Publishing Peer-Reviewed �Computable Biomedical Knowledge Artifacts

GÜNEŞ KORU -- HTTPS://DRKORU.US

GKORU@UAMS.EDU OR GK@DRKORU.US

UNIVERSITY OF ARKANSAS FOR MEDICAL SCIENCES

ASSOCIATE EDITOR FOR CBK TRACK, LEARNING HEALTH SYSTEMS JOURNAL

​

​

Thanks

• All Attendees, committee members, and presenters

​

• MCBK 2024 NA Chairs
• Kristi Holmes & Josh Richardson

​

• Organizers
• Noor Khan and Janelle Burleigh

​

• And, of course, Chuck

We have started publishing CBKs!

We started the CBK track with in late 2021 with initial rules and made timely additions and revisions.

See LHS Special Issue 2023

CBK Artifact��by �DALL-E 3

What is a CBK artifact?

• We use CBK to refer to Computable Biomedical Knowledge Artifacts
• A CBK is a mobilizable software program
• Not simply static data, documents, images, or videos.

​

​

Observable Behavior

CBKs embody behavior that can be used to transform specific inputs into observable outputs expected according to the study goals.

• Behavior can be observed through execution of CBK on any hardware/software layer.
• CBKs include
• Web services, browser apps, mobile apps, etc.
• Scripting languages (Python, R, Perl, Tcl, shell, etc.), systems programming languages (Java, C, C++), Assembly, or Machine language can be used

​

Publishable CBKs and Decomposition

• A CBK can be decomposed into multiple dynamic or static objects/modules

​

• Can be constructed by reusing and composing multiple objects/modules or CBKs

​

• Your study purpose and what behavior you want to be reviewed will determine the borders/granularity level

​

Why Publish Peer-Reviewed CBKs?

Toward answering critical questions about CBKs for a robust biomedical research enterprise

A note on openness

• We encourage releasing source code
• Intellectual property, licensing, and confidentiality concerns are also realities
• If submitters have sufficient IP permissions, exposing and archiving behavior can be possible, and it can still support research and innovation
• The above option can be an effective marketing tool for those in the industry
• Regardless, peer reviews are essential
• They were found useful in many empirical software engineering studies

Potential Benefits

Increasing CBKs’ . . .

• Tangibility
• Reproducibility
• Reuse
• Trust
• Recognition

Will benefit . . .

• Scientists
• Research Funding Agencies
• Governments
• Industry
• Patients and Families

Experiences with publishing CBKs so far

• We were fortunate to have a motivated community making submissions
• Many of those who submit convincingly demonstrated the behavior of their CBKs
• Some couldn’t do so
• Either had to be rejected with feedback useful for another time
• We worked with a couple of submissions to resolve issues
• Some submitters might still not believe we give their CBK a good run, but we do.

​

Experiences so far…

• The variety of languages and platforms can be a challenge we have to deal with
• Container technologies work
• Manuscript served as documentation
• So far, metadata has not been collected other than Journal issued metadata, but we would like to do so
• Interestingly, video recording worked well – we should consider

​

Opportunities and Challenges Ahead

• Through publishing CBKs, we also developed some ideas about the challenges researchers might experience in publishing CBKs and potential solutions
• We want to discuss them with you.

​

THANKS FOR LISTENING

VERY INTERESTED IN LEARNING FROM YOU

​

BACK TO CHUCK