Show simple item record

The Knowledge Grid: A Platform to Increase the Interoperability of Computable Knowledge and Produce Advice for Health

dc.contributor.authorFlynn, Allen
dc.date.accessioned2018-10-25T17:42:26Z
dc.date.availableNO_RESTRICTION
dc.date.available2018-10-25T17:42:26Z
dc.date.issued2018
dc.date.submitted2018
dc.identifier.urihttps://hdl.handle.net/2027.42/146073
dc.description.abstractHere we demonstrate how more highly interoperable computable knowledge enables systems to generate large quantities of evidence-based advice for health. We first provide a thorough analysis of advice. Then, because advice derives from knowledge, we turn our focus to computable, i.e., machine-interpretable, forms for knowledge. We consider how computable knowledge plays dual roles as a resource conveying content and as an advice enabler. In this latter role, computable knowledge is combined with data about a decision situation to generate advice targeted at the pending decision. We distinguish between two types of automated services. When a computer system provides computable knowledge, we say that it provides a knowledge service. When computer system combines computable knowledge with instance data to provide advice that is specific to an unmade decision we say that it provides an advice-giving service. The work here aims to increase the interoperability of computable knowledge to bring about better knowledge services and advice-giving services for health. The primary motivation for this research is the problem of missing or inadequate advice about health topics. The global demand for well-informed health advice far exceeds the global supply. In part to overcome this scarcity, the design and development of Learning Health Systems is being pursued at various levels of scale: local, regional, state, national, and international. Learning Health Systems fuse capabilities to generate new computable biomedical knowledge with other capabilities to rapidly and widely use computable biomedical knowledge to inform health practices and behaviors with advice. To support Learning Health Systems, we believe that knowledge services and advice-giving services have to be more highly interoperable. I use examples of knowledge services and advice-giving services which exclusively support medication use. This is because I am a pharmacist and pharmacy is the biomedical domain that I know. The examples here address the serious problems of medication adherence and prescribing safety. Two empirical studies are shared that demonstrate the potential to address these problems and make improvements by using advice. But primarily we use these examples to demonstrate general and critical differences between stand-alone, unique approaches to handling computable biomedical knowledge, which make it useful for one system, and common, more highly interoperable approaches, which can make it useful for many heterogeneous systems. Three aspects of computable knowledge interoperability are addressed: modularity, identity, and updateability. We demonstrate that instances of computable knowledge, and related instances of knowledge services and advice-giving services, can be modularized. We also demonstrate the utility of uniquely identifying modular instances of computable knowledge. Finally, we build on the computing concept of pipelining to demonstrate how computable knowledge modules can automatically be updated and rapidly deployed. Our work is supported by a fledgling technical knowledge infrastructure platform called the Knowledge Grid. It includes formally specified compound digital objects called Knowledge Objects, a conventional digital Library that serves as a Knowledge Object repository, and an Activator that provides an application programming interface (API) for computable knowledge. The Library component provides knowledge services. The Activator component provides both knowledge services and advice-giving services. In conclusion, by increasing the interoperability of computable biomedical knowledge using the Knowledge Grid, we demonstrate new capabilities to generate well-informed health advice at a scale. These new capabilities may ultimately support Learning Health Systems and boost health for large populations of people who would otherwise not receive well-informed health advice.
dc.language.isoen_US
dc.subjectKnowledge organization
dc.subjectComputable biomedical knowledge
dc.subjectKnowledge Grid
dc.subjectMedMinify
dc.subjectScriptNumerate
dc.subjectAdvice Studies
dc.titleThe Knowledge Grid: A Platform to Increase the Interoperability of Computable Knowledge and Produce Advice for Health
dc.typeThesisen_US
dc.description.thesisdegreenamePhDen_US
dc.description.thesisdegreedisciplineInformation
dc.description.thesisdegreegrantorUniversity of Michigan, Horace H. Rackham School of Graduate Studies
dc.contributor.committeememberLagoze, Carl
dc.contributor.committeememberFriedman, Charles P
dc.contributor.committeememberAdler-Milstein, Julia
dc.contributor.committeememberSeagull, Frank Jacob
dc.subject.hlbsecondlevelPharmacy and Pharmacology
dc.subject.hlbsecondlevelInformation and Library Science
dc.subject.hlbtoplevelHealth Sciences
dc.subject.hlbtoplevelSocial Sciences
dc.description.bitstreamurlhttps://deepblue.lib.umich.edu/bitstream/2027.42/146073/1/ajflynn_1.pdf
dc.identifier.orcid0000-0002-3471-3063
dc.identifier.name-orcidFlynn, Allen; 0000-0002-3471-3063en_US
dc.owningcollnameDissertations and Theses (Ph.D. and Master's)


Files in this item

Show simple item record

Remediation of Harmful Language

The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.

Accessibility

If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.