Mission & Vision

Background:

New knowledge in healthcare and biomedicine is being generated is at an unprecedented rate. The average time from publication to practice averages 17 years. To facilitate the translation of new evidence and best practice to ensure the optimal outcomes and reduce waste and inefficiency the status quo of the current knowledge ecosystem is primed for change. Machine readable, interpretable, and executable knowledge may be one piece of the solution. While a robust ecosystem for sharing narrative and human readable biomedical knowledge exists, much work is to be done to build and ecosystem of computable biomedical knowledge and the infrastructure to facilitate it.

Computable biomedical knowledge (CBK) is the result of an analytic and/or deliberative process about human health, or affecting human health, that is explicit and, therefore, can be represented and reasoned upon using logic, formal standards, and mathematical approaches (MCBK Manifesto, 2018). Examples of CBK include but are not limited to, predictive risk models, computable phenotypes, clinical rule sets, biomedical ontologies, business process models, and machine learning models. This growing utilization of CBK introduces many challenges related to development, dissemination, implementation, and stewardship for knowledge engineers, implementers, and stewards.

CBK and the Learning Health System

The Knowledge Systems Laboratory is exploring the challenges in knowledge management and dissemination through a learning health system (LHS) lens. LHSs are multi-stakeholder communities in healthcare that undertake cyclical activities of generating new knowledge, seamlessly integrating that knowledge to improve health, and evaluating that knowledge in action. For example, an LHS may develop a new risk-predictive model for lung cancer screening (data-to-knowledge). This model represents a form of CBK. Stakeholders in an LHS then collaborate to translate that knowledge to practice and evaluate its implementation for further revision to improve human health. As ecosystems of CBK become more robust and complicated, the ability of health systems, researchers, developers, and implementers of CBK to appropriately manage and steward their knowledge increases.



Packaging CBK as Knowledge Objects

Central to the Knowledge System Laboratory's work is our knowledge object (KO) conceptual model (see figure below).  A knowledge object represents computable knowledge as both a resource and a service. It contains descriptions of what it does, sufficient metadata to ensure it is findable, accessible, interoperable, and reusable, and the code to do it. In theory, we argue any type of computable biomedical knowledge can be packaged as a knowledge object. This may include predictive risk models, clinical rule sets, machine learning models, etc. 

The work of the Knowledge Systems Laboratory to date has focused on further developing this conceptual model by:

Learn more about our previous published work here.

Learn more about our knowledge object conceptual model here. 

Figure: Contents of a knowledge object (KO). A conceptual framework for knowledge objects as fully described in the Knowledge Object Implementation Ontology (KOIO). A KO should include a core CBK model, appropriate metadata, and a persistent identifier.

Our Vision:

We envision a high-functioning open ecosystem of computable knowledge:

... where computable biomedical knowledge is readily accessible, interoperable, implementable;

...  anyone can easily find, access, deploy, and (re)use CBK artifacts thanks to widely available infrastructure;

... this infrastructure enables an ongoing second knowledge revolution: augmenting mass access to knowledge with the capabilities for mass action that           brings knowledge to practice at scale;

.   ... and where health care, personal health management, population health, biomedical research, and health professions education are all enhanced       through mass application of CBK.

To achieve our vision, the Knowledge Systems Lab researches computable biomedical knowledge and the infrastructure that supports it. Our work addresses four fundamental interests related to CBK:


We seek to understand the nature and optimal means of engineering computable knowledge artifacts, especially CBKs. Additionally, we study issues related to computable knowledge representation and integration. This includes the readiness of computable knowledge for inclusion in knowledgebases supporting software applications and organizational readiness for integrating computable knowledge into new and existing information systems.


Infrastructure impacts the creation, application, dissemination, implementation, and stewardship of computable knowledge. We seek to understand the evolution of infrastructure that supports computable knowledge in many forms through a complete lifecycle and facilitates use for diverse communities and multiple projects at many scale levels. 


We study the principles and practices of knowledge management related to computable knowledge and how to steward CBK through its lifecycle. This includes structural, administrative, and descriptive metadata and mechanisms by which knowledge creators can communicate and knowledge consumers can assess the fitness of the knowledge. 


How to leverage the value proposition of computable knowledge is an open question. This involves a deep understanding of user needs and uses cases and the readiness of users and organizations to adopt, integrate, and manage computable knowledge.