Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Designing Data-Driven Virtual Patients for Health Sciences Education

Babichenko, Dmitriy (2019) Designing Data-Driven Virtual Patients for Health Sciences Education. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

This is the latest version of this item.

Download (19MB) | Preview


Electronic virtual patients (VPs) are interactive screen-based computer simulations of real-life clinical scenarios that are widely used for the purposes of health sciences education. Advances in computational modeling and availability of large patient cohort datasets from Electronic Medical Record (EMR) systems have created an opportunity to develop a new type of VPs, where cases are based on and simulate real patient clinical treatment processes and outcomes.

Traditional VP cases are static narrative representations of clinical scenarios that are presented to health sciences students in order to teach a clinical topic of interest. This research investigates the feasibility of authoring and presenting virtual patient cases that leverage Bayesian network (BN) models learned from EMR data to present clinical scenarios and control outcomes of learners' decisions within the context of a presented VP. Because the underlying models are based on real patient data, each decision made by a learner would affect the probability of each outcome occurring in the same way as with real patients.

Additionally, this dissertation explores the challenges related to using BN models in the context of VP case authoring and presentation, and experimentally compares a VP case based on a BN model to one created using a traditional narrative-branched VP system across multiple categories, including meeting learning objectives, accuracy in depicting the chosen clinical scenario, introducing/reinforcing relevant clinical skills, providing formative feedback, scenario realism, and learner engagement.

Furthermore, this work investigates the extent to which the use of annotated BN models in VP cases facilitates modifying an existing VP case by allowing case authors to manipulate the underlying model in such a way that the modified VP case meets alternate learning objectives.

Last, but not least, this research provides practical and methodological contributions to the body of work in the areas of health sciences education, problem-based learning, and clinical simulation design and evaluation. More specifically, this work (1) defines criteria and guidelines for designing VP cases based on BN models, (2) identifies and describes shortcomings and challenges associated with different BN modeling approaches for different types of clinical scenarios, (3) and presents a framework for evaluating VP cases.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Babichenko, Dmitriydmb72@pitt.edudmb720000-0003-1187-6684
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairDruzdzel, Marekmarek@sis.pitt.edumarek
Committee MemberLin, Yu-RuYURULIN@pitt.eduYURULIN
Committee MemberLewis, Michaelml@sis.pitt.educmlewis0000-0002-1013-9482
Committee MemberMcGee, Jamesjbm1@pitt.edujbm1
Date: 14 January 2019
Date Type: Publication
Defense Date: 21 September 2018
Approval Date: 14 January 2019
Submission Date: 1 November 2018
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 206
Institution: University of Pittsburgh
Schools and Programs: School of Computing and Information > Information Science
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: machine learning, virtual patients, bayesian network models, medical education
Date Deposited: 14 Jan 2019 13:45
Last Modified: 14 Jan 2019 13:45

Available Versions of this Item

  • Designing Data-Driven Virtual Patients for Health Sciences Education. (deposited 14 Jan 2019 13:45) [Currently Displayed]


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item