Health improvement framework for planning actionable treatment process using surrogate Bayesian model

Health improvement framework for planning actionable treatment process using surrogate Bayesian model | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Health improvement framework for planning actionable treatment process using surrogate Bayesian model Kazuki Nakamura, Ryosuke Kojima, Eiichiro Uchino, Koichi Murashita, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-107700/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 25 May, 2021 Read the published version in Nature Communications → Version 1 posted You are reading this latest preprint version Abstract Clinical decision making regarding treatments based on personal characteristics leads to effective health improvements. Machine learning (ML) has been the primary concern of diagnosis support according to comprehensive patient information. However, the remaining prominent issue is the development of objective treatment processes in clinical situations. This study proposes a novel framework to plan treatment processes in a data-driven manner. A key point of the framework is the evaluation of the "actionability" for personal health improvements by using a surrogate Bayesian model in addition to a high-performance nonlinear ML model. We first evaluated the framework from the viewpoint of its methodology using a synthetic dataset. Subsequently, the framework was applied to an actual health checkup dataset comprising data from 3,132 participants, to improve systolic blood pressure values at the individual level. We confirmed that the computed treatment processes are actionable and consistent with clinical knowledge for lowering blood pressure. These results demonstrate that our framework could contribute toward decision making in the medical field, providing clinicians with deeper insights. Artificial Intelligence and Machine Learning Preventive Medicine Clinical Decision Making Objective Treatment Processes Computed Treatment Processes Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Full Text Due to technical limitations, full-text HTML conversion of this manuscript could not be completed. However, the latest manuscript can be downloaded and accessed as a PDF. Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryInformation.docx Supplementary Information Cite Share Download PDF Status: Published Journal Publication published 25 May, 2021 Read the published version in Nature Communications → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-107700","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":4840597,"identity":"b0d9e757-4d82-4eda-af20-5ef2ac6dd827","order_by":0,"name":"Kazuki Nakamura","email":"","orcid":"","institution":"Kyoto University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Kazuki","middleName":"","lastName":"Nakamura","suffix":""},{"id":4840598,"identity":"f5501722-81fe-494e-b6a9-5ddfdb51b97b","order_by":1,"name":"Ryosuke Kojima","email":"","orcid":"","institution":"Graduate School of Medicine, Kyoto University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Ryosuke","middleName":"","lastName":"Kojima","suffix":""},{"id":4840599,"identity":"180c236e-d533-4844-93fe-b63e51dd3943","order_by":2,"name":"Eiichiro Uchino","email":"","orcid":"","institution":"Graduate School of Medicine, Kyoto University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Eiichiro","middleName":"","lastName":"Uchino","suffix":""},{"id":4840600,"identity":"985ea39c-7d76-4fab-9a70-60f61f801fdf","order_by":3,"name":"Koichi Murashita","email":"","orcid":"","institution":"Hirosaki University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Koichi","middleName":"","lastName":"Murashita","suffix":""},{"id":4840601,"identity":"33ba5805-1dde-45c3-b143-4fe062248379","order_by":4,"name":"Ken Itoh","email":"","orcid":"","institution":"Center for Advanced Medical Research","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Ken","middleName":"","lastName":"Itoh","suffix":""},{"id":4840602,"identity":"8b804fd2-6665-4cd3-80e6-72d7bc3fe518","order_by":5,"name":"Shigeyuki Nakaji","email":"","orcid":"https://orcid.org/0000-0001-8445-8606","institution":"Hirosaki University Graduate School of Medicine","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Shigeyuki","middleName":"","lastName":"Nakaji","suffix":""},{"id":4840603,"identity":"57f43396-63cf-43ce-98b9-754b4c663cdf","order_by":6,"name":"Yasushi Okuno","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABLElEQVRIie2RMUvEMBSA3y09h2jWBNH+hZRCDxHxr6QUrks3QTpJQahL5NZu9xfqcjhWCp1yuN4YOeh0w7m5nbGcWmiLjoL5II+XhC/vJQEwGP4i6Du1gOuBUXub/KwcARU6LX6tAJwCk22lB4ykp7ZQ2nYW1kqBZ7tSOuttDH4yLhWcPXYUei8mTgalk6+iCeMwdRZL4bJCagVNGVDZUdgz8o4RlKOcRBbhUHFPr5CndOcnEOkj00Hlcq4b08qOu7NG0VXwpl9ZikbxkxX3tGJxdij2CumvQkV1TTMWBrncaIVZTiarKyYluCmpWdFzF4yCBdnG5xfzu7Cmb7FlYxE8qDiGkxkOXta0+2L79m4/k49w0MTmm0qaDChw056M1Vc6eh1UDAaD4f/wDuQsZFurViCzAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-3596-4208","institution":"Kyoto University","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Yasushi","middleName":"","lastName":"Okuno","suffix":""}],"badges":[],"createdAt":"2020-11-13 12:25:53","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-107700/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-107700/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41467-021-23319-1","type":"published","date":"2021-05-25T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":3745455,"identity":"7f24b0d1-61bf-4165-8843-a0e6c8724a53","added_by":"auto","created_at":"2020-11-21 17:38:40","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":210039,"visible":true,"origin":"","legend":"Schematic representation of the framework for planning actionable paths for treatment using\nhierarchical Bayesian modeling. The framework consists of three steps. A schematic is given as an example in\nwhich a path is planned to improve the systolic blood pressure (SBP) owing to changes in blood data and body\ncomposition data. a Construction of a regression model from the dataset. A variable SBP is set as the response\nvariable in this case. b Construction of a stochastic surrogate model based on the original dataset and the predicted\nvalues of the regression model. This figure shows a schematic representation of a two-variable space regarding blood\nglucose and body mass index (BMI). The heatmap and vertical axis represent the existence probability of data in the\nvariable space, which is expressed by the stochastic surrogate model. c, d Actionable path planning is applied to\nimprove the response variable. The path is represented as a set of multistep transitions on explanatory variables. In\nour framework, the optimal path (green line in (c)) is planned on the grid graph with high probabilities in the\nvariable space based on the stochastic surrogate model. Conversely, the nonoptimal path (red line in (c)) may pass\nthrough nodes with low or zero probability.","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1/7a32e84fe0a84eee832029dd.png"},{"id":3745457,"identity":"951e4a7b-7247-4cfe-918e-b6216db54e7b","added_by":"auto","created_at":"2020-11-21 17:38:41","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":108370,"visible":true,"origin":"","legend":"Graphical model representation of stochastic surrogate model. Nodes in the graphical model are\nrepresented as follows: 𝒙cont, continuous explanatory variables; 𝒙disc, discrete explanatory variables; 𝑦, response\nvariable predicted by the regression model; 𝒛, the parameter of the mixture components; and all the others, prior\ndistributions. 𝑘 represents each mixture component, and 𝚺( is a diagonal matrix with elements according to the\nCauchy distribution. The symbol 𝑅𝑀𝑆𝐸test in the equation represents a root-mean-squared error of the regression\nmodel, and 𝑦mean and 𝑦std represent the mean and standard deviation values of the predicted response variable,\nrespectively.","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1/ece8858c54342a4898b2a197.png"},{"id":3745458,"identity":"a2ed394b-8475-433d-a5aa-0e2d49aaa981","added_by":"auto","created_at":"2020-11-21 17:38:41","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":224086,"visible":true,"origin":"","legend":"Pseudocode of path search algorithm.","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1/87c3241d1437b7d65faa8ae0.png"},{"id":3745459,"identity":"fb684b20-6b3c-474f-8b74-3fa7cdd51950","added_by":"auto","created_at":"2020-11-21 17:38:41","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":280702,"visible":true,"origin":"","legend":"Examples of actionable paths planned on synthetic dataset. The optimal paths for improving the\nresponse variable predicted by the ML model are represented for randomly selected two examples: instance 1 (a–c)\nand instance 2 (d-f). a, d The orders of changes in the explanatory variables in the optimal path and the\naccompanying changes in the predicted values. In the transition steps, the upward or downward arrow represents a\nunit increase or decrease in the explanatory variable, respectively. b, e Two-dimensional (2D) plots of the path. The\n2D plots are shown regarding the selected two variables: X1 and X2 (b), and X2 and X3 (e). c, f Three-dimensional\n(3D) plots of the path.","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1/7c073c5eb8fdfdcd3610cb1e.png"},{"id":3745460,"identity":"b35b1201-1060-437a-aa0d-d9929f02da57","added_by":"auto","created_at":"2020-11-21 17:38:41","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":109271,"visible":true,"origin":"","legend":"Results of proposed framework on the Iwaki Health Promotion Project (IHPP) dataset. a Mean\nfeature importance: these 25 features were selected by recursive feature elimination (RFE) to predict the systolic\nblood pressure (SBP). RFE was performed with five-fold cross-validation, and the importance of all the mean\nfeatures was calculated when 25 variables remained. The color of each bar represents the following: red: intervention\nvariables in path planning, gray: variables which cannot be easily intervened, and blue: other variables. Details of\nfeatures are described in Supplementary Table 1. b Plot for prediction vs. true response variable. c Widely applicable\nBayesian information criterion (WBIC) values of stochastic surrogate models with 1–8 mixture components. d\nHistogram of actionability scores at different instances. An actionability score of zero indicates that the actionability\nof the optimal path is equivalent to that of the baseline path. Only five instances scored over seven, with a maximum\nscore \u003e14. Scores of these instances were summarized in a score of seven to adjust the plot appearance.","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1/f95e1766df64ccb56a3772c7.png"},{"id":3745461,"identity":"c9b956d0-23cb-4a44-bf75-61f71f6ce5c7","added_by":"auto","created_at":"2020-11-21 17:38:41","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":159883,"visible":true,"origin":"","legend":"Examples of personal actionable paths for treatment using the IHPP dataset. The optimal paths for\nimproving the response variable predicted by the ML model are represented for randomly selected three examples:\ninstance 1 (a, b), instance 2 (c, d), and instance 3 (e, f). a, c, e The orders of changes in the explanatory variables in\nthe optimal path and the accompanying changes in the predicted values. In the transition steps, the upward or\ndownward arrow represents a unit increase or decrease in the explanatory variable, respectively. b, d, f 2D plots of\nthe path. The 2D plots are shown regarding the two influential variables in the optimal path: blood glucose and leg\nscore (b), leg score and g-GTP (d), and blood glucose and leg score (f). 3D plots of the path are shown in\nSupplementary Fig. 4.","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1/008c593a739426fa1ebfe54e.png"},{"id":15782574,"identity":"1456c7fb-2e2b-4e26-aaf5-c25ab4fb2137","added_by":"auto","created_at":"2021-11-22 15:48:22","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5205543,"visible":true,"origin":"","legend":"","description":"","filename":"MainManuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1_covered.pdf"},{"id":13557820,"identity":"3f28cc6a-def0-4f5e-9887-929d46b9bcb4","added_by":"auto","created_at":"2021-09-17 02:54:01","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5200656,"visible":true,"origin":"","legend":"","description":"","filename":"MainManuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1_covered.pdf"},{"id":3745462,"identity":"fce0e04c-4ef2-4662-bb20-b297a948c227","added_by":"auto","created_at":"2020-11-21 17:38:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":7282245,"visible":true,"origin":"","legend":"","description":"","filename":"MainManuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1_stamped.pdf"},{"id":3745456,"identity":"cb4b948a-50e5-4267-a6cf-ff050936c469","added_by":"auto","created_at":"2020-11-21 17:38:41","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":6187638,"visible":true,"origin":"","legend":"Supplementary Information","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-107700/v1/43b9e231d86836a601723c1d.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Health improvement framework for planning actionable treatment process using surrogate Bayesian model","fulltext":[{"header":"Full Text","content":"Due to technical limitations, full-text HTML conversion of this manuscript could not be completed. However, the latest manuscript can be downloaded and \u003ca href='/article/rs-107700/latest.pdf' target='_blank'\u003e accessed as a PDF.\u003c/a\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Clinical Decision Making, Objective Treatment Processes, Computed Treatment Processes","lastPublishedDoi":"10.21203/rs.3.rs-107700/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-107700/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Clinical decision making regarding treatments based on personal characteristics leads to effective health improvements. Machine learning (ML) has been the primary concern of diagnosis support according to comprehensive patient information. However, the remaining prominent issue is the development of objective treatment processes in clinical situations. This study proposes a novel framework to plan treatment processes in a data-driven manner. A key point of the framework is the evaluation of the \"actionability\" for personal health improvements by using a surrogate Bayesian model in addition to a high-performance nonlinear ML model. We first evaluated the framework from the viewpoint of its methodology using a synthetic dataset. Subsequently, the framework was applied to an actual health checkup dataset comprising data from 3,132 participants, to improve systolic blood pressure values at the individual level. We confirmed that the computed treatment processes are actionable and consistent with clinical knowledge for lowering blood pressure. These results demonstrate that our framework could contribute toward decision making in the medical field, providing clinicians with deeper insights.","manuscriptTitle":"Health improvement framework for planning actionable treatment process using surrogate Bayesian model","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2020-11-21 17:38:39","doi":"10.21203/rs.3.rs-107700/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"837dd361-ab04-44ef-902f-a7a438d1c30d","owner":[],"postedDate":"November 21st, 2020","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":1160594,"name":"Artificial Intelligence and Machine Learning"},{"id":1160595,"name":"Preventive Medicine"}],"tags":[],"updatedAt":"2021-11-22T15:42:26+00:00","versionOfRecord":{"articleIdentity":"rs-107700","link":"https://doi.org/10.1038/s41467-021-23319-1","journal":{"identity":"nature-communications","isVorOnly":false,"title":"Nature Communications"},"publishedOn":"2021-05-25 04:00:00","publishedOnDateReadable":"May 25th, 2021"},"versionCreatedAt":"2020-11-21 17:38:39","video":"","vorDoi":"10.1038/s41467-021-23319-1","vorDoiUrl":"https://doi.org/10.1038/s41467-021-23319-1","workflowStages":[]},"version":"v1","identity":"rs-107700","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-107700","identity":"rs-107700","version":["v1"]},"buildId":"WrCJVZZCHTDjtuVLN7oU0","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}