8211;Heart Failure</italic> (GWTG-HF). O n&#237;vel de signific&#226;ncia foi de p &lt; 0,05.</p>
        </sec>
        <sec>
          <title>Resultados</title>
          <p> Foram inclu&#237;dos mil cento e cinquenta e sete pacientes hospitalizados por ICA. As cinco vari&#225;veis mais importantes do escore ML-HF foram: Qualidade do Dom&#237;nio de Sa&#250;de F&#237;sica (WHOQOL-BREF), s&#243;dio s&#233;rico, ureia s&#233;rica, creatinina s&#233;rica e press&#227;o arterial sist&#243;lica na admiss&#227;o hospitalar. No conjunto de teste, o escore ML-HF apresentou calibra&#231;&#227;o de modelo adequada (valor de p do teste de Hosmer-Lemeshow = 0,056) e discrimina&#231;&#227;o (AUC = 0,722 [IC95%, 0,661-0,783]), que foi superior aos escores GWTG-HF (AUC = 0,616 [IC95%, 0,529-0,702; p = 0,014]) e ADHERE (AUC = 0,601 [IC95%, 0,511-0,691; p = 0,006]).</p>
        </sec>
        <sec>
          <title>Conclus&#227;o</title>
          <p> Desenvolvemos e validamos um escore usando aprendizado de m&#225;quina para prever morte hospitalar em pacientes com ICA, que superou os escores tradicionais.</p>
        </sec>
      </abstract>
      <abstract abstract-type="graphical">
        <p>
          <fig id="f01">
            <label>Figura Central:</label>
            <caption>
              <title>Um Novo Escore de Risco Baseado em Aprendizado de M&#225;quina (Machine Learning) em Pacientes com Insufici&#234;ncia Card&#237;aca Aguda: O Escore ML-HF</title>
            </caption>
            <alternatives>
              <graphic xlink:href="https://minio.scielo.br/documentstore/1678-4170/XVyRRmkytVYdx8hM6nd7qwM/0687dcadf16f3d59fae27cc5155ce8379c09654b.tif"/>
              <graphic xlink:href="https://minio.scielo.br/documentstore/1678-4170/XVyRRmkytVYdx8hM6nd7qwM/1172be0681b96e00a7785d4449d0930e0320620b.png" specific-use="scielo-web"/>
              <graphic xlink:href="https://minio.scielo.br/documentstore/1678-4170/XVyRRmkytVYdx8hM6nd7qwM/2abcec4dd3abd251839c8c392218b34fea2824c9.jpg" specific-use="scielo-web" content-type="scielo-267x140"/>
            </alternatives>
          </fig>
        </p>
      </abstract>
      <kwd-group xml:lang="pt">
        <title>Palavras-chave</title>
        <kwd>Insufici&#234;ncia Card&#237;aca</kwd>
        <kwd>Aprendizado de M&#225;quina</kwd>
        <kwd>Morte</kwd>
      </kwd-group>
      <funding-group>
        <award-group>
          <funding-source>American Heart Association</funding-source>
        </award-group>
        <award-group>
          <funding-source>Hospital do Cora&#231;&#227;o</funding-source>
        </award-group>
        <award-group>
          <funding-source>Minist&#233;rio da Sa&#250;de</funding-source>
        </award-group>
        <funding-statement>O presente estudo foi financiado por American Heart Association, Hospital do Cora&#231;&#227;o (HCor) e Minist&#233;rio da Sa&#250;de.</funding-statement>
      </funding-group>
      <counts>
        <fig-count count="12"/>
        <table-count count="6"/>
        <equation-count count="0"/>
        <ref-count count="26"/>
      </counts>
    </article-meta>
  </front>
  <body>
    <sec sec-type="intro">
      <title>Introdu&#231;&#227;o</title>
      <p>A insufici&#234;ncia card&#237;aca (IC) afeta 25 milh&#245;es de pessoas em todo o mundo e apresenta altas taxas de mortalidade e hospitaliza&#231;&#227;o. No Brasil, ocorrem aproximadamente 250.000 interna&#231;&#245;es hospitalares por IC por ano, resultando em uma taxa de mortalidade hospitalar de 10,9%.<sup><xref ref-type="bibr" rid="B1">1</xref></sup></p>
      <p>As ferramentas de classifica&#231;&#227;o de risco ajudam a direcionar recursos e medidas intensivas para pacientes de alto risco. No entanto, estimar o progn&#243;stico dos pacientes na admiss&#227;o ou durante a hospitaliza&#231;&#227;o tem sido desafiador no cen&#225;rio da IC. Estudos anteriores desenvolveram escores, como o <italic>Acute Decompensated Heart Failure National Registry</italic> (ADHERE) e o <italic>Get With The Guidelines-Heart Failure</italic> (GWTG-HF), que podem ser derivados de umas poucas vari&#225;veis laboratoriais e cl&#237;nicas, para predizer a mortalidade hospitalar nesses pacientes.<sup><xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref></sup> Infelizmente, esses escores t&#234;m precis&#227;o abaixo do ideal e t&#234;m sido subutilizados na pr&#225;tica cl&#237;nica.<sup><xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref></sup></p>
      <p>Uma alternativa a essas abordagens convencionais s&#227;o os modelos baseados em aprendizado de m&#225;quina (<italic>Machine Learning</italic> - ML), que t&#234;m sido estudados nos &#250;ltimos anos para o manejo de IC.<sup><xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref></sup> A capacidade dos algoritmos de ML de analisar de forma mais profunda, multidimensional e n&#227;o linear uma variedade de vari&#225;veis aumenta seu potencial para melhorar os modelos de estratifica&#231;&#227;o de pacientes.<sup><xref ref-type="bibr" rid="B7">7</xref></sup> Modelos baseados em ML foram constru&#237;dos em pacientes com IC cr&#244;nica e aguda usando registros m&#233;dicos eletr&#244;nicos (RME), registros e ensaios cl&#237;nicos.<sup><xref ref-type="bibr" rid="B8">8</xref></sup> Esses modelos parecem ter precis&#227;o superior &#224; dos escores tradicionais ADHERE e GWTG-HF.<sup><xref ref-type="bibr" rid="B9">9</xref></sup> No entanto, a maioria desses estudos foi desenvolvida e validada em pa&#237;ses de alta renda. Diferen&#231;as na popula&#231;&#227;o de pacientes, determinantes sociais e presta&#231;&#227;o de cuidados de sa&#250;de podem impedir a implementa&#231;&#227;o desses modelos em pa&#237;ses de baixa e m&#233;dia renda (PBMR). De fato, j&#225; foi demonstrado que a capacidade dos modelos de ML de prever desfechos de IC varia significativamente entre diferentes subgrupos &#233;tnicos e grupos socioecon&#244;micos.<sup><xref ref-type="bibr" rid="B10">10</xref></sup></p>
      <p>Portanto, desenvolvemos e validamos um algoritmo baseado em ML para prever mortes hospitalares em pacientes com IC aguda, utilizando um amplo estudo de coorte em hospitais p&#250;blicos no Brasil. Comparamos ainda o algoritmo final baseado em ML com os modelos ADHERE e GWTG.</p>
    </sec>
    <sec sec-type="methods">
      <title>M&#233;todos</title>
      <sec>
        <title>Popula&#231;&#227;o</title>
        <p>Trata-se de uma an&#225;lise do programa &#8220;Boas Pr&#225;ticas em Cardiologia (BPC)&#8221;, adaptado da iniciativa &#8220;<italic>Get With The Guidelines</italic>&#8221; (GWTG) para melhorar a qualidade do atendimento cardiovascular e os desfechos dos pacientes com s&#237;ndrome coronariana aguda, IC e fibrila&#231;&#227;o atrial em hospitais p&#250;blicos brasileiros.<sup><xref ref-type="bibr" rid="B11">11</xref></sup> Resumidamente, foram selecionados hospitais p&#250;blicos terci&#225;rios em todas as cinco macrorregi&#245;es brasileiras desde janeiro de 2016.<sup><xref ref-type="bibr" rid="B12">12</xref></sup> Esses hospitais foram inclu&#237;dos em um estudo quase experimental que implementou uma interven&#231;&#227;o em n&#237;vel institucional para melhorar a ades&#227;o &#224;s diretrizes e que avaliou seu impacto nas medidas de desempenho institucional. Dados sociodemogr&#225;ficos, cl&#237;nicos, laboratoriais e ecocardiogr&#225;ficos foram coletados durante a admiss&#227;o dos pacientes. Os pacientes tamb&#233;m foram solicitados a preencher o question&#225;rio de Qualidade de Vida da Organiza&#231;&#227;o Mundial da Sa&#250;de (WHOQOL-Bref), que &#233; um instrumento de 26 itens que consiste em quatro dom&#237;nios: sa&#250;de f&#237;sica (7 itens), sa&#250;de psicol&#243;gica (6 itens), rela&#231;&#245;es sociais (3 itens) e sa&#250;de ambiental (8 itens). Os itens do WHOQOL-Bref consistem em escalas Likert de 5 pontos que, combinadas, fornecem cada dom&#237;nio de sa&#250;de. O procedimento para calcular o WHOQOL-Bref foi descrito em detalhe em outro lugar.<sup><xref ref-type="bibr" rid="B13">13</xref></sup>Como parte do programa BPC, todos os participantes foram acompanhados com visitas na alta, um m&#234;s e seis meses. Este estudo foi aprovado pelo comit&#234; de &#233;tica do Centro Coordenador (n&#250;mero 48561715.5.1001.0060) e de cada hospital participante. Todos os participantes assinaram um termo de consentimento livre e esclarecido para participa&#231;&#227;o no estudo.</p>
      </sec>
      <sec>
        <title>Crit&#233;rios de inclus&#227;o e exclus&#227;o</title>
        <p>Para esta an&#225;lise, foram inclu&#237;dos pacientes com 18 anos ou mais internados com diagn&#243;stico prim&#225;rio de IC aguda (c&#243;digos CID-10 I50; I50.0; I50.1 ou I50.9) entre 14/03/2016 e 07/12/2022 em qualquer hospital participante do programa BPC, que concordaram em participar do estudo e assinaram o termo de consentimento livre e esclarecido conforme ilustrado na <xref ref-type="fig" rid="f01">Figura Central</xref>. Os participantes precisavam ter ficado hospitalizados por pelo menos 24 horas para serem eleg&#237;veis.</p>
        <p>Exclu&#237;mos pacientes transferidos para outro hospital ou com estado vital ausente na alta. Pacientes que estavam em lista de espera para transplante card&#237;aco ou em uso de dispositivo de assist&#234;ncia ventricular (DAV) tamb&#233;m foram exclu&#237;dos. Tamb&#233;m exclu&#237;mos pacientes com dados ausentes para as vari&#225;veis selecionadas do modelo (veja a sele&#231;&#227;o de vari&#225;veis abaixo).</p>
        <p>Para construir e validar o modelo, dividimos o conjunto de dados em duas coortes: Coorte 1, pacientes internados de 16/01/2016 a 15/10/2021, e coorte 2, pacientes internados de 16/10/2021 a 07/12/2022. A coorte 1 foi utilizada nos processos de sele&#231;&#227;o de vari&#225;veis, constru&#231;&#227;o do modelo e valida&#231;&#227;o. A coorte 2 foi utilizada apenas para valida&#231;&#227;o externa (veja An&#225;lise estat&#237;stica para obter detalhes).</p>
      </sec>
    </sec>
    <sec>
      <title>Desfecho</title>
      <p>O desfecho foi morte por qualquer causa durante a hospitaliza&#231;&#227;o &#237;ndice.</p>
      <sec>
        <title>Sele&#231;&#227;o de vari&#225;veis</title>
        <p>Utilizamos todas as vari&#225;veis na admiss&#227;o no modelo preditivo, que incluiu dados sociodemogr&#225;ficos, cl&#237;nicos, laboratoriais e ecocardiogr&#225;ficos, juntamente com os escores derivados das respostas do question&#225;rio WHOQOL-Bref. Primeiramente, exclu&#237;mos vari&#225;veis com mais de 40% de dados ausentes.<sup><xref ref-type="bibr" rid="B14">14</xref></sup> Veja a tabela <xref ref-type="supplementary-material" rid="suppl01">Tabela Suplementar S1 </xref>para todas as vari&#225;veis e respectivas taxas de dados ausentes. Em seguida, aplicamos o m&#233;todo de ML &#8220;Boruta&#8221; no conjunto de treinamento (veja abaixo) para detectar automaticamente as vari&#225;veis mais relevantes que foram selecionadas para o algoritmo de treinamento. O m&#233;todo Boruta gera uma distribui&#231;&#227;o de probabilidade binomial do n&#250;mero de vezes que uma vari&#225;vel excede a import&#226;ncia de uma vari&#225;vel fict&#237;cia criada, que tem um valor aleat&#243;rio (n&#227;o preditivo).<sup><xref ref-type="bibr" rid="B15">15</xref></sup> As vari&#225;veis foram selecionadas se tivessem um valor de p &lt; 0,05 na cauda superior da distribui&#231;&#227;o de probabilidade. Durante esse processo de sele&#231;&#227;o, aplicamos um m&#233;todo de imputa&#231;&#227;o K-vizinhos mais pr&#243;ximos, selecionando os K vizinhos mais pr&#243;ximos (K = 5) com dados completos com base em uma dist&#226;ncia de Gower e imputando o valor m&#233;dio da vari&#225;vel completa aos vizinhos com valores ausentes.<sup><xref ref-type="bibr" rid="B16">16</xref></sup></p>
        <p>Ap&#243;s a sele&#231;&#227;o das vari&#225;veis mais relevantes, elas foram utilizadas para construir o algoritmo de ML no banco de dados n&#227;o imputado. A Figura S1 ilustra o processo de sele&#231;&#227;o de vari&#225;veis (<xref ref-type="supplementary-material" rid="suppl01">Figura Suplementar S1 </xref>). Realizamos uma An&#225;lise de Casos Completos, incluindo apenas pacientes com dados completos para as vari&#225;veis selecionadas. Assim, pacientes com dados ausentes para qualquer uma das vari&#225;veis selecionadas foram exclu&#237;dos (<xref ref-type="fig" rid="f02">Figura 1</xref>).</p>
        <p>
          <fig id="f02">
            <label>Figura 1</label>
            <caption>
              <title>&#8211; Fluxograma de pacientes inclu&#237;dos na an&#225;lise.</title>
            </caption>
            <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf01.tif"/></fig>
        </p>
      </sec>
      <sec>
        <title>An&#225;lise estat&#237;stica</title>
        <sec>
          <title>Estat&#237;sticas descritivas</title>
          <p>As vari&#225;veis cont&#237;nuas foram avaliadas quanto &#224; normalidade por meio da inspe&#231;&#227;o visual dos histogramas, assimetria e curtose, e pela aplica&#231;&#227;o do teste de Kolmogorov-Smirnov. Os dados s&#227;o apresentados como m&#233;dia &#177; desvio-padr&#227;o ou mediana (percentil 25 - percentil 75), conforme apropriado. As compara&#231;&#245;es entre os grupos foram realizadas utilizando o teste- t de Student n&#227;o pareado para vari&#225;veis com distribui&#231;&#227;o normal e o teste U de Mann-Whitney para vari&#225;veis com distribui&#231;&#227;o anormal. As vari&#225;veis categ&#243;ricas foram expressas como contagens e porcentagens e comparadas utilizando o teste exato de Fisher.</p>
        </sec>
        <sec>
          <title>Desenvolvimento e valida&#231;&#227;o de modelos</title>
          <p>A coorte 1 foi dividida aleatoriamente em conjuntos de treinamento (70%) e teste (30%). Este procedimento foi realizado por meio de amostragem aleat&#243;ria gerada por computador, sem reposi&#231;&#227;o, garantindo igual probabilidade de sele&#231;&#227;o de cada paciente. O conjunto de treinamento foi usado para treinamento do modelo e valida&#231;&#227;o interna. O conjunto de teste combinou 30% da coorte 1 com a coorte 2 completa e foi usado para valida&#231;&#227;o externa. Uma an&#225;lise de sensibilidade foi realizada usando apenas a coorte 2 como conjunto de teste. Este processo &#233; ilustrado na <xref ref-type="fig" rid="f03">Figura 2</xref>.</p>
          <p>
            <fig id="f03">
              <label>Figura 2</label>
              <caption>
                <title>&#8211; Fluxograma ilustrando a valida&#231;&#227;o interna e externa do modelo de escore ML-HF.</title>
              </caption>
              <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf02.tif"/></fig>
          </p>
        </sec>
        <sec>
          <title>Sele&#231;&#227;o de modelos</title>
          <p>Utilizamos um m&#233;todo de valida&#231;&#227;o cruzada de 5 etapas para selecionar o algoritmo de ML mais adequado. Este m&#233;todo consiste em subdividir o conjunto de treinamento em 5 grupos aleat&#243;rios. Quatro desses subconjuntos foram usados para treinar o modelo e o subconjunto restante para test&#225;-lo. Trata-se de um processo iterativo que ocorre 5 vezes, alterando o subconjunto de teste, conforme ilustrado na <xref ref-type="fig" rid="f03">Figura 2</xref>. O algoritmo com a melhor AUC m&#233;dia foi escolhido como modelo final.</p>
        </sec>
        <sec>
          <title>Desempenho de modelo</title>
          <p>Para avaliar a discrimina&#231;&#227;o do modelo, constru&#237;mos as curvas ROC (<italic>Receiver Operating Characteristic</italic>) e comparamos a &#225;rea sob a curva ROC (AUC) usando a morte hospitalar como desfecho. A calibra&#231;&#227;o do modelo foi avaliada plotando a probabilidade prevista versus a observada de morte hospitalar em decis e calculando o teste de qualidade de ajuste de Hosmer-Lemeshow. Um valor de p &lt; 0,05 indicou ajuste inadequado dos dados, enquanto valores de p mais altos indicaram calibra&#231;&#227;o adequada.</p>
        </sec>
        <sec>
          <title>Compara&#231;&#227;o com escores tradicionais</title>
          <p>Os escores ADHERE e GWTG-HF foram calculados para cada paciente. Como o escore GWTG-HF n&#227;o fornece diretamente as probabilidades de morte, derivamos uma regress&#227;o log&#237;stica de seus resultados num&#233;ricos.<sup><xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref></sup> Comparamos a AUC-ROC entre os escores usando o teste de DeLong nos conjuntos de treinamento e teste.<sup><xref ref-type="bibr" rid="B17">17</xref></sup></p>
          <p>Seguimos a diretriz de Relat&#243;rio Transparente de um Modelo de Predi&#231;&#227;o Multivari&#225;vel para Progn&#243;stico ou Diagn&#243;stico Individual (TRIPOD) para conduzir nossa pesquisa e escrever este artigo.<sup><xref ref-type="bibr" rid="B18">18</xref></sup></p>
          <p>Todas as an&#225;lises e treinamento do modelo foram conduzidos usando o software R. Consideramos estatisticamente significativos valores de p &lt; 0,05.</p>
        </sec>
      </sec>
    </sec>
    <sec sec-type="results">
      <title>Resultados</title>
      <p>De janeiro de 2016 a dezembro de 2022, um total de 3.532 pacientes com IC aguda foram inclu&#237;dos no programa BPC. Ap&#243;s a exclus&#227;o daqueles com DAV, na lista de espera para transplante ou sem estado vital na alta, 3.043 foram considerados eleg&#237;veis para o estudo (2.309 na coorte 1 e 734 na coorte 2, como visto na <xref ref-type="fig" rid="f02">Figura 1</xref>).</p>
      <p>Das 113 vari&#225;veis dispon&#237;veis na admiss&#227;o, nove foram exclu&#237;das devido &#224; aus&#234;ncia de mais de 40% de dados (Tabela Suplementar S1). O algoritmo de Boruta foi aplicado a 70% dos pacientes (n = 1.616) selecionados aleatoriamente da coorte 1, resultando em 17 vari&#225;veis selecionadas para construir o modelo de ML.</p>
      <p>Dos 3.043 pacientes eleg&#237;veis, 1.157 foram inclu&#237;dos na an&#225;lise final devido &#224; disponibilidade completa de dados para todas as 17 vari&#225;veis. O conjunto de treinamento consistiu em 621 pacientes, o que representou 70% dos 879 indiv&#237;duos da coorte 1, enquanto o conjunto de teste consistiu em 536 pacientes, compreendendo os 258 pacientes restantes da coorte 1 e 278 pacientes mais da coorte 2.</p>
      <p>A taxa de mortalidade hospitalar foi de 10,1%, com um total de 117 &#243;bitos entre os 1.157 pacientes examinados. A idade m&#233;dia desses pacientes foi de 61 &#177; 15 anos e a fra&#231;&#227;o de eje&#231;&#227;o do ventr&#237;culo esquerdo (FEVE) foi de 42 &#177; 17% e 58% eram do sexo masculino (<xref ref-type="table" rid="t1">Tabela 1</xref>). A <xref ref-type="table" rid="t2">Tabela 2</xref> exibe as vari&#225;veis selecionadas no escore ML-HF e aquelas usadas nos escores ADHERE e GWTG para compara&#231;&#227;o.</p>
      <p>
        <table-wrap id="t1">
          <label>Tabela 1</label>
          <caption>
            <title>&#8211; Caracter&#237;sticas basais dos pacientes inclu&#237;dos</title>
          </caption>
          <table frame="hsides" rules="groups">
            <colgroup>
              <col/>
              <col/>
              <col/>
              <col/>
            </colgroup>
            <thead>
              <tr>
                <th align="left">Par&#226;metros</th>
                <th style="font-weight:normal">Alta com vida (n=1040)</th>
                <th style="font-weight:normal">Morte hospitalar (n=117)</th>
                <th>Valor-p</th>
              </tr>
            </thead>
            <tbody>
              <tr>
                <td>
                  <bold>Idade, anos</bold>
                </td>
                <td align="center">61,3 &#177; 14,5</td>
                <td align="center">61,6 &#177; 14,4</td>
                <td align="center">0,829</td>
              </tr>
              <tr>
                <td>
                  <bold>Sexo feminino, n(%)</bold>
                </td>
                <td align="center">424 (40,8)</td>
                <td align="center">53 (45,3)</td>
                <td align="center">0,398</td>
              </tr>
              <tr>
                <td>
                  <bold>Ra&#231;a branca, n(%)</bold>
                </td>
                <td align="center">674 (64,8)</td>
                <td align="center">78 (66,7)</td>
                <td align="center">0,766</td>
              </tr>
              <tr>
                <td>
                  <bold>Educa&#231;&#227;o, n(%)</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>Analfabeto</td>
                <td align="center">84 (8,1)</td>
                <td align="center">9 (7,7)</td>
                <td align="center" rowspan="7">0,242</td>
              </tr>
              <tr>
                <td>Ensino fundamental incompleto</td>
                <td align="center">399 (38,4)</td>
                <td align="center">47 (40,2)</td>
              </tr>
              <tr>
                <td>Ensino fundamental completo</td>
                <td align="center">148 (14,2)</td>
                <td align="center">26 (22,2)</td>
              </tr>
              <tr>
                <td>Ensino m&#233;dio incompleto</td>
                <td align="center">72 (6,9)</td>
                <td align="center">8 (6,8)</td>
              </tr>
              <tr>
                <td>Ensino m&#233;dio completo</td>
                <td align="center">225 (21,6)</td>
                <td align="center">20 (17,1)</td>
              </tr>
              <tr>
                <td>Ensino superior incompleto</td>
                <td align="center">30 (2,9)</td>
                <td align="center">2 (1,7)</td>
              </tr>
              <tr>
                <td>Ensino superior completo</td>
                <td align="center">82 (7,9)</td>
                <td align="center">5 (4,3)</td>
              </tr>
              <tr>
                <td>
                  <bold>Renda m&#233;dia, n(%)</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>&#8804; 1 sal&#225;rio-m&#237;nimo</td>
                <td align="center">721 (69,3)</td>
                <td align="center">97 (82,9)</td>
                <td align="center">0,009</td>
              </tr>
              <tr>
                <td>
                  <bold>Fumante, n(%)</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>N&#227;o</td>
                <td align="center">520 (50)</td>
                <td align="center">56 (47,9)</td>
                <td align="center" rowspan="3">0,889</td>
              </tr>
              <tr>
                <td>Atual</td>
                <td align="center">438 (42,1)</td>
                <td align="center">52 (44,4)</td>
              </tr>
              <tr>
                <td>Passado</td>
                <td align="center">82 (7,9)</td>
                <td align="center">9 (7,7)</td>
              </tr>
              <tr>
                <td>
                  <bold>Alco&#243;latra, n(%)</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>N&#227;o</td>
                <td align="center">654 (62,9)</td>
                <td align="center">80 (68,4)</td>
                <td align="center" rowspan="3">0,200</td>
              </tr>
              <tr>
                <td>Atual</td>
                <td align="center">246 (23,7)</td>
                <td align="center">28 (23,9)</td>
              </tr>
              <tr>
                <td>Passado</td>
                <td align="center">140 (13,5)</td>
                <td align="center">9 (7,7)</td>
              </tr>
              <tr>
                <td>
                  <bold>Etiologia da IC, n(%)</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>Isqu&#234;mica</td>
                <td align="center">200 (19,2)</td>
                <td align="center">22 (18,8)</td>
                <td align="center" rowspan="4">&lt; 0,001</td>
              </tr>
              <tr>
                <td>Doen&#231;a de Chagas</td>
                <td align="center">95 (9,1)</td>
                <td align="center">13 (11,1)</td>
              </tr>
              <tr>
                <td>Doen&#231;a valvar</td>
                <td align="center">123 (11,8)</td>
                <td align="center">30 (25,6)</td>
              </tr>
              <tr>
                <td>Outro</td>
                <td align="center">622 (59,8)</td>
                <td align="center">52 (44,4)</td>
              </tr>
              <tr>
                <td>Pr&#243;tese valvar, n(%)</td>
                <td align="center">338 (32,5)</td>
                <td align="center">52 (44,4)</td>
                <td align="center">0,013</td>
              </tr>
              <tr>
                <td>Insufici&#234;ncia renal cr&#244;nica, n(%)</td>
                <td align="center">152 (14,6)</td>
                <td align="center">27 (23,1)</td>
                <td align="center">0,024</td>
              </tr>
              <tr>
                <td>Hipotireoidismo, n(%)</td>
                <td align="center">77 (7,4)</td>
                <td align="center">19 (16,2)</td>
                <td align="center">0,002</td>
              </tr>
              <tr>
                <td>FEVE, %</td>
                <td align="center">41,7 &#177; 17,1</td>
                <td align="center">41,90 &#177; 18,5</td>
                <td align="center">0,911</td>
              </tr>
              <tr>
                <td>Di&#226;metro do &#225;trio esquerdo, mm</td>
                <td align="center">45,8 &#177; 8,5</td>
                <td align="center">48,1 &#177; 9,4</td>
                <td align="center">0,005</td>
              </tr>
              <tr>
                <td>N&#250;mero de comorbidades<sup>&#8224;</sup></td>
                <td align="center">4 (3 - 5)</td>
                <td align="center">4 (2 - 5)</td>
                <td align="center">0,665</td>
              </tr>
              <tr>
                <td>
                  <bold>N&#250;mero de procedimentos cir&#250;rgicos anteriores, n(%)</bold>
                  <sup>
                    <bold>&#8225;</bold>
                  </sup>
                </td>
                <td> </td>
                <td> </td>
                <td align="center">0,002</td>
              </tr>
              <tr>
                <td>0</td>
                <td align="center">702 (67,5)</td>
                <td align="center">65 (55,6)</td>
                <td> </td>
              </tr>
              <tr>
                <td>1</td>
                <td align="center">288 (27,7)</td>
                <td align="center">36 (30,8)</td>
                <td> </td>
              </tr>
              <tr>
                <td>2</td>
                <td align="center">46 (4,4)</td>
                <td align="center">14 (12,0)</td>
                <td> </td>
              </tr>
              <tr>
                <td>3 ou mais</td>
                <td align="center">4 (0,3)</td>
                <td align="center">2 (1,7)</td>
                <td> </td>
              </tr>
              <tr>
                <td>
                  <bold>Caracter&#237;sticas cl&#237;nicas na admiss&#227;o</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>
                  <bold>Perfil hemodin&#226;mico, n(%)</bold>
                </td>
                <td> </td>
                <td> </td>
                <td align="center">0,020</td>
              </tr>
              <tr>
                <td>A</td>
                <td align="center">78 (7,5)</td>
                <td align="center">8 (6,8)</td>
                <td> </td>
              </tr>
              <tr>
                <td>B</td>
                <td align="center">797 (76,6)</td>
                <td align="center">77 (65,8)</td>
                <td> </td>
              </tr>
              <tr>
                <td>C</td>
                <td align="center">130 (12,5)</td>
                <td align="center">25 (21,4)</td>
                <td> </td>
              </tr>
              <tr>
                <td>L</td>
                <td align="center">35 (3,4)</td>
                <td align="center">7 (6,0)</td>
                <td> </td>
              </tr>
              <tr>
                <td>Press&#227;o arterial sist&#243;lica, mmHg</td>
                <td align="center">123 &#177; 29</td>
                <td align="center">115&#177; 27</td>
                <td align="center">0,004</td>
              </tr>
              <tr>
                <td>Press&#227;o arterial diast&#243;lica, mmHg</td>
                <td align="center">76 &#177; 17</td>
                <td align="center">71 &#177; 16</td>
                <td align="center">0,001</td>
              </tr>
              <tr>
                <td>Baixa perfus&#227;o perif&#233;rica, n(%)</td>
                <td align="center">127 (12,2)</td>
                <td align="center">20 (17,1)</td>
                <td align="center">0,175</td>
              </tr>
              <tr>
                <td>Ortopneia, n(%)</td>
                <td align="center">757 (74,6)</td>
                <td align="center">78 (70,9)</td>
                <td align="center">0,471</td>
              </tr>
              <tr>
                <td>Ascite, n(%)</td>
                <td align="center">242 (24,6)</td>
                <td align="center">39 (35,5)</td>
                <td align="center">0,019</td>
              </tr>
              <tr>
                <td>B3, n(%)</td>
                <td align="center">103 (10,3)</td>
                <td align="center">13 (11,5)</td>
                <td align="center">0,806</td>
              </tr>
              <tr>
                <td>Edema perif&#233;rico, n(%)</td>
                <td align="center">779 (75,6)</td>
                <td align="center">99 (85,3)</td>
                <td align="center">0,025</td>
              </tr>
              <tr>
                <td>
                  <bold>Resultados laboratoriais na admiss&#227;o (n&#237;veis sangu&#237;neos)</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>Creatinina, mg/dL</td>
                <td align="center">1,71 &#177; 1,51</td>
                <td align="center">1,93 &#177; 1,30</td>
                <td align="center">0,136</td>
              </tr>
              <tr>
                <td>BUN, mg/dL</td>
                <td align="center">33,07 &#177; 22,8</td>
                <td align="center">46,0 &#177; 30,4</td>
                <td align="center">&lt; 0,001</td>
              </tr>
              <tr>
                <td>S&#243;dio, mmol/L</td>
                <td align="center">136,6 &#177; 5,0</td>
                <td align="center">134,2 &#177; 6,9</td>
                <td align="center">&lt; 0,001</td>
              </tr>
              <tr>
                <td>Pot&#225;ssio, mmol/L</td>
                <td align="center">4,46 &#177; 0,77</td>
                <td align="center">4,66 &#177; 0,92</td>
                <td align="center">0,010</td>
              </tr>
              <tr>
                <td>Hemoglobina, g/dL</td>
                <td align="center">12,6 &#177; 2,28</td>
                <td align="center">12,32 &#177; 2,4</td>
                <td align="center">0,207</td>
              </tr>
              <tr>
                <td>
                  <bold>Question&#225;rio WHOQOL-BREF</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>Dom&#237;nio da sa&#250;de f&#237;sica</td>
                <td align="center">50,4 &#177; 19,4</td>
                <td align="center">39,5 &#177;15,8</td>
                <td align="center">&lt; 0,001</td>
              </tr>
              <tr>
                <td>Dom&#237;nio de rela&#231;&#245;es sociais</td>
                <td align="center">64,2 &#177; 16,5</td>
                <td align="center">61,1 &#177; 17,3</td>
                <td align="center">0,059</td>
              </tr>
              <tr>
                <td>Sa&#250;de ambiental</td>
                <td align="center">58,9 &#177; 12,6</td>
                <td align="center">54,5 &#177; 12,9</td>
                <td align="center">&lt;0,001</td>
              </tr>
              <tr>
                <td>Sa&#250;de psicol&#243;gica</td>
                <td align="center">77,3 &#177; 17,9</td>
                <td align="center">68,8 &#177; 18,0</td>
                <td align="center">&lt;0,001</td>
              </tr>
            </tbody>
          </table>
          <table-wrap-foot>
            <fn id="TFN1">
              <p>Vari&#225;veis categ&#243;ricas s&#227;o apresentadas como n (%) e vari&#225;veis cont&#237;nuas como m&#233;dia &#177; desvio padr&#227;o ou mediana (percentil 25 - percentil 75). IC: insufici&#234;ncia card&#237;aca; FEVE: fra&#231;&#227;o de eje&#231;&#227;o do ventr&#237;culo esquerdo; WHOQOL-BREF: Vers&#227;o resumida do question&#225;rio de qualidade de vida da Organiza&#231;&#227;o Mundial da Sa&#250;de; BUN: Nitrog&#234;nio ureico no sangue. &#8224; As comorbidades inclu&#237;ram hipertens&#227;o, diabetes, dislipidemia, acidente vascular cerebral, doen&#231;a carot&#237;dea, doen&#231;a arterial coronariana, doen&#231;a vascular perif&#233;rica, infarto agudo do mioc&#225;rdio, fibrila&#231;&#227;o atrial, doen&#231;a de Chagas, doen&#231;a card&#237;aca reum&#225;tica, doen&#231;a valvar, doen&#231;a pulmonar obstrutiva cr&#244;nica, apneia do sono, doen&#231;a hep&#225;tica, insufici&#234;ncia renal cr&#244;nica, anemia, hipertireoidismo, hipotireoidismo, depress&#227;o, c&#226;ncer de &#243;rg&#227;o s&#243;lido e c&#226;ncer hematol&#243;gico. &#8225; Procedimentos cir&#250;rgicos anteriores inclu&#237;ram angioplastia coron&#225;ria, cirurgia de revasculariza&#231;&#227;o do mioc&#225;rdio, marcapasso ou desfibrilador card&#237;aco implant&#225;vel, terapia de ressincroniza&#231;&#227;o card&#237;aca e pr&#243;tese valvar.</p>
            </fn>
          </table-wrap-foot>
        </table-wrap>
      </p>
      <p>
        <table-wrap id="t2">
          <label>Tabela 2</label>
          <caption>
            <title>&#8211; Compara&#231;&#227;o das vari&#225;veis utilizadas nos escores ADHERE, GWTG e ML-HF</title>
          </caption>
          <table frame="hsides" rules="groups">
            <colgroup>
              <col/>
              <col/>
              <col/>
              <col/>
              <col/>
            </colgroup>
            <thead>
              <tr>
                <th align="left">VARI&#193;VEL</th>
                <th>ADHERE</th>
                <th>GWTG</th>
                <th>ML-HF</th>
              </tr>
            </thead>
            <tbody>
              <tr>
                <td>
                  <bold>Demogr&#225;fico</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>Idade, anos</td>
                <td align="center">&#10003;</td>
                <td align="center">&#10003;</td>
                <td> </td>
              </tr>
              <tr>
                <td>Ra&#231;a negra</td>
                <td> </td>
                <td align="center">&#10003;</td>
                <td> </td>
              </tr>
              <tr>
                <td>
                  <bold>Caracter&#237;sticas cl&#237;nicas</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>BUN, mg/dL</td>
                <td align="center">&#10003;</td>
                <td align="center">&#10003;</td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Creatinina, mg/dL</td>
                <td align="center">&#10003;</td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Di&#226;metro do &#225;trio esquerdo</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>DPOC</td>
                <td> </td>
                <td align="center">&#10003;</td>
                <td> </td>
              </tr>
              <tr>
                <td>Doen&#231;a Renal Cr&#244;nica</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Etiologia valvar da IC</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Frequ&#234;ncia card&#237;aca</td>
                <td> </td>
                <td align="center">&#10003;</td>
                <td> </td>
              </tr>
              <tr>
                <td>Frequ&#234;ncia respirat&#243;ria</td>
                <td> </td>
                <td align="center">&#10003;</td>
                <td> </td>
              </tr>
              <tr>
                <td>Hemoglobina, g/dL</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Hipotireoidismo</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>N&#250;mero de procedimentos card&#237;acos anteriores</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Press&#227;o arterial sist&#243;lica, mmHg</td>
                <td align="center">&#10003;</td>
                <td align="center">&#10003;</td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Press&#227;o Arterial Diast&#243;lica, mmHg</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Perfil hemodin&#226;mico B</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Perfil hemodin&#226;mico C</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Perfil hemodin&#226;mico L</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>M&#225; perfus&#227;o perif&#233;rica</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>Pr&#243;tese valvar</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>S&#243;dio, mmol/L</td>
                <td> </td>
                <td align="center">&#10003;</td>
                <td align="center">&#10003;</td>
              </tr>
              <tr>
                <td>
                  <bold>Qualidade de vida</bold>
                </td>
                <td> </td>
                <td> </td>
                <td> </td>
              </tr>
              <tr>
                <td>Dom&#237;nio da sa&#250;de f&#237;sica &#8211; WHOQOL-BREF</td>
                <td> </td>
                <td> </td>
                <td align="center">&#10003;</td>
              </tr>
            </tbody>
          </table>
          <table-wrap-foot>
            <fn id="TFN2">
              <p>BUN: Nitrog&#234;nio ureico sangu&#237;neo; DPOC: doen&#231;a pulmonar obstrutiva cr&#244;nica; WHOQOL-BREF: Vers&#227;o resumida do question&#225;rio de qualidade de vida da Organiza&#231;&#227;o Mundial da Sa&#250;de.</p>
            </fn>
          </table-wrap-foot>
        </table-wrap>
      </p>
      <sec>
        <title>Sele&#231;&#227;o de algoritmo</title>
        <p>Dentre os algoritmos de ML avaliados, o modelo Random Forest apresentou o melhor desempenho na predi&#231;&#227;o de mortalidade hospitalar em compara&#231;&#227;o com Gradient Boosting Machines e Extreme Gradient Boosting (<xref ref-type="table" rid="t3">Tabela 3</xref>). Dessa forma, o algoritmo Random Forest foi escolhido para construir o modelo final.</p>
        <p>
          <table-wrap id="t3">
            <label>Tabela 3</label>
            <caption>
              <title>&#8211; Desempenho de discrimina&#231;&#227;o no conjunto de treinamento de acordo com o algoritmo de Aprendizado de M&#225;quina</title>
            </caption>
            <table frame="hsides" rules="groups">
              <colgroup>
                <col/>
                <col/>
              </colgroup>
              <thead>
                <tr>
                  <th align="left">Algoritmo de Aprendizado de M&#225;quina</th>
                  <th style="font-weight:normal">AUC (M&#233;dia &#177; DP) N = 810</th>
                </tr>
              </thead>
              <tbody>
                <tr>
                  <td>Random Forest</td>
                  <td align="center">0,772 &#177; 0,07</td>
                </tr>
                <tr>
                  <td>Gradient Boosting Machines</td>
                  <td align="center">0,703 &#177; 0,08</td>
                </tr>
                <tr>
                  <td>Extreme Gradient Boosting</td>
                  <td align="center">0,724 &#177; 0,09</td>
                </tr>
              </tbody>
            </table>
            <table-wrap-foot>
              <fn id="TFN3">
                <p>DP: desvio padr&#227;o; AUC: &#225;rea sob a curva de opera&#231;&#227;o de recep&#231;&#227;o.</p>
              </fn>
            </table-wrap-foot>
          </table-wrap>
        </p>
      </sec>
      <sec>
        <title>Valida&#231;&#227;o interna</title>
        <p>No conjunto de treinamento (n=621), o escore ML-HF demonstrou precis&#227;o de boa a excelente para prever mortalidade hospitalar, com uma AUC de 0,906, superando os escores ADHERE e GWTG-HF. [<xref ref-type="fig" rid="f04">Figura 3A</xref>] [<xref ref-type="fig" rid="f04">Figura Central</xref>].</p>
        <p>
          <fig id="f04">
            <label>Figura 3</label>
            <caption>
              <title>&#8211; Compara&#231;&#227;o das curvas ROC (Receiver Operating Characteristic) entre os escores ML-HF, GWTG-HF e ADHERE nos conjuntos de treinamento (A) e teste (B).</title>
            </caption>
            <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf03.tif"/></fig>
        </p>
      </sec>
      <sec>
        <title>Valida&#231;&#227;o externa</title>
        <p>No conjunto de teste (n=536), o escore ML-HF manteve uma discrimina&#231;&#227;o aceit&#225;vel, com uma AUC de 0,722. Embora inferior ao conjunto de treinamento, seu desempenho permaneceu superior aos escores ADHERE e GWTG. [<xref ref-type="fig" rid="f04">Figura 3B</xref>, Figura Central]. Resultados semelhantes foram observados ao utilizar apenas a coorte 2 completa como conjunto de teste (n=280), conforme mostrado na Figura S2 Suplementar.</p>
      </sec>
      <sec>
        <title>Calibra&#231;&#227;o do modelo</title>
        <p>A <xref ref-type="fig" rid="f05">Figura 4</xref> mostra a mortalidade hospitalar observada versus prevista para cada escore no conjunto de teste. Embora o escore ML-HF tenha apresentado boa calibra&#231;&#227;o do modelo, a mortalidade prevista divergiu significativamente da mortalidade observada nos outros dois escores convencionais. Especificamente, o escore ADHERE subestimou o risco, enquanto o escore GTWG o superestimou.</p>
        <p>
          <fig id="f05">
            <label>Figura 4</label>
            <caption>
              <title>&#8211; Compara&#231;&#227;o da mortalidade hospitalar observada e prevista para avaliar a calibra&#231;&#227;o do modelo dos escores ML-HF (A), GWTG-HF (B) e ADHERE (C) no conjunto de teste. HL: Teste de Hosmer-Lemeshow. Um valor de p menor que 0,05 indica um ajuste inadequado dos dados.</title>
            </caption>
            <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf04.tif"/></fig>
        </p>
      </sec>
      <sec>
        <title>Import&#226;ncia da vari&#225;vel no escore ML-HF</title>
        <p>A <xref ref-type="fig" rid="f06">Figura 5</xref> mostra a import&#226;ncia relativa das vari&#225;veis do escore ML-HF em ordem decrescente. A vari&#225;vel com maior import&#226;ncia foi o dom&#237;nio Sa&#250;de F&#237;sica do WHOQOL, seguido pelos n&#237;veis sangu&#237;neos de s&#243;dio, ureia e creatinina, conforme mostrado na Figura Central. A press&#227;o arterial na admiss&#227;o tamb&#233;m estava entre as dez vari&#225;veis mais importantes, com a press&#227;o arterial sist&#243;lica e a press&#227;o arterial diast&#243;lica ocupando a quinta e a oitava posi&#231;&#245;es, respectivamente. Notavelmente, a FEVE n&#227;o teve import&#226;ncia suficiente para predizer a mortalidade hospitalar no m&#233;todo de Boruta. O &#250;nico par&#226;metro ecocardiogr&#225;fico no modelo foi o tamanho do &#225;trio esquerdo, ocupando o sexto lugar em ordem de import&#226;ncia.</p>
        <p>
          <fig id="f06">
            <label>Figura 5</label>
            <caption>
              <title>&#8211; Import&#226;ncia das vari&#225;veis inclu&#237;das no escore ML-HF. WHOQOL-BREF: Vers&#227;o resumida do question&#225;rio de qualidade de vida da Organiza&#231;&#227;o Mundial da Sa&#250;de; PAS: press&#227;o arterial sist&#243;lica; PAD: press&#227;o arterial diast&#243;lica; BUN: Nitrog&#234;nio ureico no sangue.</title>
            </caption>
            <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf05.tif"/></fig>
        </p>
      </sec>
    </sec>
    <sec sec-type="discussion">
      <title>Discuss&#227;o</title>
      <p>Este estudo apresenta tr&#234;s principais achados. Desenvolvemos um novo escore baseado em ML &#8211; o escore ML-HF &#8211; para prever a mortalidade hospitalar em pacientes com ICA em uma PBMR, que foi validado externamente com acur&#225;cia aceit&#225;vel e boa calibra&#231;&#227;o. O escore ML-HF superou os escores convencionais ADHERE e GWTG-HF na mesma popula&#231;&#227;o.<sup><xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B5">5</xref></sup> O estudo revelou que as informa&#231;&#245;es coletadas a partir de um question&#225;rio de qualidade de vida emergiram como o preditor mais importante de mortalidade. No geral, este estudo fornece uma nova ferramenta para prever o risco em pacientes com IC aguda e destaca o potencial do ML para aprimorar os modelos de estratifica&#231;&#227;o de risco nessa popula&#231;&#227;o.</p>
      <p>At&#233; onde sabemos, este &#233; o primeiro estudo a incorporar uma medida de qualidade de vida pessoal em um modelo baseado em ML para prever mortalidade na IC. Surpreendentemente, o dom&#237;nio sa&#250;de f&#237;sica do question&#225;rio WHOQOL-BREF emergiu como o preditor mais importante de mortalidade, superando biomarcadores s&#233;ricos e par&#226;metros ecocardiogr&#225;ficos. Esse achado novo e inesperado provavelmente s&#243; foi poss&#237;vel gra&#231;as ao uso de t&#233;cnicas de ML, pois elas permitem a an&#225;lise de vari&#225;veis de maneira n&#227;o linear e predefinida. Este &#233; um exemplo de como o ML pode analisar padr&#245;es complexos em dados para encontrar preditores contraintuitivos da perspectiva humana. Este resultado ressalta a import&#226;ncia de incorporar os resultados relatados pelo paciente (RRP) e considerar a autopercep&#231;&#227;o do paciente sobre sua sa&#250;de. Em termos mais simples, destaca a import&#226;ncia de realmente ouvir o paciente.</p>
      <p>Estudos anteriores demonstraram o valor progn&#243;stico do RRP em pacientes com IC cr&#244;nica. No registro CHAMP-HF (<italic>Change the Management of Patients With Heart Failure</italic>), uma melhora de 5 pontos no Question&#225;rio de Cardiomiopatia de Kansas City foi associada a uma redu&#231;&#227;o de 41% no risco de morte por todas as causas.<sup><xref ref-type="bibr" rid="B19">19</xref></sup> Notavelmente, o RRP provou ser um melhor preditor de eventos futuros em compara&#231;&#227;o com a classe funcional da NYHA, ressaltando a import&#226;ncia de avaliar a perspectiva do paciente sobre seu estado de sa&#250;de, al&#233;m da avalia&#231;&#227;o cl&#237;nica.<sup><xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B20">20</xref></sup> De fato, as diretrizes atuais de IC recomendam a avalia&#231;&#227;o padronizada do RRP como uma recomenda&#231;&#227;o de classe 2a.<sup><xref ref-type="bibr" rid="B10">10</xref></sup> Nosso estudo contribui para a literatura ao demonstrar o papel progn&#243;stico independente da medi&#231;&#227;o do RRP na admiss&#227;o de pacientes com IC descompensada.</p>
      <p>&#201; poss&#237;vel destacar o fato de que novas vari&#225;veis, al&#233;m do dom&#237;nio de sa&#250;de f&#237;sica do WHOQOL-Brief, foram relevantes para a predi&#231;&#227;o de &#243;bito hospitalar em nosso estudo, fatores que n&#227;o s&#227;o avaliados em outros escores, como hemoglobina, di&#226;metro do &#225;trio esquerdo, presen&#231;a de pr&#243;tese valvar, n&#250;mero de procedimentos cir&#250;rgicos e presen&#231;a de hipotireoidismo. A relev&#226;ncia da presen&#231;a de pr&#243;tese valvar pode ser devida &#224; alta preval&#234;ncia de febre reum&#225;tica em nossa popula&#231;&#227;o. Isso, mais uma vez, destaca a import&#226;ncia do desenvolvimento de escores personalizados para PBMR.<sup><xref ref-type="bibr" rid="B21">21</xref></sup></p>
      <p>Modelos de ML foram testados anteriormente para prever a mortalidade hospitalar de pacientes com IC aguda em outros pa&#237;ses. O modelo MARKER-HF (<italic>Machine learning Assessment of RisK and EaRly mortality in Heart Failure</italic>) demonstrou excelente poder discriminat&#243;rio, com uma AUC de 0,88. No entanto, esse modelo foi desenvolvido usando pacientes com IC cr&#244;nica e aguda, com dura&#231;&#245;es de acompanhamento variadas.<sup><xref ref-type="bibr" rid="B9">9</xref></sup> Segar et al. utilizaram dados do registro GWTG nos Estados Unidos para construir um modelo baseado em ML para prever a mortalidade hospitalar, que foi validado na coorte ARIC (<italic>Atherosclerosis Risk in Community</italic>).<sup><xref ref-type="bibr" rid="B5">5</xref></sup> Assim como em nossa investiga&#231;&#227;o, eles observaram que o modelo baseado em ML superou o escore GWTG-HF, apresentando um &#237;ndice C de aproximadamente 0,80. Eles puderam incorporar biomarcadores s&#233;ricos, como n&#237;veis sangu&#237;neos de troponina e pept&#237;deos natriur&#233;ticos, e determinantes sociais de sa&#250;de em n&#237;vel contextual, o que ajuda a reduzir o vi&#233;s do algoritmo e melhorar a equidade entre popula&#231;&#245;es carentes.<sup><xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref></sup></p>
      <p>Diferen&#231;as nas popula&#231;&#245;es de pacientes, no manejo cl&#237;nico e na disponibilidade de biomarcadores, como pept&#237;deos natriur&#233;ticos, podem explicar, em parte, a maior precis&#227;o do modelo de Segar et al.<sup><xref ref-type="bibr" rid="B5">5</xref></sup> em compara&#231;&#227;o ao nosso; no entanto, nosso modelo pode ser particularmente valioso em cen&#225;rios com acesso limitado a esses biomarcadores, especialmente em PBMR. Em nosso estudo, determinantes sociais individuais foram utilizados, mas n&#227;o selecionados a partir do algoritmo baseado em ML. Por outro lado, demonstramos pela primeira vez que uma medida de qualidade de vida ajuda a melhorar a previs&#227;o de morte hospitalar de pacientes com IC em um modelo baseado em ML.</p>
      <p>Nosso estudo tem pontos fortes que devem ser destacados. Primeiramente, analisamos os determinantes sociais da sa&#250;de individualmente, embora n&#227;o tenham sido selecionados como vari&#225;veis importantes. Esse achado est&#225; em linha com a literatura atual, que afirma que os determinantes sociais podem ser relevantes apenas em subgrupos mais homog&#234;neos, o que pode n&#227;o ser o caso da popula&#231;&#227;o brasileira.<sup><xref ref-type="bibr" rid="B4">4</xref></sup> Em segundo lugar, tamb&#233;m realizamos valida&#231;&#227;o externa usando uma coorte diferente daquela usada para o treinamento do modelo, o que deve ser incentivado em estudos que utilizam ML.<sup><xref ref-type="bibr" rid="B7">7</xref></sup> Em terceiro lugar, nossa amostra cont&#233;m pacientes de todas as macrorregi&#245;es do pa&#237;s, abrangendo uma ampla gama de caracter&#237;sticas demogr&#225;ficas e status sociais. Essa diversidade aumenta a generaliza&#231;&#227;o e a valida&#231;&#227;o de nossos achados em diferentes popula&#231;&#245;es. Em quarto lugar, nossos dados foram coletados sistematicamente como parte de um estudo de coorte prospectivo, em vez de depender de RME. que s&#227;o mais suscet&#237;veis ao vi&#233;s de classifica&#231;&#227;o incorreta.<sup><xref ref-type="bibr" rid="B22">22</xref>-<xref ref-type="bibr" rid="B26">26</xref></sup></p>
      <p>Este estudo tamb&#233;m apresenta limita&#231;&#245;es que devem ser consideradas. Dados ausentes levaram a uma redu&#231;&#227;o no tamanho da amostra inicial, diminuindo-a para aproximadamente 25% do tamanho original, pois removemos pacientes com dados ausentes em qualquer uma das vari&#225;veis mais significativas. Inclu&#237;mos apenas casos completos, assumindo que os dados estavam completamente ausentes aleatoriamente. Optamos por n&#227;o imputar m&#250;ltiplas vari&#225;veis para evitar a introdu&#231;&#227;o de vi&#233;s potencial de estimativas baseadas em modelos em um ambiente preditivo. Al&#233;m disso, os pept&#237;deos natriur&#233;ticos n&#227;o foram utilizados para construir o modelo devido &#224; falta de disponibilidade na maioria dos centros, por se tratar de um banco de dados em que a grande maioria dos locais &#233; financiada publicamente. Reconhecemos que a aus&#234;ncia desses marcadores pode prejudicar a an&#225;lise, visto que atualmente s&#227;o testes essenciais para a confirma&#231;&#227;o do diagn&#243;stico de IC. Em pa&#237;ses com maior disponibilidade, acreditamos que essa vari&#225;vel pode ser ainda mais relevante e aumentar a precis&#227;o do algoritmo. Por outro lado, isso poderia melhorar a aplicabilidade do escore ML em outros PBMR, onde o acesso ao BNP tamb&#233;m pode ser limitado.</p>
      <p>Apesar dos resultados promissores, a implementa&#231;&#227;o de escores de risco baseados em ML na pr&#225;tica cl&#237;nica continua desafiadora. Ao contr&#225;rio dos escores de risco tradicionais, que podem usar tabelas ou calculadoras simples, os modelos de ML exigem infraestrutura digital para entrada de dados, processamento e integra&#231;&#227;o de resultados em fluxos de trabalho cl&#237;nicos. Em nosso estudo, o desenvolvimento e a valida&#231;&#227;o do modelo foram realizados em um computador centralizado contendo dados de todos os centros participantes, e a an&#225;lise foi computacionalmente intensiva. Para uso no mundo real, a implanta&#231;&#227;o exigiria sistemas compat&#237;veis e entrada de dados padronizada. Estrat&#233;gias como o desenvolvimento de aplicativos m&#243;veis offline ou a integra&#231;&#227;o do modelo em prontu&#225;rios eletr&#244;nicos de sa&#250;de de baixa complexidade poderiam ser exploradas para facilitar a implementa&#231;&#227;o em ambientes com recursos limitados.</p>
      <p>Embora o modelo atual possa ser aplicado sem retreinamento, os modelos baseados em ML podem ser atualizados ou retreinados periodicamente para melhorar o desempenho e se adaptar &#224; evolu&#231;&#227;o das popula&#231;&#245;es cl&#237;nicas. Isso levanta importantes considera&#231;&#245;es &#233;ticas e regulat&#243;rias, incluindo privacidade de dados, consentimento informado para uso secund&#225;rio de dados de pacientes e transpar&#234;ncia em rela&#231;&#227;o aos procedimentos de atualiza&#231;&#227;o do modelo. A implementa&#231;&#227;o eficaz depender&#225; n&#227;o apenas de uma infraestrutura t&#233;cnica adequada para atender &#224;s demandas computacionais, mas tamb&#233;m de fortes estruturas de supervis&#227;o e governan&#231;a para garantir o uso seguro, equitativo e respons&#225;vel.</p>
    </sec>
    <sec sec-type="conclusions">
      <title>Conclus&#227;o</title>
      <p>N&#243;s desenvolvemos e validamos um modelo baseado em aprendizado de m&#225;quina - o escore ML-HF - para prever o risco de morte hospitalar em pacientes com IC aguda. Essa ferramenta preditiva apresentou poder discriminativo superior aos escores tradicionais ADHERE e GWTG.</p>
    </sec>
    <sec sec-type="supplementary-material">
      <title>*Material suplementar</title>
      <supplementary-material id="suppl01">
        <p>Para informa&#231;&#227;o adicional, por favor, clique aqui.<ext-link ext-link-type="uri" xlink:href="https://abccardiol.org/supplementary-material/2025/12211/2025-0136_supplementary_material.pdf">https://abccardiol.org/supplementary-material/2025/12211/2025-0136_supplementary_material.pdf</ext-link> </p>
        <media mime-subtype="pdf" mimetype="application" xlink:href="https://minio.scielo.br/documentstore/1678-4170/XVyRRmkytVYdx8hM6nd7qwM/3ed1c116bd46f6527e95586bd4b3302acc3a5505.pdf"/>
      </supplementary-material>
    </sec>
  </body>
  <back>
    <ack>
      <title>Agradecimentos</title>
      <p>Os autores gostariam de expressar sua gratid&#227;o aos colegas participantes que apoiaram o recrutamento para o programa BPC.</p>
    </ack>
    <ref-list>
      <title>Refer&#234;ncias</title>
      <ref id="B1">
        <label>1</label>
        <mixed-citation>1. Albuquerque DC, Barros e Silva PGM, Lopes RD, Hoffmann-Filho CR, Nogueira PR, Reis H, et al. In-Hospital Management and Long-Term Clinical Outcomes and Adherence in Patients with Acute Decompensated Heart Failure: Primary Results of the First Brazilian Registry of Heart Failure (BREATHE). J Card Fail. 2024;30(5):639-50. doi: 10.1016/j.cardfail.2023.08.014.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Albuquerque</surname>
              <given-names>DC</given-names>
            </name>
            <name>
              <surname>Barros e Silva</surname>
              <given-names>PGM</given-names>
            </name>
            <name>
              <surname>Lopes</surname>
              <given-names>RD</given-names>
            </name>
            <name>
              <surname>Hoffmann-Filho</surname>
              <given-names>CR</given-names>
            </name>
            <name>
              <surname>Nogueira</surname>
              <given-names>PR</given-names>
            </name>
            <name>
              <surname>Reis</surname>
              <given-names>H</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>In-Hospital Management and Long-Term Clinical Outcomes and Adherence in Patients with Acute Decompensated Heart Failure: Primary Results of the First Brazilian Registry of Heart Failure (BREATHE)</article-title>
          <source>J Card Fail</source>
          <year>2024</year>
          <volume>30</volume>
          <issue>5</issue>
          <fpage>639</fpage>
          <lpage>650</lpage>
          <pub-id pub-id-type="doi">10.1016/j.cardfail.2023.08.014</pub-id>
        </element-citation>
      </ref>
      <ref id="B2">
        <label>2</label>
        <mixed-citation>2. Royston P, Altman DG. Risk Stratification for in-Hospital Mortality in Acutely Decompensated Heart Failure. JAMA. 2005;293(20):2467-8. doi: 10.1001/jama.293.20.2467-c.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Royston</surname>
              <given-names>P</given-names>
            </name>
            <name>
              <surname>Altman</surname>
              <given-names>DG</given-names>
            </name>
          </person-group>
          <article-title>Risk Stratification for in-Hospital Mortality in Acutely Decompensated Heart Failure</article-title>
          <source>JAMA</source>
          <year>2005</year>
          <volume>293</volume>
          <issue>20</issue>
          <fpage>2467</fpage>
          <lpage>2468</lpage>
          <pub-id pub-id-type="doi">10.1001/jama.293.20.2467-c</pub-id>
        </element-citation>
      </ref>
      <ref id="B3">
        <label>3</label>
        <mixed-citation>3. Peterson PN, Rumsfeld JS, Liang L, Albert NM, Hernandez AF, Peterson ED, et al. A Validated Risk Score for in-Hospital Mortality in Patients with Heart Failure from the American Heart Association Get with the Guidelines Program. Circ Cardiovasc Qual Outcomes. 2010;3(1):25-32. doi: 10.1161/CIRCOUTCOMES.109.854877.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Peterson</surname>
              <given-names>PN</given-names>
            </name>
            <name>
              <surname>Rumsfeld</surname>
              <given-names>JS</given-names>
            </name>
            <name>
              <surname>Liang</surname>
              <given-names>L</given-names>
            </name>
            <name>
              <surname>Albert</surname>
              <given-names>NM</given-names>
            </name>
            <name>
              <surname>Hernandez</surname>
              <given-names>AF</given-names>
            </name>
            <name>
              <surname>Peterson</surname>
              <given-names>ED</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>A Validated Risk Score for in-Hospital Mortality in Patients with Heart Failure from the American Heart Association Get with the Guidelines Program</article-title>
          <source>Circ Cardiovasc Qual Outcomes</source>
          <year>2010</year>
          <volume>3</volume>
          <issue>1</issue>
          <fpage>25</fpage>
          <lpage>32</lpage>
          <pub-id pub-id-type="doi">10.1161/CIRCOUTCOMES.109.854877</pub-id>
        </element-citation>
      </ref>
      <ref id="B4">
        <label>4</label>
        <mixed-citation>4. Li Y, Wang H, Luo Y. Improving Fairness in the Prediction of Heart Failure Length of Stay and Mortality by Integrating Social Determinants of Health. Circ Heart Fail. 2022;15(11):e009473. doi: 10.1161/CIRCHEARTFAILURE.122.009473.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Li</surname>
              <given-names>Y</given-names>
            </name>
            <name>
              <surname>Wang</surname>
              <given-names>H</given-names>
            </name>
            <name>
              <surname>Luo</surname>
              <given-names>Y</given-names>
            </name>
          </person-group>
          <article-title>Improving Fairness in the Prediction of Heart Failure Length of Stay and Mortality by Integrating Social Determinants of Health</article-title>
          <source>Circ Heart Fail</source>
          <year>2022</year>
          <volume>15</volume>
          <issue>11</issue>
          <elocation-id>e009473</elocation-id>
          <pub-id pub-id-type="doi">10.1161/CIRCHEARTFAILURE.122.009473</pub-id>
        </element-citation>
      </ref>
      <ref id="B5">
        <label>5</label>
        <mixed-citation>5. Segar MW, Hall JL, Jhund PS, Powell-Wiley TM, Morris AA, Kao D, et al. Machine Learning-Based Models Incorporating Social Determinants of Health vs Traditional Models for Predicting in-Hospital Mortality in Patients with Heart Failure. JAMA Cardiol. 2022;7(8):844-54. doi: 10.1001/jamacardio.2022.1900.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Segar</surname>
              <given-names>MW</given-names>
            </name>
            <name>
              <surname>Hall</surname>
              <given-names>JL</given-names>
            </name>
            <name>
              <surname>Jhund</surname>
              <given-names>PS</given-names>
            </name>
            <name>
              <surname>Powell-Wiley</surname>
              <given-names>TM</given-names>
            </name>
            <name>
              <surname>Morris</surname>
              <given-names>AA</given-names>
            </name>
            <name>
              <surname>Kao</surname>
              <given-names>D</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Machine Learning-Based Models Incorporating Social Determinants of Health vs Traditional Models for Predicting in-Hospital Mortality in Patients with Heart Failure</article-title>
          <source>JAMA Cardiol</source>
          <year>2022</year>
          <volume>7</volume>
          <issue>8</issue>
          <fpage>844</fpage>
          <lpage>854</lpage>
          <pub-id pub-id-type="doi">10.1001/jamacardio.2022.1900</pub-id>
        </element-citation>
      </ref>
      <ref id="B6">
        <label>6</label>
        <mixed-citation>6. Deo RC. Machine Learning in Medicine. Circulation. 2015;132(20):1920-30. doi: 10.1161/CIRCULATIONAHA.115.001593.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Deo</surname>
              <given-names>RC</given-names>
            </name>
          </person-group>
          <article-title>Machine Learning in Medicine</article-title>
          <source>Circulation</source>
          <year>2015</year>
          <volume>132</volume>
          <issue>20</issue>
          <fpage>1920</fpage>
          <lpage>1930</lpage>
          <pub-id pub-id-type="doi">10.1161/CIRCULATIONAHA.115.001593</pub-id>
        </element-citation>
      </ref>
      <ref id="B7">
        <label>7</label>
        <mixed-citation>7. Shin S, Austin PC, Ross HJ, Abdel-Qadir H, Freitas C, Tomlinson G, et al. Machine Learning vs. Conventional Statistical Models for Predicting Heart Failure Readmission and Mortality. ESC Heart Fail. 2021;8(1):106-15. doi: 10.1002/ehf2.13073.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Shin</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Austin</surname>
              <given-names>PC</given-names>
            </name>
            <name>
              <surname>Ross</surname>
              <given-names>HJ</given-names>
            </name>
            <name>
              <surname>Abdel-Qadir</surname>
              <given-names>H</given-names>
            </name>
            <name>
              <surname>Freitas</surname>
              <given-names>C</given-names>
            </name>
            <name>
              <surname>Tomlinson</surname>
              <given-names>G</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Machine Learning vs. Conventional Statistical Models for Predicting Heart Failure Readmission and Mortality</article-title>
          <source>ESC Heart Fail</source>
          <year>2021</year>
          <volume>8</volume>
          <issue>1</issue>
          <fpage>106</fpage>
          <lpage>115</lpage>
          <pub-id pub-id-type="doi">10.1002/ehf2.13073</pub-id>
        </element-citation>
      </ref>
      <ref id="B8">
        <label>8</label>
        <mixed-citation>8. Angraal S, Mortazavi BJ, Gupta A, Khera R, Ahmad T, Desai NR, et al. Machine Learning Prediction of Mortality and Hospitalization in Heart Failure with Preserved Ejection Fraction. JACC Heart Fail. 2020;8(1):12-21. doi: 10.1016/j.jchf.2019.06.013.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Angraal</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Mortazavi</surname>
              <given-names>BJ</given-names>
            </name>
            <name>
              <surname>Gupta</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Khera</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Ahmad</surname>
              <given-names>T</given-names>
            </name>
            <name>
              <surname>Desai</surname>
              <given-names>NR</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Machine Learning Prediction of Mortality and Hospitalization in Heart Failure with Preserved Ejection Fraction</article-title>
          <source>JACC Heart Fail</source>
          <year>2020</year>
          <volume>8</volume>
          <issue>1</issue>
          <fpage>12</fpage>
          <lpage>21</lpage>
          <pub-id pub-id-type="doi">10.1016/j.jchf.2019.06.013</pub-id>
        </element-citation>
      </ref>
      <ref id="B9">
        <label>9</label>
        <mixed-citation>9. Adler ED, Voors AA, Klein L, Macheret F, Braun OO, Urey MA, et al. Improving Risk Prediction in Heart Failure Using Machine Learning. Eur J Heart Fail. 2020;22(1):139-47. doi: 10.1002/ejhf.1628. Epub 2019 Nov 12.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Adler</surname>
              <given-names>ED</given-names>
            </name>
            <name>
              <surname>Voors</surname>
              <given-names>AA</given-names>
            </name>
            <name>
              <surname>Klein</surname>
              <given-names>L</given-names>
            </name>
            <name>
              <surname>Macheret</surname>
              <given-names>F</given-names>
            </name>
            <name>
              <surname>Braun</surname>
              <given-names>OO</given-names>
            </name>
            <name>
              <surname>Urey</surname>
              <given-names>MA</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Improving Risk Prediction in Heart Failure Using Machine Learning</article-title>
          <source>Eur J Heart Fail</source>
          <year>2020</year>
          <volume>22</volume>
          <issue>1</issue>
          <fpage>139</fpage>
          <lpage>147</lpage>
          <pub-id pub-id-type="doi">10.1002/ejhf.1628</pub-id>
          <comment>Epub 2019 Nov 12</comment>
        </element-citation>
      </ref>
      <ref id="B10">
        <label>10</label>
        <mixed-citation>10. Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, et al. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J Am Coll Cardiol. 2022;79(17):e263-e421. doi: 10.1016/j.jacc.2021.12.012.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Heidenreich</surname>
              <given-names>PA</given-names>
            </name>
            <name>
              <surname>Bozkurt</surname>
              <given-names>B</given-names>
            </name>
            <name>
              <surname>Aguilar</surname>
              <given-names>D</given-names>
            </name>
            <name>
              <surname>Allen</surname>
              <given-names>LA</given-names>
            </name>
            <name>
              <surname>Byun</surname>
              <given-names>JJ</given-names>
            </name>
            <name>
              <surname>Colvin</surname>
              <given-names>MM</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines</article-title>
          <source>J Am Coll Cardiol</source>
          <year>2022</year>
          <volume>79</volume>
          <issue>17</issue>
          <fpage>e263</fpage>
          <lpage>e421</lpage>
          <pub-id pub-id-type="doi">10.1016/j.jacc.2021.12.012</pub-id>
        </element-citation>
      </ref>
      <ref id="B11">
        <label>11</label>
        <mixed-citation>11. Taniguchi FP, Bernardez-Pereira S, Silva SA, Ribeiro ALP, Morgan L, Curtis AB, et al. Implementation of a Best Practice in Cardiology (BPC) Program Adapted from Get with the Guidelines&#174;in Brazilian Public Hospitals: Study Design and Rationale. Arq Bras Cardiol. 2020;115(1):92-9. doi: 10.36660/abc.20190393.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Taniguchi</surname>
              <given-names>FP</given-names>
            </name>
            <name>
              <surname>Bernardez-Pereira</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Silva</surname>
              <given-names>SA</given-names>
            </name>
            <name>
              <surname>Ribeiro</surname>
              <given-names>ALP</given-names>
            </name>
            <name>
              <surname>Morgan</surname>
              <given-names>L</given-names>
            </name>
            <name>
              <surname>Curtis</surname>
              <given-names>AB</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Implementation of a Best Practice in Cardiology (BPC) Program Adapted from Get with the Guidelines&#174;in Brazilian Public Hospitals: Study Design and Rationale</article-title>
          <source>Arq Bras Cardiol</source>
          <year>2020</year>
          <volume>115</volume>
          <issue>1</issue>
          <fpage>92</fpage>
          <lpage>99</lpage>
          <pub-id pub-id-type="doi">10.36660/abc.20190393</pub-id>
        </element-citation>
      </ref>
      <ref id="B12">
        <label>12</label>
        <mixed-citation>12. Taniguchi FP, Bernardez-Pereira S, Ribeiro ALP, Morgan L, Curtis AB, Taubert K, et al. A Nationwide Initiative to Improve Cardiology Quality: The Best Practice in Cardiology Program in Brazil. Arq Bras Cardiol. 2023;120(10):e20230375. doi: 10.36660/abc.20230375.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Taniguchi</surname>
              <given-names>FP</given-names>
            </name>
            <name>
              <surname>Bernardez-Pereira</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Ribeiro</surname>
              <given-names>ALP</given-names>
            </name>
            <name>
              <surname>Morgan</surname>
              <given-names>L</given-names>
            </name>
            <name>
              <surname>Curtis</surname>
              <given-names>AB</given-names>
            </name>
            <name>
              <surname>Taubert</surname>
              <given-names>K</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>A Nationwide Initiative to Improve Cardiology Quality: The Best Practice in Cardiology Program in Brazil</article-title>
          <source>Arq Bras Cardiol</source>
          <year>2023</year>
          <volume>120</volume>
          <issue>10</issue>
          <elocation-id>e20230375</elocation-id>
          <pub-id pub-id-type="doi">10.36660/abc.20230375</pub-id>
        </element-citation>
      </ref>
      <ref id="B13">
        <label>13</label>
        <mixed-citation>13. Haugen AS, Sevdalis N, S&#248;fteland E. Impact of the World Health Organization Surgical Safety Checklist on Patient Safety. Anesthesiology. 2019;131(2):420-5. doi: 10.1097/ALN.0000000000002674.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Haugen</surname>
              <given-names>AS</given-names>
            </name>
            <name>
              <surname>Sevdalis</surname>
              <given-names>N</given-names>
            </name>
            <name>
              <surname>S&#248;fteland</surname>
              <given-names>E</given-names>
            </name>
          </person-group>
          <article-title>Impact of the World Health Organization Surgical Safety Checklist on Patient Safety</article-title>
          <source>Anesthesiology</source>
          <year>2019</year>
          <volume>131</volume>
          <issue>2</issue>
          <fpage>420</fpage>
          <lpage>425</lpage>
          <pub-id pub-id-type="doi">10.1097/ALN.0000000000002674</pub-id>
        </element-citation>
      </ref>
      <ref id="B14">
        <label>14</label>
        <mixed-citation>14. Jakobsen JC, Gluud C, Wetterslev J, Winkel P. When and How Should Multiple Imputation be Used for Handling Missing Data in Randomised Clinical Trials - A Practical Guide with Flowcharts. BMC Med Res Methodol. 2017;17(1):162. doi: 10.1186/s12874-017-0442-1.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Jakobsen</surname>
              <given-names>JC</given-names>
            </name>
            <name>
              <surname>Gluud</surname>
              <given-names>C</given-names>
            </name>
            <name>
              <surname>Wetterslev</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Winkel</surname>
              <given-names>P</given-names>
            </name>
          </person-group>
          <article-title>When and How Should Multiple Imputation be Used for Handling Missing Data in Randomised Clinical Trials - A Practical Guide with Flowcharts</article-title>
          <source>BMC Med Res Methodol</source>
          <year>2017</year>
          <volume>17</volume>
          <issue>1</issue>
          <size units="pages">162</size>
          <pub-id pub-id-type="doi">10.1186/s12874-017-0442-1</pub-id>
        </element-citation>
      </ref>
      <ref id="B15">
        <label>15</label>
        <mixed-citation>15. Kursa MB, Rudnicki WR. Feature Selection with the Boruta Package. J Stat Softw. 2010;36(11):1-13. doi: 10.18637/jss.v036.i11.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Kursa</surname>
              <given-names>MB</given-names>
            </name>
            <name>
              <surname>Rudnicki</surname>
              <given-names>WR</given-names>
            </name>
          </person-group>
          <article-title>Feature Selection with the Boruta Package</article-title>
          <source>J Stat Softw</source>
          <year>2010</year>
          <volume>36</volume>
          <issue>11</issue>
          <fpage>1</fpage>
          <lpage>13</lpage>
          <pub-id pub-id-type="doi">10.18637/jss.v036.i11</pub-id>
        </element-citation>
      </ref>
      <ref id="B16">
        <label>16</label>
        <mixed-citation>16. Kowarik A, Templ M. Imputation with the R Package VIM. J Stat Softw. 2016;74(7):1-16. doi: 10.18637/jss.v074.i07.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Kowarik</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Templ</surname>
              <given-names>M</given-names>
            </name>
          </person-group>
          <article-title>Imputation with the R Package VIM</article-title>
          <source>J Stat Softw</source>
          <year>2016</year>
          <volume>74</volume>
          <issue>7</issue>
          <fpage>1</fpage>
          <lpage>16</lpage>
          <pub-id pub-id-type="doi">10.18637/jss.v074.i07</pub-id>
        </element-citation>
      </ref>
      <ref id="B17">
        <label>17</label>
        <mixed-citation>17. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the Areas Under Two or More Correlated Receiver Operating Characteristic Curves: A Nonparametric Approach. Biometrics. 1988;44(3):837-45.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>DeLong</surname>
              <given-names>ER</given-names>
            </name>
            <name>
              <surname>DeLong</surname>
              <given-names>DM</given-names>
            </name>
            <name>
              <surname>Clarke-Pearson</surname>
              <given-names>DL</given-names>
            </name>
          </person-group>
          <article-title>Comparing the Areas Under Two or More Correlated Receiver Operating Characteristic Curves: A Nonparametric Approach</article-title>
          <source>Biometrics</source>
          <year>1988</year>
          <volume>44</volume>
          <issue>3</issue>
          <fpage>837</fpage>
          <lpage>845</lpage>
        </element-citation>
      </ref>
      <ref id="B18">
        <label>18</label>
        <mixed-citation>18. Collins GS, Reitsma JB, Altman DG, Moons KGM; members of the TRIPOD group. Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD): The TRIPOD Statement. Eur Urol. 2015;67(6):1142-51. doi: 10.1016/j.eururo.2014.11.025.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Collins</surname>
              <given-names>GS</given-names>
            </name>
            <name>
              <surname>Reitsma</surname>
              <given-names>JB</given-names>
            </name>
            <name>
              <surname>Altman</surname>
              <given-names>DG</given-names>
            </name>
            <name>
              <surname>Moons</surname>
              <given-names>KGM</given-names>
            </name>
            <collab>members of the TRIPOD group</collab>
          </person-group>
          <article-title>Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD): The TRIPOD Statement</article-title>
          <source>Eur Urol</source>
          <year>2015</year>
          <volume>67</volume>
          <issue>6</issue>
          <fpage>1142</fpage>
          <lpage>1151</lpage>
          <pub-id pub-id-type="doi">10.1016/j.eururo.2014.11.025</pub-id>
        </element-citation>
      </ref>
      <ref id="B19">
        <label>19</label>
        <mixed-citation>19. Greene SJ, Butler J, Spertus JA, Hellkamp AS, Vaduganathan M, DeVore AD, et al. Comparison of New York Heart Association Class and Patient-Reported Outcomes for Heart Failure with Reduced Ejection Fraction. JAMA Cardiol. 2021;6(5):522-31. doi: 10.1001/jamacardio.2021.0372.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Greene</surname>
              <given-names>SJ</given-names>
            </name>
            <name>
              <surname>Butler</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Spertus</surname>
              <given-names>JA</given-names>
            </name>
            <name>
              <surname>Hellkamp</surname>
              <given-names>AS</given-names>
            </name>
            <name>
              <surname>Vaduganathan</surname>
              <given-names>M</given-names>
            </name>
            <name>
              <surname>DeVore</surname>
              <given-names>AD</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Comparison of New York Heart Association Class and Patient-Reported Outcomes for Heart Failure with Reduced Ejection Fraction</article-title>
          <source>JAMA Cardiol</source>
          <year>2021</year>
          <volume>6</volume>
          <issue>5</issue>
          <fpage>522</fpage>
          <lpage>531</lpage>
          <pub-id pub-id-type="doi">10.1001/jamacardio.2021.0372</pub-id>
        </element-citation>
      </ref>
      <ref id="B20">
        <label>20</label>
        <mixed-citation>20. Hawwa N, Vest AR, Kumar R, Lahoud R, Young JB, Wu Y, et al. Comparison between the Kansas City Cardiomyopathy Questionnaire and New York Heart Association in Assessing Functional Capacity and Clinical Outcomes. J Card Fail. 2017;23(4):280-5. doi: 10.1016/j.cardfail.2016.12.002.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Hawwa</surname>
              <given-names>N</given-names>
            </name>
            <name>
              <surname>Vest</surname>
              <given-names>AR</given-names>
            </name>
            <name>
              <surname>Kumar</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Lahoud</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Young</surname>
              <given-names>JB</given-names>
            </name>
            <name>
              <surname>Wu</surname>
              <given-names>Y</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Comparison between the Kansas City Cardiomyopathy Questionnaire and New York Heart Association in Assessing Functional Capacity and Clinical Outcomes</article-title>
          <source>J Card Fail</source>
          <year>2017</year>
          <volume>23</volume>
          <issue>4</issue>
          <fpage>280</fpage>
          <lpage>285</lpage>
          <pub-id pub-id-type="doi">10.1016/j.cardfail.2016.12.002</pub-id>
        </element-citation>
      </ref>
      <ref id="B21">
        <label>21</label>
        <mixed-citation>21. Coffey S, Cairns BJ, Iung B. The Modern Epidemiology of Heart Valve Disease. Heart. 2016;102(1):75-85. doi: 10.1136/heartjnl-2014-307020.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Coffey</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Cairns</surname>
              <given-names>BJ</given-names>
            </name>
            <name>
              <surname>Iung</surname>
              <given-names>B</given-names>
            </name>
          </person-group>
          <article-title>The Modern Epidemiology of Heart Valve Disease</article-title>
          <source>Heart</source>
          <year>2016</year>
          <volume>102</volume>
          <issue>1</issue>
          <fpage>75</fpage>
          <lpage>85</lpage>
          <pub-id pub-id-type="doi">10.1136/heartjnl-2014-307020</pub-id>
        </element-citation>
      </ref>
      <ref id="B22">
        <label>22</label>
        <mixed-citation>22. Panahiazar M, Taslimitehrani V, Pereira N, Pathak J. Using EHRs and Machine Learning for Heart Failure Survival Analysis. Stud Health Technol Inform. 2015;216:40-4.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Panahiazar</surname>
              <given-names>M</given-names>
            </name>
            <name>
              <surname>Taslimitehrani</surname>
              <given-names>V</given-names>
            </name>
            <name>
              <surname>Pereira</surname>
              <given-names>N</given-names>
            </name>
            <name>
              <surname>Pathak</surname>
              <given-names>J</given-names>
            </name>
          </person-group>
          <article-title>Using EHRs and Machine Learning for Heart Failure Survival Analysis</article-title>
          <source>Stud Health Technol Inform</source>
          <year>2015</year>
          <volume>216</volume>
          <fpage>40</fpage>
          <lpage>44</lpage>
        </element-citation>
      </ref>
      <ref id="B23">
        <label>23</label>
        <mixed-citation>23. Desai RJ, Wang SV, Vaduganathan M, Evers T, Schneeweiss S. Comparison of Machine Learning Methods with Traditional Models for Use of Administrative Claims with Electronic Medical Records to Predict Heart Failure Outcomes. JAMA Netw Open. 2020;3(1):e1918962. doi: 10.1001/jamanetworkopen.2019.18962.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Desai</surname>
              <given-names>RJ</given-names>
            </name>
            <name>
              <surname>Wang</surname>
              <given-names>SV</given-names>
            </name>
            <name>
              <surname>Vaduganathan</surname>
              <given-names>M</given-names>
            </name>
            <name>
              <surname>Evers</surname>
              <given-names>T</given-names>
            </name>
            <name>
              <surname>Schneeweiss</surname>
              <given-names>S</given-names>
            </name>
          </person-group>
          <article-title>Comparison of Machine Learning Methods with Traditional Models for Use of Administrative Claims with Electronic Medical Records to Predict Heart Failure Outcomes</article-title>
          <source>JAMA Netw Open</source>
          <year>2020</year>
          <volume>3</volume>
          <issue>1</issue>
          <elocation-id>e1918962</elocation-id>
          <pub-id pub-id-type="doi">10.1001/jamanetworkopen.2019.18962</pub-id>
        </element-citation>
      </ref>
      <ref id="B24">
        <label>24</label>
        <mixed-citation>24. Negassa A, Ahmed S, Zolty R, Patel SR. Prediction Model Using Machine Learning for Mortality in Patients with Heart Failure. Am J Cardiol. 2021;153:86-93. doi: 10.1016/j.amjcard.2021.05.044.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Negassa</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Ahmed</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Zolty</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Patel</surname>
              <given-names>SR</given-names>
            </name>
          </person-group>
          <article-title>Prediction Model Using Machine Learning for Mortality in Patients with Heart Failure</article-title>
          <source>Am J Cardiol</source>
          <year>2021</year>
          <volume>153</volume>
          <fpage>86</fpage>
          <lpage>93</lpage>
          <pub-id pub-id-type="doi">10.1016/j.amjcard.2021.05.044</pub-id>
        </element-citation>
      </ref>
      <ref id="B25">
        <label>25</label>
        <mixed-citation>25. Lv H, Yang X, Wang B, Wang S, Du X, Tan Q, et al. Machine Learning-Driven Models to Predict Prognostic Outcomes in Patients Hospitalized with Heart Failure Using Electronic Health Records: Retrospective Study. J Med Internet Res. 2021;23(4):e24996. doi: 10.2196/24996.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Lv</surname>
              <given-names>H</given-names>
            </name>
            <name>
              <surname>Yang</surname>
              <given-names>X</given-names>
            </name>
            <name>
              <surname>Wang</surname>
              <given-names>B</given-names>
            </name>
            <name>
              <surname>Wang</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Du</surname>
              <given-names>X</given-names>
            </name>
            <name>
              <surname>Tan</surname>
              <given-names>Q</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Machine Learning-Driven Models to Predict Prognostic Outcomes in Patients Hospitalized with Heart Failure Using Electronic Health Records: Retrospective Study</article-title>
          <source>J Med Internet Res</source>
          <year>2021</year>
          <volume>23</volume>
          <issue>4</issue>
          <elocation-id>e24996</elocation-id>
          <pub-id pub-id-type="doi">10.2196/24996</pub-id>
        </element-citation>
      </ref>
      <ref id="B26">
        <label>26</label>
        <mixed-citation>26. Barbalho IMP, Fernandes F, Barros DMS, Paiva JC, Henriques J, Morais AHF, et al. Electronic Health Records in Brazil: Prospects and Technological Challenges. Front Public Health. 2022;10:963841. doi: 10.3389/fpubh.2022.963841.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Barbalho</surname>
              <given-names>IMP</given-names>
            </name>
            <name>
              <surname>Fernandes</surname>
              <given-names>F</given-names>
            </name>
            <name>
              <surname>Barros</surname>
              <given-names>DMS</given-names>
            </name>
            <name>
              <surname>Paiva</surname>
              <given-names>JC</given-names>
            </name>
            <name>
              <surname>Henriques</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Morais</surname>
              <given-names>AHF</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Electronic Health Records in Brazil: Prospects and Technological Challenges</article-title>
          <source>Front Public Health</source>
          <year>2022</year>
          <volume>10</volume>
          <size units="pages">963841</size>
          <pub-id pub-id-type="doi">10.3389/fpubh.2022.963841</pub-id>
        </element-citation>
      </ref>
    </ref-list>
    <fn-group>
      <fn fn-type="other">
        <label>Vincula&#231;&#227;o acad&#234;mica:</label>
        <p> Este artigo &#233; parte de disserta&#231;&#227;o de mestrado de Gustavo Sarot Pereira pela Universidade Federal do Paran&#225;.</p>
      </fn>
      <fn fn-type="other">
        <label>Aprova&#231;&#227;o &#233;tica e consentimento informado:</label>
        <p> Este estudo foi aprovado pelo Comit&#234; de &#201;tica do Hospital do Cora&#231;&#227;o - HCOR / Associa&#231;&#227;o Beneficiente S&#237;ria sob o n&#250;mero de protocolo 48561715.5.1001.0060. Todos os procedimentos envolvidos nesse estudo est&#227;o de acordo com a Declara&#231;&#227;o de Helsinki de 1975, atualizada em 2013. O consentimento informado foi obtido de todos os participantes inclu&#237;dos no estudo.</p>
      </fn>
      <fn fn-type="other">
        <label>Uso de Intelig&#234;ncia Artificial:</label>
        <p>Os autores n&#227;o utilizaram ferramentas de intelig&#234;ncia artificial no desenvolvimento deste trabalho.</p>
      </fn>
      <fn fn-type="data-availability" specific-use="data-available-upon-request">
        <label>Disponibilidade de Dados:</label>
        <p> Os dados estar&#227;o dispon&#237;veis sob demanda aos autores.</p>
      </fn>
      <fn fn-type="supported-by">
        <label>Fontes de financiamento:</label>
        <p> O presente estudo foi financiado por American Heart Association, Hospital do Cora&#231;&#227;o (HCor) e Minist&#233;rio da Sa&#250;de.</p>
      </fn>
    </fn-group>
  </back>
  <sub-article article-type="translation" id="TRen" xml:lang="en">
    <front-stub>
      <article-id pub-id-type="doi">10.36660/abc.20250136i</article-id>
      <article-categories>
        <subj-group subj-group-type="heading">
          <subject>Original Article</subject>
        </subj-group>
      </article-categories>
      <title-group>
        <article-title>A New Risk Score Based on Machine Learning in Patients with Acute Heart Failure: The ML-HF score</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0009-0003-4069-8781</contrib-id>
          <name>
            <surname>Ferreira</surname>
            <given-names>Matheus Bissa Duarte</given-names>
          </name>
          <role>Conception and design of the research</role>
          <role>Acquisition of data</role>
          <role>Analysis and interpretation of the data</role>
          <role>Statistical analysis</role>
          <role>Writing of the manuscript</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0002-2385-0361</contrib-id>
          <name>
            <surname>Daikubara</surname>
            <given-names>Jorge Tadashi</given-names>
            <suffix>Neto</suffix>
          </name>
          <role>Conception and design of the research</role>
          <role>Acquisition of data</role>
          <role>Analysis and interpretation of the data</role>
          <role>Statistical analysis</role>
          <role>Writing of the manuscript</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Cunha</surname>
            <given-names>Gustavo S. Pereira da</given-names>
          </name>
          <role>Acquisition of data</role>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Moretti</surname>
            <given-names>Rafael</given-names>
          </name>
          <role>Acquisition of data</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Reichert</surname>
            <given-names>Jessica Tamires</given-names>
          </name>
          <role>Acquisition of data</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Prado</surname>
            <given-names>Lucas M&#252;ller</given-names>
          </name>
          <role>Acquisition of data</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Cirino</surname>
            <given-names>Raphael Henrique D&#233;a</given-names>
          </name>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Smith</surname>
            <given-names>Sidney C</given-names>
            <suffix>Jr</suffix>
          </name>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff2001">
            <sup>2</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0003-0674-5997</contrib-id>
          <name>
            <surname>Taniguchi</surname>
            <given-names>F&#225;bio Papa</given-names>
          </name>
          <role>Conception and design of the research</role>
          <role>Obtaining financing</role>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff3001">
            <sup>3</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Sposito</surname>
            <given-names>Andrei C.</given-names>
          </name>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff4001">
            <sup>4</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Silvestre</surname>
            <given-names>Odilson M.</given-names>
          </name>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff5001">
            <sup>5</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Nadruz</surname>
            <given-names>Wilson</given-names>
          </name>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff5001">
            <sup>5</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0002-0095-9049</contrib-id>
          <name>
            <surname>Fernandes-Silva</surname>
            <given-names>Miguel Morita</given-names>
          </name>
          <role>Conception and design of the research</role>
          <role>Analysis and interpretation of the data</role>
          <role>Writing of the manuscript</role>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
      </contrib-group>
      <aff id="aff1001">
        <label>1</label>
        <country country="BR">Brazil </country>
        <institution content-type="original">Universidade Federal do Paran&#225;, Curitiba, PR &#8211; Brazil </institution>
      </aff>
      <aff id="aff2001">
        <label>2</label>
        <country country="US">USA </country>
        <institution content-type="original">University of North Carolina At Chapel Hill, Chapel Hill, North Carolina &#8211; USA </institution>
      </aff>
      <aff id="aff3001">
        <label>3</label>
        <country country="BR">Brazil </country>
        <institution content-type="original">Hospital do Cora&#231;&#227;o (HCor), S&#227;o Paulo, SP &#8211; Brazil </institution>
      </aff>
      <aff id="aff4001">
        <label>4</label>
        <country country="BR">Brazil</country>
        <institution content-type="original">Universidade Estadual de Campinas, Campinas, SP &#8211; Brazil</institution>
      </aff>
      <aff id="aff5001">
        <label>5</label>
        <country country="BR">Brazil</country>
        <institution content-type="original">Universidade Federal do Acre, Rio Branco, AC &#8211; Brazil</institution>
      </aff>
      <author-notes>
        <corresp id="c01001"><label>Mailing Address:</label> Miguel Morita Fernandes-Silva Universidade Federal do Paran&#225; &#8211; Rua General Carneiro, 181, 10&#186; andar. Postal Code 80060-900, Curitiba, PR &#8211; Brazil E-mail: miguelmorita@ufpr.br </corresp>
        <fn fn-type="coi-statement">
          <label>Potential conflict of interest:</label>
          <p>No potential conflict of interest relevant to this article was reported.</p>
        </fn>
        <fn fn-type="edited-by">
          <label>Editor responsible for the review:</label>
          <p>Gl&#225;ucia Maria Moraes de Oliveira</p>
        </fn>
      </author-notes>
      <abstract>
        <title>Abstract</title>
        <sec>
          <title>Background</title>
          <p> Conventional prognostic assessment scores often fall short in performance to predict mortality in patients with acute heart failure (AHF).</p>
        </sec>
        <sec>
          <title>Objective</title>
          <p> To develop and validate a machine learning-based prognostic score to predict in-hospital death in patients with AHF and compare its performance with the main traditional scores.</p>
        </sec>
        <sec>
          <title>Methods</title>
          <p> Patients admitted for AHF in Brazilian hospitals from the &#8220;Best Practices in Cardiology Program&#8221; from 2016 - 2022 were included. Clinical data, laboratory results, and the World Health Organization Quality of Life (WHOQOL-Bref) at hospital admission were collected. The outcome was in-hospital death. The model was trained using 70% of admissions (training set) and validated with the remaining 30% (test set). The ML-HF score area under the ROC curve (AUC) was compared with the AUC of the traditional scores Acute Decompensated Heart Failure National Registry (ADHERE) and Get With the Guidelines&#8211;Heart Failure (GWTG-HF). The level of significance was p&lt;0.05.</p>
        </sec>
        <sec>
          <title>Results</title>
          <p> One thousand and one hundred fifty-seven patients hospitalized for AHF were included. The five most important variables of the ML-HF score were: Physical Health Domain Quality (WHOQOL-BREF), serum sodium, serum urea, serum creatinine, and systolic blood pressure at hospital admission. In the test set, the ML-HF score showed an adequate model calibration (Hosmer-Lemeshow test p value=0.056) and discrimination (AUC=0.722 [CI95%, 0.661-0.783]), which was superior to GWTG-HF (AUC=0.616 [CI95%, 0.529-0.702; p=0.014]) and the ADHERE (AUC=0.601 [CI95%, 0.511-0.691; p=0.006]) scores.</p>
        </sec>
        <sec>
          <title>Conclusion</title>
          <p> We developed and validated a score using machine learning to predict in-hospital death in patients with AHF, which outperformed the traditional scores.</p>
        </sec>
      </abstract>
      <abstract abstract-type="graphical">
        <p>
          <fig id="f01001">
            <label>Central Illustration:</label>
            <caption>
              <title>A New Risk Score Based on Machine Learning in Patients with Acute Heart Failure: The ML-HF Score</title>
            </caption>
            <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf01-en.tif"/></fig>
        </p>
      </abstract>
      <kwd-group xml:lang="en">
        <title>Keywords</title>
        <kwd>Heart Failure</kwd>
        <kwd>Machine Learning</kwd>
        <kwd>Death</kwd>
      </kwd-group>
      <funding-group>
        <award-group>
          <funding-source>American Heart Association</funding-source>
        </award-group>
        <award-group>
          <funding-source>Hospital do Cora&#231;&#227;o</funding-source>
        </award-group>
        <award-group>
          <funding-source>Minist&#233;rio da Sa&#250;de</funding-source>
        </award-group>
        <funding-statement>This study was partially funded by American Heart Association, Hospital do Cora&#231;&#227;o (HCor) and Minist&#233;rio da Sa&#250;de.</funding-statement>
      </funding-group>
    </front-stub>
    <body>
      <sec sec-type="intro">
        <title>Introduction</title>
        <p>Heart failure (HF) affects 25 million people worldwide and has high mortality and hospitalization rates. In Brazil, there are approximately 250,000 hospital admissions due to HF per year, resulting in a 10.9% in-hospital mortality rate.<sup><xref ref-type="bibr" rid="B1">1</xref></sup></p>
        <p>Risk classification tools help direct resources and intensive measures to higher-risk patients. However, estimating the prognosis of patients upon admission or during hospitalization has been challenging in the HF scenario. Previous studies have developed scores, such as the Acute Decompensated Heart Failure National Registry (ADHERE) and the Get With The Guidelines-Heart Failure (GWTG - HF) scores, which can be derived from a few laboratory and clinical variables, to predict in-hospital mortality in these patients.<sup><xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref></sup> Unfortunately, these scores have sub-optimal accuracy and have been underused in clinical practice.<sup><xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref></sup></p>
        <p>An alternative to these conventional approaches are the models based on Machine Learning (ML), which have been studied in the last years for the management of HF.<sup><xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref></sup> The ability of ML algorithms to analyze in a deeper, multi-dimensional, and non-linear way a variety of variables increases their potential to improve patient stratification models.<sup><xref ref-type="bibr" rid="B7">7</xref></sup> ML-based models have been built in patients with chronic and acute HF (AHF) using electronic medical records (EMR), registries, and clinical trials.<sup><xref ref-type="bibr" rid="B8">8</xref></sup> These models appear to have superior accuracy than the traditional ADHERE and GWTG-HF scores.<sup><xref ref-type="bibr" rid="B9">9</xref></sup> However, most of these studies have been developed and validated in high-income countries. Differences in patient population, social determinants, and healthcare delivery may prevent these models from being implemented in low- and middle-income countries (LMIC). In fact, it has already been shown that the ability of ML models to predict HF outcomes significantly changes across different ethnic and socioeconomic subgroups.<sup><xref ref-type="bibr" rid="B10">10</xref></sup></p>
        <p>Therefore, we developed and validated an algorithm based on ML to predict in-hospital deaths in patients with AHF using a large cohort study of public hospitals in Brazil. We further compared the final ML-based algorithm with the ADHERE and GWTG models.</p>
      </sec>
      <sec sec-type="methods">
        <title>Methods</title>
        <sec>
          <title>Population</title>
          <p>This is an analysis of the "Best Practice in Cardiology (BPC)" program, which was adapted from the "Get With The Guidelines" (GWTG) initiative to improve the quality of cardiovascular care and patients' outcomes in acute coronary syndrome, HF, and atrial fibrillation in Brazilian public hospitals. The study design, rationale, procedures, and main results have been previously published.<sup><xref ref-type="bibr" rid="B11">11</xref></sup> Briefly, tertiary public hospitals in all five Brazilian macro-regions have been selected since January 2016.<sup><xref ref-type="bibr" rid="B12">12</xref></sup> These hospitals were included in a quasi-experimental study which implemented an institutional-level intervention to improve adherence to guidelines and evaluated its impact on measures of institutional performance. Sociodemographic, clinical, laboratory, and echocardiographic data were collected during patient admission. Patients were also asked to complete the World Health Organization Quality Of Life (WHOQOL-Bref) questionnaire, which is a 26-item instrument consisting of four domains: physical health (7 items), psychological health (6 items), social relationships (3 items), and environmental health (8 items). The items of the WHOQOL-Bref consist of 5-point Likert scales that are combined to provide each health domain. The procedure to calculate the WHOLQOL-BREF was described in detail elsewhere.<sup><xref ref-type="bibr" rid="B13">13</xref></sup> As part of the BPC program, all participants were followed up with visits at discharge, one month, and six-month. This study was approved by the ethics committee of the Coordinating Center (number 48561715.5.1001.0060) and of each participating hospital. All participants gave written informed consent for participation in the study.</p>
        </sec>
        <sec>
          <title>Inclusion and exclusion criteria</title>
          <p>For this analysis, we included patients aged 18 years or older hospitalized with a primary diagnosis of AHF (ICD-10 codes I50; I50.0; I50.1 or I50.9) between 14/03/2016 and 07/12/2022 in any participating hospital of the BPC program, who agreed to participate in the study and signed the informed consent form as illustrated in the <xref ref-type="fig" rid="f01001">Central Illustration</xref>. Participants were required to have been hospitalized for at least 24 hours to be eligible.</p>
          <p>We excluded patients transferred to another hospital or with missing vital status at discharge. Patients who were on the waiting list for a heart transplant or using a ventricular assist device were also excluded. We also excluded patients with missing data for the selected variables of the model (see variable selection below).</p>
          <p>To build and validate the model, we divided the dataset into 2 Cohorts: Cohort 1, patients admitted from 16/01/2016 to 15/10/2021, and Cohort 2, patients admitted from 16/10/2021 to 07/12/2022. Cohort 1 was used in the variable selection, model building, and validation processes. Cohort 2 was used only for external validation (see Statistical analysis for details).</p>
        </sec>
        <sec>
          <title>Outcome</title>
          <p>The outcome was death from any cause during the index hospitalization.</p>
          <sec>
            <title>Variable selection</title>
            <p>We used all variables at admission in the predictive model, which included sociodemographic, clinical, laboratory, and echocardiographic data, along with the scores derived from the responses of the WHOQOL-Bref questionnaire. First, we excluded variables with more than 40% of data missing.<sup><xref ref-type="bibr" rid="B14">14</xref></sup> See <xref ref-type="supplementary-material" rid="suppl01-en">supplemental table S1 </xref> for all variables and respective missing data rates. Then, we applied the ML method "Boruta" in the training set (see below) to automatically detect the most relevant variables that were selected for the training algorithm. The Boruta method generates a binomial probability distribution of the number of times a variable exceeds the importance of a created dummy variable, which has a random (non-predictive) value.<sup><xref ref-type="bibr" rid="B15">15</xref></sup> Variables were selected if they had a p-value &lt; 0.05 from the upper tail of the probability distribution. During this selection process, we applied a K-Nearest Neighbor imputation method, selecting the K nearest (K = 5) neighbors with complete data based on a Gower distance and inputting the mean value of the complete variable to the neighbors with missing values.<sup><xref ref-type="bibr" rid="B16">16</xref></sup></p>
            <p>After selecting the most relevant variables, they were used to build the ML algorithm in the non-imputed database. Figure S1 illustrates the variable selection process (<xref ref-type="supplementary-material" rid="suppl01-en">Supplementary Figure S1</xref>). We conducted a Complete Case Analysis, including only patients with complete data for the selected variables. Thus, patients with missing data for any of the selected variables were excluded (<xref ref-type="fig" rid="f02001">Figure 1</xref>).</p>
            <p>
              <fig id="f02001">
                <label>Figure 1</label>
                <caption>
                  <title>&#8211; Flowchart of patients included in the analysis.</title>
                </caption>
                <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf02-en.tif"/></fig>
            </p>
          </sec>
        </sec>
      </sec>
      <sec>
        <title>Statistical analysis</title>
        <sec>
          <title>Summary Statistics</title>
          <p>Continuous variables were assessed for normality through visual inspection of histograms, skewness, and kurtosis, and by applying the Kolmogorov-Smirnov test. Data are presented as mean &#177; standard deviation or median (25<sup>th</sup> percentile - 75<sup>th</sup> percentile), as appropriate. Between-group comparisons were performed using the unpaired Student's t-test for normally distributed variables and the Mann-Whitney U test for non-normally distributed variables. Categorical variables were expressed as counts and percentages and compared using Fisher's exact test.</p>
        </sec>
        <sec>
          <title>Model development and validation</title>
          <p>Cohort 1 was randomly divided into training (70%) and test (30%) sets. This procedure was performed using a computer-generated random sampling without replacement, ensuring equal probability of selection of each patient. The training set was used for model training and internal validation. The test set combined 30% of Cohort 1 with the full Cohort 2 and was used for external validation. A sensitivity analysis was performed using only Cohort 2 as the test set. This process is illustrated in <xref ref-type="fig" rid="f03001">Figure 2</xref>.</p>
          <p>
            <fig id="f03001">
              <label>Figure 2</label>
              <caption>
                <title>&#8211; Flowchart illustrating the internal and external validation of the ML-HF score model.</title>
              </caption>
              <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf03-en.tif"/></fig>
          </p>
        </sec>
        <sec>
          <title>Model selection</title>
          <p>We employed a 5-fold cross-validation method to select the most suitable ML algorithm. This method consists of subsetting the training set into 5 random groups. Four of these subsets were used to train the model, and the remaining subset to test it. It is an iterative process that occurs 5 times, changing the test subset, as illustrated in <xref ref-type="fig" rid="f03001">Figure 2</xref>. The algorithm with the best mean AUC was chosen as the final model.</p>
        </sec>
        <sec>
          <title>Model performance</title>
          <p>To evaluate model discrimination, we constructed the receiver-operating characteristic (ROC) curves and compared the area under the ROC curve (AUC) using in-hospital death as the outcome. Model calibration was evaluated by plotting the predicted versus observed probability of in-hospital death across deciles and calculating the Hosmer-Lemeshow goodness-of-fit test. A p-value &lt; 0.05 indicated inadequate fit of the data, while higher p-values indicated adequate calibration.</p>
        </sec>
        <sec>
          <title>Comparison with traditional scores</title>
          <p>The ADHERE and GWTG-HF scores were calculated for each patient. Since the GWTG-HF score does not directly output the probabilities of death, we derived a logistic regression of its numeric outputs.<sup><xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref></sup> We compared the AUC-ROC between scores using DeLong's test in both training and test sets.<sup><xref ref-type="bibr" rid="B17">17</xref></sup></p>
          <p>We followed the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) guideline for conducting our research and writing this article.<sup><xref ref-type="bibr" rid="B18">18</xref></sup></p>
          <p>All analyses and model training were conducted using R software. We considered statistically significant p-values &lt; 0.05.</p>
        </sec>
      </sec>
      <sec>
        <title>Results</title>
        <p>From January 2016 to December 2022, a total of 3,532 patients with AHF were included in the BPC program. After excluding those with VAD, on the transplant waiting list, or with missing vital status at discharge, 3,043 were considered eligible for the study (2,309 in Cohort 1 and 734 in Cohort 2, as seen in <xref ref-type="fig" rid="f02001">Figure 1</xref>).</p>
        <p>Of 113 variables available at admission, nine were excluded due to more than 40% missing data (Supplemental Table S1). The Boruta algorithm was applied to 70% patients (n=1,616) randomly selected from Cohort 1, which yielded 17 variables that were selected to build the ML model.</p>
        <p>Out of the 3,043 eligible patients, 1,157 patients were included for the final analysis due to complete data for all 17 variables. The training set consisted of 621 patients, which accounted for 70% of the 879 individuals in Cohort 1, while the test set consisted of 536 patients, comprising 258 remaining patients of Cohort 1 and an additional 278 patients from Cohort 2.</p>
        <p>The in-hospital mortality rate was 10.1%, with a total of 117 deaths among the 1,157 patients examined. The mean age of these patients was 61 &#177; 15 years, LVEF was 42 &#177; 17 % and 58% were male (<xref ref-type="table" rid="t1001">Table 1</xref>). <xref ref-type="table" rid="t2001">Table 2</xref> displays these selected variables along with variables used in the ADHERE and GWTG scores for comparison.</p>
        <p>
          <table-wrap id="t1001">
            <label>Table 1</label>
            <caption>
              <title>&#8211; Baseline characteristics of included patients</title>
            </caption>
            <table frame="hsides" rules="groups">
              <colgroup>
                <col/>
                <col/>
                <col/>
                <col/>
              </colgroup>
              <thead>
                <tr>
                  <th align="left">Parameters</th>
                  <th style="font-weight:normal">Discharged alive (n=1040)</th>
                  <th style="font-weight:normal">In-hospital death (n=117)</th>
                  <th>p-value</th>
                </tr>
              </thead>
              <tbody>
                <tr>
                  <td>
                    <bold>Age, years</bold>
                  </td>
                  <td align="center">61.3 &#177; 14.5</td>
                  <td align="center">61.6 &#177; 14.4</td>
                  <td align="center">0.829</td>
                </tr>
                <tr>
                  <td>
                    <bold>Female sex, n (%)</bold>
                  </td>
                  <td align="center">424 (40.8)</td>
                  <td align="center">53 (45.3)</td>
                  <td align="center">0.398</td>
                </tr>
                <tr>
                  <td>
                    <bold>White race, n(%)</bold>
                  </td>
                  <td align="center">674 (64.8)</td>
                  <td align="center">78 (66.7)</td>
                  <td align="center">0.766</td>
                </tr>
                <tr>
                  <td>
                    <bold>Education, n(%)</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Illiterate</td>
                  <td align="center">84 (8.1)</td>
                  <td align="center">9 (7.7)</td>
                  <td align="center" rowspan="7">0.242</td>
                </tr>
                <tr>
                  <td>Incomplete elementary school</td>
                  <td align="center">399 (38.4)</td>
                  <td align="center">47 (40.2)</td>
                </tr>
                <tr>
                  <td>Elementary school</td>
                  <td align="center">148 (14.2)</td>
                  <td align="center">26 (22.2)</td>
                </tr>
                <tr>
                  <td>Incomplete high school</td>
                  <td align="center">72 (6.9)</td>
                  <td align="center">8 (6.8)</td>
                </tr>
                <tr>
                  <td>High school</td>
                  <td align="center">225 (21.6)</td>
                  <td align="center">20 (17.1)</td>
                </tr>
                <tr>
                  <td>Incomplete higher education</td>
                  <td align="center">30 (2.9)</td>
                  <td align="center">2 (1.7)</td>
                </tr>
                <tr>
                  <td>Complete higher education</td>
                  <td align="center">82 (7.9)</td>
                  <td align="center">5 (4.3)</td>
                </tr>
                <tr>
                  <td>
                    <bold>Average income, n(%)</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>&#8804; 1 minimum wage</td>
                  <td align="center">721 (69.3)</td>
                  <td align="center">97 (82.9)</td>
                  <td align="center">0.009</td>
                </tr>
                <tr>
                  <td>
                    <bold>Smoker, n(%)</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>No</td>
                  <td align="center">520 (50)</td>
                  <td align="center">56 (47.9)</td>
                  <td align="center" rowspan="3">0.889</td>
                </tr>
                <tr>
                  <td>Current</td>
                  <td align="center">438 (42.1)</td>
                  <td align="center">52 (44.4)</td>
                </tr>
                <tr>
                  <td>Past</td>
                  <td align="center">82 (7.9)</td>
                  <td align="center">9 (7.7)</td>
                </tr>
                <tr>
                  <td>
                    <bold>Alcoholic, n(%)</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>No</td>
                  <td align="center">654 (62.9)</td>
                  <td align="center">80 (68.4)</td>
                  <td align="center" rowspan="3">0.200</td>
                </tr>
                <tr>
                  <td>Current</td>
                  <td align="center">246 (23.7)</td>
                  <td align="center">28 (23.9)</td>
                </tr>
                <tr>
                  <td>Past</td>
                  <td align="center">140 (13.5)</td>
                  <td align="center">9 (7.7)</td>
                </tr>
                <tr>
                  <td>
                    <bold>HF etiology, n(%)</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Ischemic</td>
                  <td align="center">200 (19.2)</td>
                  <td align="center">22 (18.8)</td>
                  <td align="center" rowspan="4">&lt; 0.001</td>
                </tr>
                <tr>
                  <td>Chagas disease</td>
                  <td align="center">95 (9.1)</td>
                  <td align="center">13 (11.1)</td>
                </tr>
                <tr>
                  <td>Valvar disease</td>
                  <td align="center">123 (11.8)</td>
                  <td align="center">30 (25.6)</td>
                </tr>
                <tr>
                  <td>Other</td>
                  <td align="center">622 (59.8)</td>
                  <td align="center">52 (44.4)</td>
                </tr>
                <tr>
                  <td>Valvar prosthesis, n(%)</td>
                  <td align="center">338 (32.5)</td>
                  <td align="center">52 (44.4)</td>
                  <td align="center">0.013</td>
                </tr>
                <tr>
                  <td>Chronic kidney failure, n(%)</td>
                  <td align="center">152 (14.6)</td>
                  <td align="center">27 (23.1)</td>
                  <td align="center">0.024</td>
                </tr>
                <tr>
                  <td>Hypothyroidism, n(%)</td>
                  <td align="center">77 (7.4)</td>
                  <td align="center">19 (16.2)</td>
                  <td align="center">0.002</td>
                </tr>
                <tr>
                  <td>LVEF, %</td>
                  <td align="center">41.7 &#177; 17.1</td>
                  <td align="center">41.90 &#177; 18.5</td>
                  <td align="center">0.911</td>
                </tr>
                <tr>
                  <td>Left atrium diameter, mm</td>
                  <td align="center">45.8 &#177; 8.5</td>
                  <td align="center">48.1 &#177; 9.4</td>
                  <td align="center">0.005</td>
                </tr>
                <tr>
                  <td>Number of comorbidities&#8224;</td>
                  <td align="center">4 (3 - 5)</td>
                  <td align="center">4 (2 - 5)</td>
                  <td align="center">0.665</td>
                </tr>
                <tr>
                  <td>
                    <bold>Number of previous surgical procedures, n(%)</bold>
                    <sup>
                      <bold>&#8225;</bold>
                    </sup>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td align="center">0.002</td>
                </tr>
                <tr>
                  <td>0</td>
                  <td align="center">702 (67.5)</td>
                  <td align="center">65 (55.6)</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>1</td>
                  <td align="center">288 (27.7)</td>
                  <td align="center">36 (30.8)</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>2</td>
                  <td align="center">46 (4.4)</td>
                  <td align="center">14 (12.0)</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>3 or more</td>
                  <td align="center">4 (0.3)</td>
                  <td align="center">2 (1.7)</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>
                    <bold>Clinical characteristics at admission</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>
                    <bold>Hemodynamic profile, n(%)</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td align="center">0.020</td>
                </tr>
                <tr>
                  <td>A</td>
                  <td align="center">78 (7.5)</td>
                  <td align="center">8 (6.8)</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>B</td>
                  <td align="center">797 (76.6)</td>
                  <td align="center">77 (65.8)</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>C</td>
                  <td align="center">130 (12.5)</td>
                  <td align="center">25 (21.4)</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>L</td>
                  <td align="center">35 (3.4)</td>
                  <td align="center">7 (6.0)</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Systolic blood pressure, mmHg</td>
                  <td align="center">123 &#177; 29</td>
                  <td align="center">115&#177; 27</td>
                  <td align="center">0.004</td>
                </tr>
                <tr>
                  <td>Diastolic blood pressure, mmHg</td>
                  <td align="center">76 &#177; 17</td>
                  <td align="center">71 &#177; 16</td>
                  <td align="center">0.001</td>
                </tr>
                <tr>
                  <td>Low peripheral perfusion, n(%)</td>
                  <td align="center">127 (12.2)</td>
                  <td align="center">20 (17.1)</td>
                  <td align="center">0.175</td>
                </tr>
                <tr>
                  <td>Orthopnea, n(%)</td>
                  <td align="center">757 (74.6)</td>
                  <td align="center">78 (70.9)</td>
                  <td align="center">0.471</td>
                </tr>
                <tr>
                  <td>Ascites, n(%)</td>
                  <td align="center">242 (24.6)</td>
                  <td align="center">39 (35.5)</td>
                  <td align="center">0.019</td>
                </tr>
                <tr>
                  <td>B3, n(%)</td>
                  <td align="center">103 (10.3)</td>
                  <td align="center">13 (11.5)</td>
                  <td align="center">0.806</td>
                </tr>
                <tr>
                  <td>Peripheral edema, n(%)</td>
                  <td align="center">779 (75.6)</td>
                  <td align="center">99 (85.3)</td>
                  <td align="center">0.025</td>
                </tr>
                <tr>
                  <td>
                    <bold>Laboratory results at admission (blood levels)</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Creatinine, mg/dL</td>
                  <td align="center">1.71 &#177; 1.51</td>
                  <td align="center">1.93 &#177; 1.30</td>
                  <td align="center">0.136</td>
                </tr>
                <tr>
                  <td>BUN, mg/dL</td>
                  <td align="center">33.07 &#177; 22.8</td>
                  <td align="center">46.0 &#177; 30.4</td>
                  <td align="center">&lt; 0.001</td>
                </tr>
                <tr>
                  <td>Sodium, mmol/L</td>
                  <td align="center">136.6 &#177; 5.0</td>
                  <td align="center">134.2 &#177; 6.9</td>
                  <td align="center">&lt; 0.001</td>
                </tr>
                <tr>
                  <td>Potassium, mmol/L</td>
                  <td align="center">4.46 &#177; 0.77</td>
                  <td align="center">4.66 &#177; 0.92</td>
                  <td align="center">0.010</td>
                </tr>
                <tr>
                  <td>Hemoglobin, g/dL</td>
                  <td align="center">12.6 &#177; 2.28</td>
                  <td align="center">12.32 &#177; 2.4</td>
                  <td align="center">0.207</td>
                </tr>
                <tr>
                  <td>
                    <bold>WHOQOL-BREF questionnaire</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Physical health domain</td>
                  <td align="center">50.4 &#177; 19.4</td>
                  <td align="center">39.5 &#177;15.8</td>
                  <td align="center">&lt; 0.001</td>
                </tr>
                <tr>
                  <td>Social relationships domain</td>
                  <td align="center">64.2 &#177; 16.5</td>
                  <td align="center">61.1 &#177; 17.3</td>
                  <td align="center">0.059</td>
                </tr>
                <tr>
                  <td>Environmental health</td>
                  <td align="center">58.9 &#177; 12.6</td>
                  <td align="center">54.5 &#177; 12.9</td>
                  <td align="center">&lt;0.001</td>
                </tr>
                <tr>
                  <td>Psychological health</td>
                  <td align="center">77.3 &#177; 17.9</td>
                  <td align="center">68.8 &#177; 18.0</td>
                  <td align="center">&lt;0.001</td>
                </tr>
              </tbody>
            </table>
            <table-wrap-foot>
              <fn id="TFN1001">
                <p>Categorical variables are shown as n (%) and continuous variables as mean &#177; standard deviation or median (25<sup>th</sup> percentile - 75<sup>th</sup> percentile). HF: heart failure; LVEF: left ventricular ejection fraction; WHOQOL-BREF: World Health Organization quality of life brief version; BUN: Blood Urea Nitrogen. &#8224; Comorbidities included hypertension, diabetes, dyslipidemia, cerebrovascular accident, carotid disease, coronary artery disease, peripheral vascular disease, acute myocardial infarction, atrial fibrillation, Chagas disease, rheumatic heart disease, valve disease, chronic obstructive pulmonary disease, sleep apnea, liver disease, chronic kidney failure, anemia, hyperthyroidism, hypothyroidism, depression, solid organ cancer, and hematologic cancer. &#8225; Previous surgical procedures included coronary angioplasty, myocardial revascularization surgery, pacemaker or implantable cardiac defibrillator, cardiac resynchronization therapy, and valve prosthesis.</p>
              </fn>
            </table-wrap-foot>
          </table-wrap>
        </p>
        <p>
          <table-wrap id="t2001">
            <label>Table 2</label>
            <caption>
              <title>&#8211; Comparison of variables used in the ADHERE, GWTG, and ML-HF scores</title>
            </caption>
            <table frame="hsides" rules="groups">
              <colgroup>
                <col/>
                <col/>
                <col/>
                <col/>
                <col/>
              </colgroup>
              <thead>
                <tr>
                  <th align="left">VARIABLE</th>
                  <th>ADHERE</th>
                  <th>GWTG</th>
                  <th>ML-HF</th>
                </tr>
              </thead>
              <tbody>
                <tr>
                  <td>
                    <bold>Demographic</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Age, years</td>
                  <td align="center">&#10003;</td>
                  <td align="center">&#10003;</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Black race</td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>
                    <bold>Clinical characteristics</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>BUN, mg/dL</td>
                  <td align="center">&#10003;</td>
                  <td align="center">&#10003;</td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Creatinine, mg/dL</td>
                  <td align="center">&#10003;</td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Left atrium diameter</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>COPD</td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Chronic Kidney Disease</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Valve etiology of HF</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Heart rate</td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Respiratory frequency</td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Hemoglobin, g/dL</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Hypothyroidism</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Number of previous cardiac procedures</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Systolic Blood Pressure, mmHg</td>
                  <td align="center">&#10003;</td>
                  <td align="center">&#10003;</td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Diastolic Blood Pressure, mmHg</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Hemodynamic profile B</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Hemodynamic profile C</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Hemodynamic profile L</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Bad peripheral perfusion</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Valve prosthesis</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>Sodium, mmol/L</td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                  <td align="center">&#10003;</td>
                </tr>
                <tr>
                  <td>
                    <bold>Quality of life</bold>
                  </td>
                  <td> </td>
                  <td> </td>
                  <td> </td>
                </tr>
                <tr>
                  <td>Physical health domain &#8211; WHOQOL-BREF</td>
                  <td> </td>
                  <td> </td>
                  <td align="center">&#10003;</td>
                </tr>
              </tbody>
            </table>
            <table-wrap-foot>
              <fn id="TFN2001">
                <p>BUN: Blood urea nitrogen; COPD: chronic obstructive pulmonary disease; WHOQOL-BREF: World Health Organization quality of life brief version.</p>
              </fn>
            </table-wrap-foot>
          </table-wrap>
        </p>
        <sec>
          <title>Algorithm selection</title>
          <p>Among the ML algorithms evaluated, the Random Forest model showed the highest performance in predicting in-hospital mortality compared to Gradient Boosting Machines and Extreme Gradient Boosting (<xref ref-type="table" rid="t3001">Table 3</xref>). Accordingly, the Random Forest algorithm was chosen to build the final model.</p>
          <p>
            <table-wrap id="t3001">
              <label>Table 3</label>
              <caption>
                <title>&#8211; Discrimination performance in the training set according to the Machine Learning algorithm</title>
              </caption>
              <table frame="hsides" rules="groups">
                <colgroup>
                  <col/>
                  <col/>
                </colgroup>
                <thead>
                  <tr>
                    <th align="left">Machine Learning Algorithm</th>
                    <th style="font-weight:normal">AUC (Mean &#177; SD) N = 810</th>
                  </tr>
                </thead>
                <tbody>
                  <tr>
                    <td>Random Forest</td>
                    <td align="center">0.772 &#177; 0.07</td>
                  </tr>
                  <tr>
                    <td>Gradient Boosting Machines</td>
                    <td align="center">0.703 &#177; 0.08</td>
                  </tr>
                  <tr>
                    <td>Extreme Gradient Boosting</td>
                    <td align="center">0.724 &#177; 0.09</td>
                  </tr>
                </tbody>
              </table>
              <table-wrap-foot>
                <fn id="TFN3001">
                  <p>SD: standard deviation; AUC: area under the receiving operating curve.</p>
                </fn>
              </table-wrap-foot>
            </table-wrap>
          </p>
        </sec>
        <sec>
          <title>Internal validation</title>
          <p>In the training set (n=621), the ML-HF score demonstrated good to excellent accuracy for predicting in-hospital mortality, with an AUC of 0.906, outperforming both the ADHERE and GWTG-HF scores. [<xref ref-type="fig" rid="f04001">Figure 3A</xref>] [<xref ref-type="fig" rid="f04001">Central Illustration</xref>].</p>
          <p>
            <fig id="f04001">
              <label>Figure 3</label>
              <caption>
                <title>&#8211; Comparison of the receiver operating characteristic (ROC) curves among the ML-HF, GWTG-HF, and ADHERE scores in the training (A) and testing (B) sets.</title>
              </caption>
              <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf04-en.tif"/></fig>
          </p>
        </sec>
        <sec>
          <title>External validation</title>
          <p>In the test set (n=536), the ML-HF score maintained acceptable discrimination, with an AUC of 0.722. Although lower than in the training set, its performance remained superior to the ADHERE and GWTG scores (<xref ref-type="fig" rid="f04001">Figure 3B</xref>, Central Illustration).. Similar results were observed when using the full Cohort 2 alone as the test set (n=280), as shown in Supplemental Figure S2.</p>
        </sec>
        <sec>
          <title>Model calibration</title>
          <p><xref ref-type="fig" rid="f05001">Figure 4</xref> displays the observed vs predicted in-hospital mortality of each score in the test set. While the ML-HF score showed good model calibration, predicted mortality significantly diverged from the observed mortality in the other two conventional scores. Specifically, the ADHERE score underestimated the risk, while the GTWG score overestimated it.</p>
          <p>
            <fig id="f05001">
              <label>Figure 4</label>
              <caption>
                <title>&#8211; Comparison of observed and predicted in-hospital mortality to assess model calibration of the ML-HF (A), GWTG-HF (B), and ADHERE (C) scores in the testing set. HL: Hosmer-Lemeshow test. A p-value less than 0.05 indicates an inadequate fit of the data.</title>
              </caption>
              <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf05-en.tif"/></fig>
          </p>
        </sec>
        <sec>
          <title>Variable importance in ML-HF score</title>
          <p><xref ref-type="fig" rid="f06001">Figure 5</xref> shows the relative importance of the ML-HF score variables in decreasing order. The variable with the highest importance was the Physical Health domain of the WHOQOL, followed by the sodium, BUN, and creatinine blood levels, as showed in the Central Illustration. Blood pressure at admission was also among the ten most important variables, with SBP and DBP ranking in the fifth and eighth positions, respectively. Notably, LVEF did not have sufficient importance to predict in-hospital mortality in the Boruta method. The only echocardiographic parameter in the model was left atrium size, ranking sixth in order of importance.</p>
          <p>
            <fig id="f06001">
              <label>Figure 5</label>
              <caption>
                <title>&#8211; Importance of the variables included in the ML-HF score. WHOQOL-BREF: World Health Organization quality of life brief version; SBP: systolic blood pressure; DBP: diastolic blood pressure; BUN: Blood Urea Nitrogen..</title>
              </caption>
              <graphic xlink:href="1678-4170-abc-122-11-e20250136-50136-gf06-en.tif"/></fig>
          </p>
        </sec>
      </sec>
      <sec sec-type="results">
        <title>Discussion</title>
        <p>This study has three main findings. We developed a new ML-based score &#8211; the ML-HF score - to predict in-hospital mortality in patients with AHF in a LMIC, which was externally validated with acceptable accuracy and good calibration. The ML-HF score outperformed the conventional ADHERE and GWTG-HF scores in the same population.<sup><xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B5">5</xref></sup> The study revealed that information gathered from a quality-of-life questionnaire emerged as the most crucial predictor of mortality. Overall, this study provides a new tool to predict risk in AHF patients and highlights the potential of ML to improve risk stratification models within this population.</p>
        <p>To our knowledge, this is the first study to incorporate a personal quality-of-life measure in an ML-based model to predict mortality in HF. Surprisingly, the physical health domain of the WHOQOL-BREF questionnaire emerged as the most important predictor of mortality, surpassing serum biomarkers and echocardiographic parameters. This new and unexpected finding was likely made possible only through the use of ML techniques, as they allow for the analysis of variables in a non-linear and predefined manner. This is an example of how ML can analyze intricate patterns in data to find counterintuitive predictors from the human perspective. This result underscores the importance of incorporating patient-reported outcomes (PRO) and considering the patient's self-perception about their health. In simpler terms, it highlights the value of truly listening to the patient.</p>
        <p>Previous studies have shown the prognosis value of PRO in patients with chronic HF. In the CHAMP-HF (Change the Management of Patients With Heart Failure) registry, an improvement of 5 points in the Kansas City Cardiomyopathy Questionnaire was associated with a 41% lower risk of all-cause death.<sup><xref ref-type="bibr" rid="B19">19</xref></sup> Notably, PRO has proven to be a better predictor of future events compared to the NYHA functional class, underscoring the importance of assessing the patient's perspective on their health status in addition to the clinician assessment.<sup><xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B20">20</xref></sup> In fact, current HF guidelines recommend standardized assessment of PRO as a class 2a recommendation.<sup><xref ref-type="bibr" rid="B10">10</xref></sup> Our study adds to the literature by demonstrating the independent prognostic role of measuring PRO at the admission of patients experiencing decompensated HF.</p>
        <p>It is possible to highlight the fact that new variables, in addition to the WHOQOL-Brief physical health domain, were relevant for predicting in-hospital death in our study, factors that are not assessed in other scores, such as hemoglobin, left atrium diameter, presence of prosthetic valve, number of surgical procedures, and presence of hypothyroidism. The relevance of the presence of a prosthetic valve may be due to the high prevalence of rheumatic fever in our population. This once again highlights the importance of developing personalized scores for LMIC countries.<sup><xref ref-type="bibr" rid="B21">21</xref></sup></p>
        <p>Machine-learning models have been previously tested to predict in-hospital mortality of patients with AHF in other countries. The MARKER-HF (Machine learning Assessment of RisK and EaRly mortality in Heart Failure) model demonstrated excellent discriminatory power, with an AUC of 0.88. However, this model was developed using both chronic and AHF patients, with varying follow-up durations.<sup><xref ref-type="bibr" rid="B9">9</xref></sup> Segar et al. utilized data from the GWTG registry in the United States to construct an ML-based model to predict in-hospital mortality, which was validated in the Atherosclerosis Risk in Community (ARIC) cohort.<sup><xref ref-type="bibr" rid="B5">5</xref></sup> Much like our investigation, they observed that the ML-based model outperformed the GWTG-HF score, boasting a C-index of approximately 0.80. They were able to incorporate serum biomarkers, such as troponin and natriuretic peptides blood levels, and contextual-level social determinants of health, which helps reduce algorithm bias and improve fairness among underserved populations.<sup><xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref></sup></p>
        <p>Differences in patient populations, clinical management, and availability of biomarkers such as natriuretic peptides may partly explain the higher accuracy of the Segar et al.<sup><xref ref-type="bibr" rid="B5">5</xref></sup> model compared to ours; however, our model may be particularly valuable in settings with limited access to these biomarkers, especially in LMIC. In our study, individual-level social determinants were used but not selected from the ML-based algorithm. On the other hand, we have shown for the first time that a quality-of-life measure helps improve the prediction of in-hospital death of HF patients in an ML-based model.</p>
        <p>Our study has strengths that should be highlighted. Firstly, we analyzed social determinants of health individually, even though they were not selected as important variables. This finding is in line with current literature, which states that social determinants may be relevant only in more homogeneous subgroups, which may not be the case for the Brazilian population.<sup><xref ref-type="bibr" rid="B4">4</xref></sup> Secondly, we also performed external validation using a different cohort from that used for model training, which should be encouraged in studies using ML.<sup><xref ref-type="bibr" rid="B7">7</xref></sup> Thirdly, our sample contains patients from all macro regions of the country, encompassing a wide range of demographic characteristics and social statuses. This diversity enhances the generalizability and validation of our findings across different populations. Fourthly, our data was systematically collected as part of a prospective cohort study, rather than relying on EMR, which are more susceptible to misclassification bias.<sup><xref ref-type="bibr" rid="B22">22</xref>-<xref ref-type="bibr" rid="B26">26</xref></sup></p>
        <p>This study also has limitations that should be considered. Missing data led to a reduction in the initial sample size, decreasing it to approximately 25% of the original size as we removed patients with missing data in any of the most significant variables. We included only complete cases, assuming data was missing completely at random (MCAR). We opted against multiple imputations to avoid introducing potential bias from model-driven estimates in a predictive setting. In addition, natriuretic peptides were not used to build the model due to a lack of availability in most centers, because it is a database in which the vast majority of sites are publicly funded. We recognize that the absence of these markers may impair the analysis, as these are currently essential tests for confirming the diagnosis of HF. In countries with greater availability, we believe that this variable may be even more relevant and increase the accuracy of the algorithm. On the other hand, this could improve the applicability of the ML score in other LMIC, where access to BNP may also be limited.</p>
        <p>Despite promising results, implementing ML-based risk scores in clinical practice remains challenging. Unlike traditional risk scores, which can use simple tables or calculators, ML models require digital infrastructure for data input, processing, and result integration into clinical workflows. In our study, model development and validation were performed on a centralized computer containing data from all participating centers, and analysis was computationally intensive. For real-world use, deployment would require compatible systems and standardized input data. Strategies such as developing offline mobile applications or integrating the model into low-complexity electronic health records could be explored to facilitate implementation in resource-limited settings.</p>
        <p>While the current model can be applied without retraining, ML-based models may be periodically updated or retrained to improve performance and adapt to evolving clinical populations. This raises important ethical and regulatory considerations, including data privacy, informed consent for secondary use of patient data, and transparency around model updating procedures. Effective implementation will depend not only on adequate technical infrastructure to support computational demands but also on strong oversight and governance frameworks to ensure safe, equitable, and responsible use.</p>
      </sec>
      <sec sec-type="conclusions">
        <title>Conclusion</title>
        <p>We developed and validated a machine learning-based model - the ML-HF score - to predict in-hospital mortality among patients with AHF. This predictive tool presented a discriminative power superior to traditional ADHERE and GWTG scores.</p>
      </sec>
      <sec sec-type="supplementary-material">
        <title>*Supplemental Materials</title>
        <supplementary-material id="suppl01-en">
          <p>For additional information, please click here. <ext-link ext-link-type="uri" xlink:href="https://abccardiol.org/supplementary-material/2025/12211/2025-0136_supplementary_material.pdf">https://abccardiol.org/supplementary-material/2025/12211/2025-0136_supplementary_material.pdf</ext-link> </p>
          <media mime-subtype="pdf" mimetype="application" xlink:href="https://minio.scielo.br/documentstore/1678-4170/XVyRRmkytVYdx8hM6nd7qwM/3ed1c116bd46f6527e95586bd4b3302acc3a5505.pdf"/>
        </supplementary-material>
      </sec>
    </body>
    <back>
      <ack>
        <title>Acknowledgments</title>
        <p>The authors would like to express gratitude to the participating colleagues who supported recruitment for the BPC program.</p>
      </ack>
      <fn-group>
        <fn fn-type="other">
          <label>Study association:</label>
          <p> This article is part of the thesis of master submitted by Gustavo Sarot Pereira, from Universidade Federal do Paran&#225;.</p>
        </fn>
        <fn fn-type="other">
          <label>Ethics approval and consent to participate:</label>
          <p> This study was approved by the Ethics Committee of the Hospital do Cora&#231;&#227;o - HCOR / Associa&#231;&#227;o Beneficiente S&#237;ria under the protocol number 48561715.5.1001.0060. All the procedures in this study were in accordance with the 1975 Helsinki Declaration, updated in 2013. Informed consent was obtained from all participants included in the study.</p>
        </fn>
        <fn fn-type="other">
          <label>Use of Artificial Intelligence:</label>
          <p> The authors did not use any artificial intelligence tools in the development of this work.</p>
        </fn>
        <fn fn-type="data-availability" specific-use="data-available-upon-request">
          <label>Data Availability:</label>
          <p> The data will be made available to the authors upon request.</p>
        </fn>
        <fn fn-type="supported-by">
          <label>Sources of funding:</label>
          <p>This study was partially funded by American Heart Association, Hospital do Cora&#231;&#227;o (HCor) and Minist&#233;rio da Sa&#250;de.</p>
        </fn>
      </fn-group>
    </back>
  </sub-article>
</article>
