Deployment and Validation of a Smart System for Screening of Language Disorders in Primary Care
Identifieur interne : 000571 ( Pmc/Corpus ); précédent : 000570; suivant : 000572Deployment and Validation of a Smart System for Screening of Language Disorders in Primary Care
Auteurs : María Luisa Martín-Ruiz ; Miguel Ngel Valero Duboy ; Iván Pau De La CruzSource :
- Sensors (Basel, Switzerland) [ 1424-8220 ] ; 2013.
Abstract
Neuro-evolutive development from birth until the age of six years is a decisive factor in a child's quality of life. Early detection of development disorders in early childhood can facilitate necessary diagnosis and/or treatment. Primary-care pediatricians play a key role in its detection as they can undertake the preventive and therapeutic actions requested to promote a child's optimal development. However, the lack of time and little specific knowledge at primary-care avoid to applying continuous early-detection anomalies procedures. This research paper focuses on the deployment and evaluation of a smart system that enhances the screening of language disorders in primary care. Pediatricians get support to proceed with early referral of language disorders. The proposed model provides them with a decision-support tool for referral actions to trigger essential diagnostic and/or therapeutic actions for a comprehensive individual development. The research was conducted by starting from a sample of 60 cases of children with language disorders. Validation was carried out through two complementary steps: first, by including a team of seven experts from the fields of neonatology, pediatrics, neurology and language therapy, and, second, through the evaluation of 21 more previously diagnosed cases. The results obtained show that therapist positively accepted the system proposal in 18 cases (86%) and suggested system redesign for single referral to a speech therapist in three remaining cases.
Url:
DOI: 10.3390/s130607522
PubMed: 23752564
PubMed Central: 3715251
Links to Exploration step
PMC:3715251Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en">Deployment and Validation of a Smart System for Screening of Language Disorders in Primary Care</title>
<author><name sortKey="Martin Ruiz, Maria Luisa" sort="Martin Ruiz, Maria Luisa" uniqKey="Martin Ruiz M" first="María Luisa" last="Martín-Ruiz">María Luisa Martín-Ruiz</name>
</author>
<author><name sortKey="Duboy, Miguel Ngel Valero" sort="Duboy, Miguel Ngel Valero" uniqKey="Duboy M" first="Miguel Ngel Valero" last="Duboy">Miguel Ngel Valero Duboy</name>
</author>
<author><name sortKey="De La Cruz, Ivan Pau" sort="De La Cruz, Ivan Pau" uniqKey="De La Cruz I" first="Iván Pau" last="De La Cruz">Iván Pau De La Cruz</name>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">PMC</idno>
<idno type="pmid">23752564</idno>
<idno type="pmc">3715251</idno>
<idno type="url">http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3715251</idno>
<idno type="RBID">PMC:3715251</idno>
<idno type="doi">10.3390/s130607522</idno>
<date when="2013">2013</date>
<idno type="wicri:Area/Pmc/Corpus">000571</idno>
<idno type="wicri:explorRef" wicri:stream="Pmc" wicri:step="Corpus" wicri:corpus="PMC">000571</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en" level="a" type="main">Deployment and Validation of a Smart System for Screening of Language Disorders in Primary Care</title>
<author><name sortKey="Martin Ruiz, Maria Luisa" sort="Martin Ruiz, Maria Luisa" uniqKey="Martin Ruiz M" first="María Luisa" last="Martín-Ruiz">María Luisa Martín-Ruiz</name>
</author>
<author><name sortKey="Duboy, Miguel Ngel Valero" sort="Duboy, Miguel Ngel Valero" uniqKey="Duboy M" first="Miguel Ngel Valero" last="Duboy">Miguel Ngel Valero Duboy</name>
</author>
<author><name sortKey="De La Cruz, Ivan Pau" sort="De La Cruz, Ivan Pau" uniqKey="De La Cruz I" first="Iván Pau" last="De La Cruz">Iván Pau De La Cruz</name>
</author>
</analytic>
<series><title level="j">Sensors (Basel, Switzerland)</title>
<idno type="eISSN">1424-8220</idno>
<imprint><date when="2013">2013</date>
</imprint>
</series>
</biblStruct>
</sourceDesc>
</fileDesc>
<profileDesc><textClass></textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en"><p>Neuro-evolutive development from birth until the age of six years is a decisive factor in a child's quality of life. Early detection of development disorders in early childhood can facilitate necessary diagnosis and/or treatment. Primary-care pediatricians play a key role in its detection as they can undertake the preventive and therapeutic actions requested to promote a child's optimal development. However, the lack of time and little specific knowledge at primary-care avoid to applying continuous early-detection anomalies procedures. This research paper focuses on the deployment and evaluation of a smart system that enhances the screening of language disorders in primary care. Pediatricians get support to proceed with early referral of language disorders. The proposed model provides them with a decision-support tool for referral actions to trigger essential diagnostic and/or therapeutic actions for a comprehensive individual development. The research was conducted by starting from a sample of 60 cases of children with language disorders. Validation was carried out through two complementary steps: first, by including a team of seven experts from the fields of neonatology, pediatrics, neurology and language therapy, and, second, through the evaluation of 21 more previously diagnosed cases. The results obtained show that therapist positively accepted the system proposal in 18 cases (86%) and suggested system redesign for single referral to a speech therapist in three remaining cases.</p>
</div>
</front>
<back><div1 type="bibliography"><listBibl><biblStruct></biblStruct>
<biblStruct></biblStruct>
<biblStruct></biblStruct>
<biblStruct></biblStruct>
<biblStruct><analytic><author><name sortKey="Narbona, J" uniqKey="Narbona J">J. Narbona</name>
</author>
<author><name sortKey="Chevrie Muller, C" uniqKey="Chevrie Muller C">C. Chevrie-Muller</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Mossabeb, R" uniqKey="Mossabeb R">R. Mossabeb</name>
</author>
<author><name sortKey="Wade, K C" uniqKey="Wade K">K.C. Wade</name>
</author>
<author><name sortKey="Finnegan, K" uniqKey="Finnegan K">K. Finnegan</name>
</author>
<author><name sortKey="Sivieri, E" uniqKey="Sivieri E">E. Sivieri</name>
</author>
<author><name sortKey="Abbasi, S" uniqKey="Abbasi S">S. Abbasi</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Arrabal Teran, M C" uniqKey="Arrabal Teran M">M.C. Arrabal Terán</name>
</author>
<author><name sortKey="Arizcun Pineda, J" uniqKey="Arizcun Pineda J">J. Arizcun Pineda</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Nelson, H D" uniqKey="Nelson H">H.D. Nelson</name>
</author>
<author><name sortKey="Nygren, P" uniqKey="Nygren P">P. Nygren</name>
</author>
<author><name sortKey="Walker, M" uniqKey="Walker M">M. Walker</name>
</author>
<author><name sortKey="Panoscha, R" uniqKey="Panoscha R">R. Panoscha</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Fejerman, N" uniqKey="Fejerman N">N. Fejerman</name>
</author>
<author><name sortKey="Fernandez Lvarez, E" uniqKey="Fernandez Lvarez E">E. Fernández Álvarez</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Paul, R" uniqKey="Paul R">R. Paul</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Paul, R" uniqKey="Paul R">R. Paul</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Castro Rebolledo, R" uniqKey="Castro Rebolledo R">R. Castro-Rebolledo</name>
</author>
<author><name sortKey="Giraldo Prieto, M" uniqKey="Giraldo Prieto M">M. Giraldo-Prieto</name>
</author>
<author><name sortKey="Hincapie Henao, L" uniqKey="Hincapie Henao L">L. Hincapié-Henao</name>
</author>
<author><name sortKey="Lopera, F" uniqKey="Lopera F">F Lopera</name>
</author>
<author><name sortKey="Pineda, D A" uniqKey="Pineda D">D.A. Pineda</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Przybylski, L" uniqKey="Przybylski L">L. Przybylski</name>
</author>
<author><name sortKey="Bedoin, N" uniqKey="Bedoin N">N. Bedoin</name>
</author>
<author><name sortKey="Krifi Papoz, S" uniqKey="Krifi Papoz S">S. Krifi-Papoz</name>
</author>
<author><name sortKey="Herbillon, V" uniqKey="Herbillon V">V. Herbillon</name>
</author>
<author><name sortKey="Roch, D" uniqKey="Roch D">D. Roch</name>
</author>
<author><name sortKey="Leculier, L" uniqKey="Leculier L">L. Léculier</name>
</author>
<author><name sortKey="Kotz, S A" uniqKey="Kotz S">S.A. Kotz</name>
</author>
<author><name sortKey="Tillmann, B" uniqKey="Tillmann B">B. Tillmann</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Naddy, C" uniqKey="Naddy C">C. Naddy</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Mclellan, M C" uniqKey="Mclellan M">M.C. McLellan</name>
</author>
<author><name sortKey="Connor, J A" uniqKey="Connor J">J.A. Connor</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Parshuram, C" uniqKey="Parshuram C">C. Parshuram</name>
</author>
<author><name sortKey="Hutchison, J" uniqKey="Hutchison J">J. Hutchison</name>
</author>
<author><name sortKey="Middaugh, K" uniqKey="Middaugh K">K. Middaugh</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Haines, C" uniqKey="Haines C">C. Haines</name>
</author>
<author><name sortKey="Perrott, M" uniqKey="Perrott M">M. Perrott</name>
</author>
<author><name sortKey="Weir, P" uniqKey="Weir P">P. Weir</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Bates, D W" uniqKey="Bates D">D.W. Bates</name>
</author>
<author><name sortKey="Kuperman, G J" uniqKey="Kuperman G">G.J. Kuperman</name>
</author>
<author><name sortKey="Wang, S" uniqKey="Wang S">S. Wang</name>
</author>
<author><name sortKey="Gandhi, T" uniqKey="Gandhi T">T. Gandhi</name>
</author>
<author><name sortKey="Kittler, A" uniqKey="Kittler A">A. Kittler</name>
</author>
<author><name sortKey="Volk, L" uniqKey="Volk L">L. Volk</name>
</author>
<author><name sortKey="Spurr, C" uniqKey="Spurr C">C. Spurr</name>
</author>
<author><name sortKey="Khorasani, R" uniqKey="Khorasani R">R. Khorasani</name>
</author>
<author><name sortKey="Tanasijevic, M" uniqKey="Tanasijevic M">M. Tanasijevic</name>
</author>
<author><name sortKey="Middleton, B" uniqKey="Middleton B">B. Middleton</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Shortlife, E H" uniqKey="Shortlife E">E.H. Shortlife</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Alonso Betanzos, A" uniqKey="Alonso Betanzos A">A. Alonso Betanzos</name>
</author>
<author><name sortKey="Guijarro Berdi As, B" uniqKey="Guijarro Berdi As B">B. Guijarro Berdiñas</name>
</author>
<author><name sortKey="Lozano Tello, A" uniqKey="Lozano Tello A">A. Lozano Tello</name>
</author>
<author><name sortKey="Palma Mendez, J T" uniqKey="Palma Mendez J">J.T. Palma Méndez</name>
</author>
<author><name sortKey="Taboada Iglesias, M J" uniqKey="Taboada Iglesias M">M.J. Taboada Iglesias</name>
</author>
</analytic>
</biblStruct>
<biblStruct></biblStruct>
<biblStruct><analytic><author><name sortKey="Carretero Diaz, L E" uniqKey="Carretero Diaz L">L.E. Carretero Díaz</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Mate Hernandez, J L" uniqKey="Mate Hernandez J">J.L. Maté Hernández</name>
</author>
<author><name sortKey="Pazos Sierra, J" uniqKey="Pazos Sierra J">J. Pazos Sierra</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Lopez Sanchez, J I" uniqKey="Lopez Sanchez J">J.I. Lopez Sanchez</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Corcho, O" uniqKey="Corcho O">O. Corcho</name>
</author>
<author><name sortKey="Fernandez L Pez, M" uniqKey="Fernandez L Pez M">M. Fernández-López</name>
</author>
<author><name sortKey="G Mez Perez, A" uniqKey="G Mez Perez A">A. Gómez-Pérez</name>
</author>
<author><name sortKey="L Pez Cima, A" uniqKey="L Pez Cima A">A. López-Cima</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Pe Afiel Puerto, M" uniqKey="Pe Afiel Puerto M">M. Peñafiel Puerto</name>
</author>
</analytic>
</biblStruct>
<biblStruct></biblStruct>
<biblStruct></biblStruct>
<biblStruct></biblStruct>
<biblStruct><analytic><author><name sortKey="Velasquez, J D" uniqKey="Velasquez J">J.D. Velásquez</name>
</author>
<author><name sortKey="Palade, V" uniqKey="Palade V">V. Palade</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Torsun, I S" uniqKey="Torsun I">I.S. Torsun</name>
</author>
</analytic>
</biblStruct>
<biblStruct></biblStruct>
<biblStruct><analytic><author><name sortKey="Frankenburg, W K" uniqKey="Frankenburg W">W.K. Frankenburg</name>
</author>
<author><name sortKey="Dodds, J B" uniqKey="Dodds J">J.B. Dodds</name>
</author>
<author><name sortKey="Archer, P" uniqKey="Archer P">P. Archer</name>
</author>
</analytic>
</biblStruct>
<biblStruct><analytic><author><name sortKey="Hervas, R" uniqKey="Hervas R">R. Hervás</name>
</author>
<author><name sortKey="Bravo, J" uniqKey="Bravo J">J. Bravo</name>
</author>
<author><name sortKey="Fontecha, J" uniqKey="Fontecha J">J. Fontecha</name>
</author>
</analytic>
</biblStruct>
</listBibl>
</div1>
</back>
</TEI>
<pmc article-type="research-article"><pmc-dir>properties open_access</pmc-dir>
<front><journal-meta><journal-id journal-id-type="nlm-ta">Sensors (Basel)</journal-id>
<journal-id journal-id-type="iso-abbrev">Sensors (Basel)</journal-id>
<journal-title-group><journal-title>Sensors (Basel, Switzerland)</journal-title>
</journal-title-group>
<issn pub-type="epub">1424-8220</issn>
<publisher><publisher-name>Molecular Diversity Preservation International (MDPI)</publisher-name>
</publisher>
</journal-meta>
<article-meta><article-id pub-id-type="pmid">23752564</article-id>
<article-id pub-id-type="pmc">3715251</article-id>
<article-id pub-id-type="doi">10.3390/s130607522</article-id>
<article-id pub-id-type="publisher-id">sensors-13-07522</article-id>
<article-categories><subj-group subj-group-type="heading"><subject>Article</subject>
</subj-group>
</article-categories>
<title-group><article-title>Deployment and Validation of a Smart System for Screening of Language Disorders in Primary Care</article-title>
</title-group>
<contrib-group><contrib contrib-type="author"><name><surname>Martín-Ruiz</surname>
<given-names>María Luisa</given-names>
</name>
<xref rid="c1-sensors-13-07522" ref-type="corresp"><sup>*</sup>
</xref>
</contrib>
<contrib contrib-type="author"><name><surname>Duboy</surname>
<given-names>Miguel Ángel Valero</given-names>
</name>
</contrib>
<contrib contrib-type="author"><name><surname>de la Cruz</surname>
<given-names>Iván Pau</given-names>
</name>
</contrib>
<aff id="af1-sensors-13-07522">Departamento de Ingeniería y Arquitecturas Telemáticas, Universidad Politécnica de Madrid, Carretera de Valencia km. 7, Madrid 28031, Spain; E-Mails:<email>mavalero@diatel.upm.es</email>
(M.A.V.D.);<email>ipau@diatel.upm.es</email>
(I.P.C.)</aff>
</contrib-group>
<author-notes><corresp id="c1-sensors-13-07522"><label>*</label>
Author to whom correspondence should be addressed; E-Mail: <email>marisam@diatel.upm.es</email>
; Tel.: +34-91-336-3770; Fax: +34-91-336-7817.</corresp>
</author-notes>
<pub-date pub-type="collection"><month>6</month>
<year>2013</year>
</pub-date>
<pub-date pub-type="epub"><day>10</day>
<month>6</month>
<year>2013</year>
</pub-date>
<volume>13</volume>
<issue>6</issue>
<fpage>7522</fpage>
<lpage>7545</lpage>
<history><date date-type="received"><day>19</day>
<month>4</month>
<year>2013</year>
</date>
<date date-type="rev-recd"><day>10</day>
<month>5</month>
<year>2013</year>
</date>
<date date-type="accepted"><day>04</day>
<month>6</month>
<year>2013</year>
</date>
</history>
<permissions><copyright-statement>© 2013 by the authors; licensee MDPI, Basel, Switzerland.</copyright-statement>
<copyright-year>2013</copyright-year>
<license><license-p><pmc-comment>CREATIVE COMMONS</pmc-comment>
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (<ext-link ext-link-type="uri" xlink:href="http://creativecommons.org/licenses/by/3.0/">http://creativecommons.org/licenses/by/3.0/</ext-link>
).</license-p>
</license>
</permissions>
<abstract><p>Neuro-evolutive development from birth until the age of six years is a decisive factor in a child's quality of life. Early detection of development disorders in early childhood can facilitate necessary diagnosis and/or treatment. Primary-care pediatricians play a key role in its detection as they can undertake the preventive and therapeutic actions requested to promote a child's optimal development. However, the lack of time and little specific knowledge at primary-care avoid to applying continuous early-detection anomalies procedures. This research paper focuses on the deployment and evaluation of a smart system that enhances the screening of language disorders in primary care. Pediatricians get support to proceed with early referral of language disorders. The proposed model provides them with a decision-support tool for referral actions to trigger essential diagnostic and/or therapeutic actions for a comprehensive individual development. The research was conducted by starting from a sample of 60 cases of children with language disorders. Validation was carried out through two complementary steps: first, by including a team of seven experts from the fields of neonatology, pediatrics, neurology and language therapy, and, second, through the evaluation of 21 more previously diagnosed cases. The results obtained show that therapist positively accepted the system proposal in 18 cases (86%) and suggested system redesign for single referral to a speech therapist in three remaining cases.</p>
</abstract>
<kwd-group><kwd>intelligent healthcare</kwd>
<kwd>knowledge management</kwd>
<kwd>early attention</kwd>
<kwd>e-health</kwd>
<kwd>language disorders</kwd>
</kwd-group>
</article-meta>
</front>
<body><sec><label>1.</label>
<title>Introduction</title>
<p>Pediatricians that work in the Spanish Healthcare System (SHS) are primary care physicians who provide care in health centers to children between the ages of 0 and 14. These professionals mainly use information systems that fulfill the WONCA standards [<xref rid="b1-sensors-13-07522" ref-type="bibr">1</xref>
] with different degrees of compliance. A widely deployed system is OMI, a software solution aligned with WONCA to register medical data that does not cover key aspects to follow-up the children's neuro-evolutive and social development.</p>
<p>The Early Intervention Program (New York State Department of Health [<xref rid="b2-sensors-13-07522" ref-type="bibr">2</xref>
]) reinforces the aspect of permanent monitoring of children: “The early years of a child's life are very important. In these years, children grow quickly and have so much to learn. Developmental delay means that a child has not attained developmental milestones expected for the child's age adjusted for prematurity in one or more of the following areas of development: cognitive, physical (including vision and hearing), communication, social-emotional, or adaptive development”. Furthermore, the Speech-language pathology is included as a key area for the Early Intervention Official [<xref rid="b2-sensors-13-07522" ref-type="bibr">2</xref>
].</p>
<p>As stated in the Spanish white paper on early attention [<xref rid="b3-sensors-13-07522" ref-type="bibr">3</xref>
], pediatricians must screen and correctly refer children with potential disorders towards diagnostic and monitoring centers. Present rates of detection of development disorders are lower than their real incidence [<xref rid="b4-sensors-13-07522" ref-type="bibr">4</xref>
], which means that early identification of children with such disorders remains a pending task. In certain cases, language disorders are the first symptoms a child will manifest of a possible development disorder [<xref rid="b5-sensors-13-07522" ref-type="bibr">5</xref>
] and expressive language delay is present in many cases of preterm infants [<xref rid="b6-sensors-13-07522" ref-type="bibr">6</xref>
].</p>
<p>However, in clinical practice, pediatricians usually lack enough time to perform proper screening of the children's neuro-evolutive development, and their background in disabilities is not always as complete as it would be desirable [<xref rid="b7-sensors-13-07522" ref-type="bibr">7</xref>
]. For these reasons, a scientific and technological contribution to this problem is the deployment and validation a Knowledge Based System (KBS) in primary care, capable of providing pediatricians with valuable support to enable efficient screening of children's disorders. A smart screening system in such healthcare scenario does require the construction of a substantial and comprehensive Knowledge Base (KB) to test effective approaches to this problem. This research paper describes the process of building such a KBS based on multi-disciplinary work with experts and the analysis of 21 actual proven cases diagnosed in the Language Intervention Center (LIC). Moreover, the result stage presents the KBS verification with 21 cases previously diagnosed in the LIC.</p>
<p>Section 2 provides details on the background of this research from two complementary perspectives: the standpoint of experts in the detection of neurological disorders in childhood, and application of expert systems in medicine. The next section explains the methodology used in the development of the KBS for primary pediatric care. A description is provided of the process of system knowledge education with Knowledge Based System tools, the procedures for extracting information managed by experts and the stages of construction of the ontology used in the reasoning model. Sections 4 and 5 present the development and the deployment of the system, respectively. Section 6 reports the main results of the research which includes two complementary approaches for the platform validation: firstly, through the participation of seven experts from the fields of neonatology, pediatrics, neurology and language therapy, and, second, by evaluating 21 previously diagnosed cases. Section 7 concludes the paper and present future work directions.</p>
</sec>
<sec><label>2.</label>
<title>Background</title>
<sec><label>2.1.</label>
<title>Early Detection Systems of Neurological Disorders in Children</title>
<p>The Spanish white paper on early attention emphasizes the importance of intervention for the transitory nature of a certain development disorder [<xref rid="b3-sensors-13-07522" ref-type="bibr">3</xref>
]. Such disorders must be considered a significant deviation from the “course” of development resulting from health or relational events that compromise biological, psychological and social development. Hence, detection of possible alterations in child development is essential for early attention, as it will foster the activation of a number of requested mechanisms of action.</p>
<p>The utilization of smart systems in primary care can improve the detection of neurological disorders in children by enabling the prevention of added pathologies and fostering functional improvements that allow a more adaptive relationship between a child and his or her surroundings. Narbona highlights that a delay in the correct acquisition of speech and language is a first alarm sign of a future neurological disorder although it is not the only symptom to have into account to diagnose a neurological disorder [<xref rid="b5-sensors-13-07522" ref-type="bibr">5</xref>
]. The paper published by Nelson <italic>et al.</italic>
explains why the correct acquisition of language is of vital importance [<xref rid="b8-sensors-13-07522" ref-type="bibr">8</xref>
], while Fejerman stresses that a complete neurological and pediatric evaluation can reveal related developmental disorders, starting with detection of a language disorder [<xref rid="b9-sensors-13-07522" ref-type="bibr">9</xref>
]. Although medical procedures are available to detect a number of neurological disorders in children [<xref rid="b4-sensors-13-07522" ref-type="bibr">4</xref>
,<xref rid="b10-sensors-13-07522" ref-type="bibr">10</xref>
–<xref rid="b12-sensors-13-07522" ref-type="bibr">12</xref>
] these procedures are difficult to apply in primary pediatric care, as many require significant time and specialized knowledge. The review in this research found no solutions that exploit the potential of Information Systems in combination with artificial intelligence to provide pediatricians with efficient and effective assistance in the early detection of these disorders.</p>
<p>The systematic review of systems that may improve early attention on children was elaborated in this research according to the followings related sources: PubMed, JAMIA and ModernMedicine. The most valuable keywords used were “early attention”, “language disorders” and “eHealth”.</p>
<p><list list-type="bullet"><list-item><p>Przybylski <italic>et al.</italic>
presented a study tested the influence of external rhythmic auditory stimulation (<italic>i.e.</italic>
, musical rhythm) on syntax processing in children with specific language impairment (SLI) and dyslexia. Results points to potential avenues in using rhythmic structures (even in nonverbal materials) to boost linguistic structure processing [<xref rid="b13-sensors-13-07522" ref-type="bibr">13</xref>
].</p>
</list-item>
<list-item><p>Naddy described a pediatric early warning tool used with routine nursing observations will alert staff to the need for increased monitoring, the support of an associated outreach team or emergency medical attention [<xref rid="b14-sensors-13-07522" ref-type="bibr">14</xref>
].</p>
</list-item>
<list-item><p>McLellan and Connor describe the modification of a pediatric early warning scoring system for cardiovascular patients, the implementation of the tool, and its companion Escalation of Care Algorithm on an inpatient pediatric cardiovascular unit [<xref rid="b15-sensors-13-07522" ref-type="bibr">15</xref>
].</p>
</list-item>
<list-item><p>Parshuram, Hutchison and Middaugh developed and validated a simple bedside score to quantify severity of illness in hospitalized children [<xref rid="b16-sensors-13-07522" ref-type="bibr">16</xref>
].</p>
</list-item>
<list-item><p>Haines, Perrott and Weir developed and evaluated a physiologically based system for the identification of acutely ill children in hospital environments [<xref rid="b17-sensors-13-07522" ref-type="bibr">17</xref>
].</p>
</list-item>
</list>
</p>
</sec>
<sec><label>2.2.</label>
<title>Knowledge-Based Systems in Medicine</title>
<p>Medicine is one of the fields to have benefited most from the use of computers, as a pioneer in the use of Knowledge Based Systems [<xref rid="b18-sensors-13-07522" ref-type="bibr">18</xref>
]. Two landmarks that showed the validity of these technologies in medicine are ELIZA and MYCIN. The laboratories of the Massachusetts Institute of Technology (MIT) developed in the 1960s the ELIZA project, simulating the behavior of a psychoanalyst. This was the first application in psychiatry that offered a “credible” answer to patients' questions by constructing generated sentences to these questions by changing certain words or phrases. In the mid-1970s, the MYCIN project emerged, and over time it became one of the most influential systems in the history of KBS [<xref rid="b19-sensors-13-07522" ref-type="bibr">19</xref>
,<xref rid="b20-sensors-13-07522" ref-type="bibr">20</xref>
]. MYCIN was a system designed for diagnostics and therapy of infectious blood diseases.</p>
<p>The success of aforementioned past experiences positively contributed to the use of these technologies in medicine and healthcare. <xref rid="t1-sensors-13-07522" ref-type="table">Table 1</xref>
contains a list of applications to support decision-making on monitoring of multiple diseases, including the name of the system, a brief description of its use and the date and country of creation [<xref rid="b21-sensors-13-07522" ref-type="bibr">21</xref>
]. As we may observe, only the LISA project and the SimulConsult tool are addressed to detect health problems in children. SimulConsult can perform neurological evaluations of children. In our study of KBS in medicine, no tool was found for the early detection of language disorders in children.</p>
<p>Carretero Díaz [<xref rid="b22-sensors-13-07522" ref-type="bibr">22</xref>
] defined an Expert System (ES) as “a system that solves problems in a certain area, with the skill and precision of an expert” in [<xref rid="b20-sensors-13-07522" ref-type="bibr">20</xref>
,<xref rid="b23-sensors-13-07522" ref-type="bibr">23</xref>
] they can find other definitions of these type of systems. An ES is an ideal system in order that the experts share knowledge with not expert people, therefore the system must get up-to-date constant on the part of experts which will allow that it should be used by specialists who do not have sufficient knowledge in the area of application. The most important advantages of an ES in order to contribute with regard to other information systems are [<xref rid="b24-sensors-13-07522" ref-type="bibr">24</xref>
]:
<list list-type="bullet"><list-item><p>It can support the decisions of many persons simultaneously thanks to the system's terminals (a human expert cannot to be available at all time and in several places at the same time).</p>
</list-item>
<list-item><p>It can improve the productivity of the system (minor time of response).</p>
</list-item>
<list-item><p>It can provide stability and consistency to the process of decision in a certain area.</p>
</list-item>
<list-item><p>It reduces the dependence opposite to the employees (for vacations, to leave working places, <italic>etc.</italic>
).</p>
</list-item>
<list-item><p>End with the shortage of available experts and reduce the cost of access to the knowledge of the experts.</p>
</list-item>
<list-item><p>It is an excellent tool of training, it justifies the decisions.</p>
</list-item>
</list>
</p>
</sec>
</sec>
<sec><label>3.</label>
<title>Methods, Materials and Design</title>
<sec><label>3.1.</label>
<title>Methods</title>
<p>The acquisition and systematization of needed knowledge required for the early referral KBS of a smart system proposed in this paper is a critical aspect that determines its effective use in primary care. The process of Knowledge Acquisition (KA) is the first step for creating a KBS and it strongly influences the conditions for correct operation. This process covers up to the final stage of KBS development.</p>
<sec><label>3.1.1.</label>
<title>Method to Knowledge Acquisition and Formalization</title>
<p>The methodology for KA requires consideration of both the definition of the knowledge to be systematized and the conceptualization and formal design of the information compiled from human and materials sources in order to model the functioning of the KBS. For this reason, the main methodologies available for extracting knowledge were studied, with a comparison of GROVER, CommonKADS (CK), Methontology and IDEAL. The conclusion of this study was to use a combination of CK and Methontology, as these methods offer the greatest potential for application in certain phases of the KBS construction. The relation between the KA collected from humans' expertise and its translation into a formal ontology is successfully achieved in a cyclic way.</p>
<p>CommonKADS is a knowledge engineering methodology for the design and development of KBS based on knowledge extracted from human experts and its codification to allow for its processing by a system [<xref rid="b20-sensors-13-07522" ref-type="bibr">20</xref>
]. The application of CK for the system design provided a set of early detection items to be considered by the pediatrician. This structured knowledge reflects all important aspects of the KBS to be implemented and verified through a user tool. CK was used in KA meetings as it is most suitable for modeling the knowledge extracted from language specialists in the form of ontology.</p>
<p>Methontology is a methodology oriented towards the implementation of ontology in the activity of conceptualization of the KBS that has been successfully used by many authors [<xref rid="b25-sensors-13-07522" ref-type="bibr">25</xref>
]. Methontology defines a set of tasks that enabled moving from an informal specification of the domain of application, collected with the language specialists, to a semi-formal specification of the domain. This facility makes easier the understanding of the ontology for smart screening of language disorders by consulted neuropediatricians and language therapists as well as the system developer.</p>
<p>One key point to guarantee the success of the KBS construction is the selection of a suitable team of experts. Next, the members of the team of experts are presented:
<list list-type="bullet"><list-item><p>A neonatologist with high expertise in development disorders and child disability, former director of the neonatology department of San Carlos Hospital in Madrid,</p>
</list-item>
<list-item><p>Five primary care pediatricians, in two Primary Attention Centers,</p>
</list-item>
<list-item><p>A neuro-pediatrician presently working in the Quirón Hospital of Madrid,</p>
</list-item>
<list-item><p>Two experts in specific language impairment who are therapists at the Language Intervention Center (LIC) at La Salle Campus (UAM) of Madrid.</p>
</list-item>
<list-item><p>A therapist working at Legamar School.</p>
</list-item>
</list>
</p>
</sec>
<sec><label>3.1.2.</label>
<title>Knowledge Base System Development</title>
<p>Once the objectives, methods, and experts have been identified a strategy for KBS development is needed. <xref rid="f1-sensors-13-07522" ref-type="fig">Figure 1</xref>
shows the whole process for the development and validation of the final solution.</p>
<p>KBS development has been carried out in five steps:
<list list-type="bullet"><list-item><p><bold>Problem Inception</bold>
: its aim is to define the problem to be solved for early intervention. The group of experts involved in the KA process in this phase were a pediatrician, a neuropediatrician and a neonatologist. Five open meetings, held between September 2009 and May 2011, supported the problem definition and led to specifically work with early screening of language disorders, by the construction of a smart detection system.</p>
</list-item>
<list-item><p><bold>System Development</bold>
: it covers the implementation phase of the KBS and its core KB. In this phase, the group of experts included a neuropediatrician and two therapists in specific language impairment who work at the Language Intervention Center (LIC) at La Salle Campus (UAM) of Madrid. They are experts on neurological disorders in children and helped to create the KB through ten structured meetings. The process of building and purging the KB was developed through a retrospective analysis of information on levels of language acquisition of 21 children who received therapy in the LIC [<xref rid="b26-sensors-13-07522" ref-type="bibr">26</xref>
]. The KB has been developed using Protégé (ontology creation platform) [<xref rid="b27-sensors-13-07522" ref-type="bibr">27</xref>
] and Pellet (reasoning engine) [<xref rid="b28-sensors-13-07522" ref-type="bibr">28</xref>
]. The verification of the KB required a usable tool so that specialists might interact with the KBS in an efficient way.</p>
</list-item>
<list-item><p><bold>Platform Verification</bold>
: experts should be able to evaluate whether the system proposal to refer to a specialist arising from a detection of language development disorders was correct or not. This process contributed in a satisfactory way to improve the graphical user interface and the reasoning rules of the KB. A web tool (Gades) has been built in the verification stage to facilitate the work of experts in specific language impairment who are working at the LIC and a therapist working at Legamar School [<xref rid="b29-sensors-13-07522" ref-type="bibr">29</xref>
]. This phase has two planned stages:
<list list-type="order"><list-item><p>KB verification by means of children's retrospective cases who have received therapy in the LIC. This process of check it has fulfilled Mrs. Maria Peñafiel, who until August, 2012 was working as therapist of the LIC and currently she is a principal in the Legamar School. The 21 cases used in this stage differ with regard to the conceptualization stage. The obtained results are gathered in the results chapter. These results will be used for the iterative refinement of the KB in next steps.</p>
</list-item>
<list-item><p>KB verification by means of children's Legamar School. This checking stages runs from March to September, 2013 along six months. This validation aims to get ready for the summative evaluation to be performed in clinical routine with all kinds of children having a normative development or with a linguistic or development delay.</p>
</list-item>
</list>
</p>
</list-item>
<list-item><p><bold>Service Evaluation</bold>
: this controlled evaluation is scheduled for autumn 2013 in cases considered of interest by the five pediatricians already involved in the final stage. Thus, a web tool (Pegasus) is being built to facilitate the work of experts and primary care pediatricians. Six doctors (five pediatricians and one neuropediatrician) will be involve in the verification stages both for usability and system performance tests along six months. End users pointed out that they could use the tool by themselves in clinical routine.</p>
</list-item>
<list-item><p><bold>Knowledge Acquisition</bold>
: acquisition and formalization processes were developed on the basis of information gathered in open meetings and then structured with the team of experts. The process of acquisition of expert knowledge relied on the use of additional techniques such as questionnaires, surveys and interviews designed in accordance with the objectives to be met by the smart system, with a view to support pediatricians in primary care working in the public health system.</p>
</list-item>
</list>
</p>
<p>Knowledge Representation (KR) is the first task to take into account to develop a system. KR must globally consider the transformations of the web interface, <italic>i.e.</italic>
, changes in the web site structure and content, as well as user interaction models. It is not a trivial task to find out an effective KR method. Velasquez and Palade suggested a Knowledge Base (KB) creation methodology to implement a web-based computerized recommendation system [<xref rid="b30-sensors-13-07522" ref-type="bibr">30</xref>
]. The information of the KB is a key part in the KBS and this work presents the KB result in the section four.</p>
</sec>
</sec>
<sec><label>3.2.</label>
<title>Materials</title>
<p>Children's data involved in the KA process and KB verification came from 21 LIC cases. <xref rid="t2-sensors-13-07522" ref-type="table">Table 2</xref>
outlines ratio of children who could not speak or had unclear speech that were seen by the pediatrician as requested by their parents. Professionals sounded alarms by themselves along routine child visits.</p>
<p>The development of the Gades web tool supported the verification process with retrospective cases from LIC. Experts benefit from this tool to validate the responses of the system for the developmental items provided by the KB [<xref rid="b29-sensors-13-07522" ref-type="bibr">29</xref>
]. Primary care pediatricians will validate the KB on autumn 2013 by using an <italic>ad-hoc</italic>
web interface called the Pegasus web tool.</p>
<p>The knowledge extracted in meetings should be represented in a well-structured and comprehensible manner in a KB to make the knowledge useful and relatable for solving problems arising in the domain of execution. The solution of problems detected in the execution domain request a KR of data acquired in meetings. A well-structured and comprehensible manner is needed to provide useful and relatable knowledge through the KB. As noted by Torsun in 1995, KR requires formalisms or structures that represent it, either through declarative logic, mathematical formulas or concept maps [<xref rid="b31-sensors-13-07522" ref-type="bibr">31</xref>
]. The use of logical basis, as a tool for KR, requires adapting formal languages to models of knowledge expressed in natural language. Out of all the KR available models, this research used description logic as it directly supports the development of ontologies for integration in the semantic web, as it is done extensively with Ontology Web Language (OWL) language [<xref rid="b32-sensors-13-07522" ref-type="bibr">32</xref>
].</p>
</sec>
<sec><label>3.3.</label>
<title>Functional Specification of the System (Design)</title>
<p>Pegasus system was designed according to the use case model of <xref rid="f2-sensors-13-07522" ref-type="fig">Figure 2</xref>
, written in Unified Modeling Language (UML). The KBS was modeled with this use cases diagram, which describes its proposed functionality as offered by the system. Main systems actors are:
<list list-type="bullet"><list-item><p>The primary attention pediatrician</p>
</list-item>
<list-item><p>The specialist the child has been a derivative after the introduction of his case in the system for primary attention pediatrician</p>
</list-item>
<list-item><p>The knowledge engineer entrusted to manage the information of the users of the system and of managing the information of the KB</p>
</list-item>
</list>
</p>
<p>A description of the actions for every use case is provided below (see <xref rid="f2-sensors-13-07522" ref-type="fig">Figure 2</xref>
):
<list list-type="order"><list-item><p><bold>Language Evaluation</bold>
: the primary care pediatrician will perform the language acquisition evaluation in a child. The pediatrician will connect to corresponding Web page and will update the information of the child requested by the system. The system will verify if the validity of this information and will ask for modifications in case of error.</p>
</list-item>
<list-item><p><bold>Obtain Evaluation Result</bold>
: when the introduced information is correct, the system shows question suggested to check the language status of the child corresponding with the child's age in months. If the state is normal the system will indicate it with a message, in other case the system can propose bringing forward visit (for re-evaluation of the level of language acquisition) or suggest referral to an appropriate specialist.</p>
</list-item>
<list-item><p><bold>Language Evaluation Consultation</bold>
: the specialist to whom the child is derived by his/her primary care pediatrician will be able to consult the study case information and verify the answers given by the pediatrician according to the age of the child. In this way, the specialist has information to support the diagnosis procurement tasks.</p>
</list-item>
<list-item><p><bold>Users' Management</bold>
: the knowledge engineer will be administrator of the information of the users with the different profiles that they can access to the system.</p>
</list-item>
<list-item><p><bold>KB' Management</bold>
: there will exist the possibility of modifying the information of the domain, in order to introduce new inferences that modify the reasoning of the KBS.</p>
</list-item>
</list>
</p>
</sec>
</sec>
<sec><label>4.</label>
<title>Development</title>
<p>This section contains the development of the research. First, it describes the general System Architecture. Secondly, it describes the construction of the KB and then the process of formalization of the KB in OWL with the tool Protégé.</p>
<sec><label>4.1.</label>
<title>General System Architecture (Pegasus Web Tool)</title>
<p>The functional architecture of the resulting system must facilitate dynamic interaction between the actors involved; distributed platforms for the management of information, models of reasoning and processes in line with the health care model in which it is located (see <xref rid="f3-sensors-13-07522" ref-type="fig">Figure 3</xref>
).</p>
<p>Next a description of the steps given in <xref rid="f3-sensors-13-07522" ref-type="fig">Figure 3</xref>
is provided:
<list list-type="order"><list-item><p>The child goes to the family pediatrician accompanied by a family member.</p>
</list-item>
<list-item><p>Primary health-care doctor decides to use the KBS to assess whether there is a language disorder in the child. The doctor will introduce the required information.</p>
</list-item>
<list-item><p>The KBS returns the result to the pediatrician. Three possibilities:
<list list-type="bullet"><list-item><p>The result is that everything is normal.</p>
</list-item>
<list-item><p>The result changes the visitation schedule the child's pediatrician.</p>
</list-item>
<list-item><p>The KBS proposed to be derived from the relevant specialist hospital. In this case the doctor decides whether to accept the decision of KBS and pursue the request for appointment at the hospital, or decide it is not necessary.</p>
<p>As always the KBS decision supports the pediatrician, but is never an imposition on the decisions that the doctor makes.</p>
</list-item>
</list>
</p>
</list-item>
<list-item><p>Health Center Administration requests an appointment with the specialist.</p>
</list-item>
<list-item><p>An appointment with the specialist is requested from the hospital.</p>
</list-item>
<list-item><p>Response is received to the request for appointment at the hospital,</p>
</list-item>
<list-item><p>The details of the appointment with the specialist will be received by the pediatrician.</p>
</list-item>
<list-item><p>The appointment details reach the child and his family.</p>
</list-item>
<list-item><p>The child goes to the specialist.</p>
</list-item>
<list-item><p>The specialist checks the KBS response for the corresponding case.</p>
</list-item>
<list-item><p>The DSS returns the result of the evaluation process for that case.</p>
</list-item>
<list-item><p>The DSS notifies the pediatrician that the specialist has accessed to the system. The specialist will validate the KBS outcome in order to improve the KB.</p>
</list-item>
</list>
</p>
</sec>
<sec><label>4.2.</label>
<title>Knowledge Base (KB) for Language Disorders</title>
<p>The KB has been built through an iterative process of structured meetings between September 2011 and May 2012. Only the two language therapists and the neuro-pediatrician participated in the meetings, using CK techniques to extract information such as structured interviews to complete the KB with questions to be asked by the primary care pediatrician children's tutor upon arriving at the care facility. The starting point for building the KB is the Denver Test, as it is extensively used in primary care [<xref rid="b33-sensors-13-07522" ref-type="bibr">33</xref>
].</p>
<p>With this basis, questions were sharpened to focus on language, with a view to enabling the primary care pediatrician to detect possible delays in children's development that require closer attention or immediate referral to an appropriate specialist. The process of building and purging the KB is based on the experience of LIC, who checked developmental items against the appearance of language disorders through a retrospective analysis of information on levels of language acquisition of 21 children who received therapy in the LIC [<xref rid="b26-sensors-13-07522" ref-type="bibr">26</xref>
]. In materials sub-section summarizes the data of the 21 LIC cases analyzed in the KA process (<xref rid="t2-sensors-13-07522" ref-type="table">Table 2</xref>
), which lays the groundwork for the KB by indicating the most frequent alarm periods in detection of disorders and the resulting diagnosis.</p>
<p>Observation of the data gets back to the hypothesis that diagnosis in all cases is related to the level of language acquisition, or problems deriving from it, when children begin to read and write. In most cases, it is the family which belatedly detects delays in language when a child barely speaks at the age of two years. In two of the 21 cases, parents did not become alarmed until the child has surpassed the age of three years. Information systematized with the LIC therapists is significant for mastering the problem [<xref rid="b26-sensors-13-07522" ref-type="bibr">26</xref>
] and enabled relating KB questions with the demand for intervention in cases of either alert or alarm.</p>
<p>The study showed that the questions would have a negative answer in many of the cases under study and, for this reason; therapists consider them to be significant. The structuring of the final KB consists of 136 questions between month 1 and month 72 in the life of the child, and questions may be of two main types:
<list list-type="bullet"><list-item><p>Questions called Alert Milestones that imply bringing forward the visit. A negative answer to these developmental milestones means that the child makes return visit within three months to allow for re-evaluation of the level of language acquisition.</p>
</list-item>
<list-item><p>Questions called Alarm Milestones that imply referral. These items could be considered as reasons for alarm and suggest referral to an appropriate specialist.</p>
</list-item>
</list>
</p>
<p><xref rid="t3-sensors-13-07522" ref-type="table">Table 3</xref>
details the KB questions for a child between the ages of 1 and 4 months. The first column indicates the child's age in months at the time of evaluation and the question type (Alert or Alarm). The second column shows the question the pediatrician asks the child's tutor to evaluate the child's state of language acquisition and the “System decision” column contains the system's answer in the event of a negative answer to the question (referral to specialist or bring forward visit).</p>
</sec>
<sec><label>4.3.</label>
<title>Formalization of Primary Care KB in OWL</title>
<p>As noted by Hervas <italic>et al.</italic>
, “An Ontology can be specified through several formal mechanisms”. In order to allow the computation of the conceptual knowledge gathered from experts a formal representation is needed [<xref rid="b34-sensors-13-07522" ref-type="bibr">34</xref>
]. One common approach to represent ontological concepts is using basic modeling languages such as entity-relationship model. More powerful languages, as frame-based languages, allow the definition of concepts and relationships. Currently, the most expressive languages are based on logic-based models, as First-order Logic for example, which allow the specification of concepts, relationships and restrictions. The challenge to choose one or another language is the agreement between expressiveness and computability. Next, the formalization process of the conceptualization model is presented.</p>
<p>In the revision carried out for this research, the construction of the ontology according to Methontology required categorizing the questions the pediatrician must ask according to the months of age of the child at the time of evaluation.</p>
<p>Protégé was used in formalization of the knowledge model to create the KB and the inference engine needed to support decision making. Protégé offers an open and useful environment for the design, modeling, implementation, manipulation and viewing of ontologies.</p>
<p>The ontology of Protégé was built with a class hierarchy for the first six years, including a sub-hierarchy of classes for each month corresponding to the questions to be asked by the pediatrician. The class hierarchy in AvanceSL, in each month, includes the questions for each month as classes, as shown in <xref rid="f4-sensors-13-07522" ref-type="fig">Figure 4</xref>
for months 2 and 3 of the first year. The OWL of the binary relationships between ontology classes is defined as follows:
<array><tbody><tr><td align="left" valign="top" rowspan="1" colspan="1"><bold>NextVisitInThreeMonths</bold>
”></td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> <bold>AV_NoEmitsOOOAAH_2M</bold>
”/></td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> <bold>AV_NoScreamsToInteract_2M</bold>
”/></td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> <bold>ThereIsNegativeResponseIn”</bold>
/></td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"> </td>
</tr>
<tr><td align="left" valign="top" rowspan="1" colspan="1"></td>
</tr>
</tbody>
</array>
</p>
<p>The definition in OWL of the binary relationships established between ontology classes will sustain the system's reasoning process through axioms such as: If the child is 2 months old and we get a negative answer to the question: “Emits OOO/AAH” or “Screams to interact”, then “To anticipate the next visit in three months” (see <xref rid="f5-sensors-13-07522" ref-type="fig">Figure 5</xref>
).</p>
<p>Hence, the class hierarchy has been created to make inferences through the Decision System class, where this class contains motor decisions according to year and the type of milestone to which the decision belongs. <xref rid="f6-sensors-13-07522" ref-type="fig">Figure 6</xref>
shows the logical formulation of the correspondence of these axioms with inferences in Protégé.</p>
</sec>