Show simple item record

dc.contributor.advisorRichard R. Fletcher and Peter Szolovits.en_US
dc.contributor.authorChamberlain, Danielen_US
dc.contributor.otherMassachusetts Institute of Technology. Department of Electrical Engineering and Computer Science.en_US
dc.date.accessioned2017-05-11T19:57:48Z
dc.date.available2017-05-11T19:57:48Z
dc.date.copyright2017en_US
dc.date.issued2017en_US
dc.identifier.urihttp://hdl.handle.net/1721.1/108957
dc.descriptionThesis: S.M. in Technology and Policy, Massachusetts Institute of Technology, School of Engineering, Institute for Data, Systems, and Society, Technology and Policy Program, 2017.en_US
dc.descriptionThesis: S.M., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017.en_US
dc.descriptionCataloged from PDF version of thesis.en_US
dc.descriptionIncludes bibliographical references (pages 107-112).en_US
dc.description.abstractPulmonary diseases are responsible for more than 15% of deaths worldwide. Much of this burden is concentrated in the developing world, where these diseases cause 19% of deaths. In much of the developing world, pulmonary diseases are under-diagnosed and misdiagnosed because the correct equipment is not available or health care is provided by workers with insufficient training. To help improve the diagnosis of pulmonary disease, we built a pulmonary diagnostic kit that consists of an electronic stethoscope, an augmented reality peak flow meter, and an electronic questionnaire. Using this kit, we collected data from patients who visited the Chest Research Foundation, a pulmonary clinic in Pune, India. Using the data collected from these patients, we pursued several of avenues of research. First, we trained algorithms to automatically detect two adventitious breath sounds: wheezes and crackles. We used two approaches to detect these sounds: traditional signal processing methods and new techniques from deep semi-supervised learning. Both techniques showed moderate success at identifying wheezes and crackles. Second, we evaluated the diagnostic potential of detecting wheezes and crackles and compared it to using signal processing analysis of lung sounds to directly detect pulmonary disease. We showed that this new technique leads to improved diagnostic accuracy. This finding indicates that future research should focus less on lung sound identification. Third, we combined measurements from all three components of our kit to predict the diagnosis of patients with pulmonary disease. We showed that most of the diagnostic accuracy of the kit was provided by the peak flow meter and questionnaire combination. Together, these two devices were able to accurately detect patients with asthma and COPD. After developing the diagnostic algorithms, we built an Android application to guide a user through the necessary data collection to arrive at a diagnosis. The application was designed to create questionnaires and data queries from an externally defined model definition file, allowing the application to be easily repurposed for different classification tasks in medicine and other fields. Future research will expand the use of the pulmonary diagnostic kit to include additional pulmonary diseases and will test its use in a large-scale field study to determine its accuracy as a screening tool for asthma and COPD. If the results of future trials are consistent with the findings in this thesis, the kit and algorithm combination may provide useful information for improving diagnosis of pulmonary disease.en_US
dc.description.statementofresponsibilityby Daniel Chamberlain.en_US
dc.format.extent113 pagesen_US
dc.language.isoengen_US
dc.publisherMassachusetts Institute of Technologyen_US
dc.rightsMIT theses are protected by copyright. They may be viewed, downloaded, or printed from this source but further reproduction or distribution in any format is prohibited without written permission.en_US
dc.rights.urihttp://dspace.mit.edu/handle/1721.1/7582en_US
dc.subjectInstitute for Data, Systems, and Society.en_US
dc.subjectTechnology and Policy Program.en_US
dc.subjectElectrical Engineering and Computer Science.en_US
dc.titleDesign and validation of mobile kit and machine learning algorithms for pulmonary disease screening and diagnosisen_US
dc.typeThesisen_US
dc.description.degreeS.M. in Technology and Policyen_US
dc.description.degreeS.M.en_US
dc.contributor.departmentMassachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
dc.contributor.departmentMassachusetts Institute of Technology. Engineering Systems Division
dc.contributor.departmentMassachusetts Institute of Technology. Institute for Data, Systems, and Society
dc.contributor.departmentMassachusetts Institute of Technology. Technology and Policy Program
dc.identifier.oclc986481917en_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record