On the Robustness of Bayesian Network Based Spatial Event Surveillance

Jiang X, Neill DB, Cooper GF. On the robustness of Bayesian network based spatial event surveillance. International Journal of Approximate Reasoning, 51 (2010) p 224-239.  http://dx.doi.org/10.1016/j.ijar.2009.01.001.

Methods for spatial cluster detection attempt to locate spatial subregions of some larger region where the count of some occurrences is higher than expected. Event surveillance consists of monitoring a region in order to detect emerging patterns that are indicative of some event of interest. In spatial event surveillance, we search for emerging patterns in spatial subregions. A well-known method for spatial cluster detection is Kulldorff’s [M. Kulldorff, A spatial scan statistic, Communications in Statistics: Theory and Methods 26 (6) (1997)] spatial scan statistic, which directly analyzes the counts of occurrences in the subregions. Neill et al. [D.B. Neill, A.W. Moore, G.F. Cooper, A Bayesian spatial scan statistic, Advances in Neural Information Processing Systems (NIPS) 18 (2005)] developed a Bayesian spatial scan statistic called BSS, which also directly analyzes the counts. We developed a new Bayesian-network-based spatial scan statistic, called BNetScan, which models the relationships among the events of interest and the observable events using a Bayesian network. BNetScan is an entity-based Bayesian network that models the underlying state and observable variables for each individual in a population. We compared the performance of BNetScan to Kulldorff’s spatial scan statistic and BSS using simulated outbreaks of influenza and cryptosporidiosis injected into real Emergency Department data from Allegheny County, Pennsylvania. It is an open question whether we can obtain acceptable results using a Bayesian network if the probability distributions in the network do not closely reflect reality, and thus, we examined the robustness of BNetScan relative to the probability distributions used to generate the data in the experiments. Our results indicate that BNetScan outperforms the other methods and its performance is robust relative to the probability distribution that is used to generate the data.

Publication Year: 
Publication Credits: 
Jiang X, Neill DB, Cooper GF

Bruce L. Rollman, MD, MPH

Directory Listing Information
Rollman, Bruce
230 McKee Place, Suite 600
Professor of Medicine, Psychiatry, and Clinical and Translational Science

The Feasibility of Automating Audit and Feedback for ART Guideline Adherence in Malawi

Landis Lewis Z, Mello-Thoms C, Gadabu OJ, Gillespie EM, Douglas GP, Crowley RS. The Feasibility of Automating Audit and Feedback for ART Guideline Adherence in Malawi. Accepted to J Am Med Inform Assoc,  (JAMIA) April 19, 2011.

ABSTRACT Objective: To determine the feasibility of using electronic medical record (EMR) data to provide audit and feedback of anti-retroviral therapy (ART) clinical guideline adherence to healthcare workers (HCWs) in Malawi. Materials and methods: We evaluated recommendations from Malawi’s ART guidelines using GuideLine Implementability Appraisal criteria. Recommendations that passed selected criteria were converted into ratio-based performance measures. We queried representative EMR data to determine the feasibility of generating feedback for each performance measure, summed clinical encounters representing each performance measure’s denominator, and then measured the distribution of encounter frequency for individual HCWs across nurse and clinical officer groups. Results: We analyzed 423 831 encounters in the EMR data and generated automated feedback for 21 recommendations (12%) from Malawi’s ART guidelines. We identified 11 nurse recommendations and eight clinical officer recommendations. Individual nurses and clinical officers had an average of 45 and 59 encounters per month, per recommendation, respectively. Another 37 recommendations (21%) would support audit and feedback if additional routine EMR data are captured and temporal constraints are modeled. Discussion: It appears feasible to implement automated guideline adherence feedback that could potentially improve HCW performance and supervision. Feedback reports may support workplace learning by increasing HCWs’ opportunities to reflect on their performance. Conclusion: A moderate number of recommendations from Malawi’s ART guidelines can be used to generate automated guideline adherence feedback using existing EMR data. Further study is needed to determine the receptivity of HCWs to peer comparison feedback and barriers to implementation of automated audit and feedback in low-resource settings.

Publication Year: 
Faculty Author: 
Publication Credits: 
Landis Lewis Z, Mello-Thoms C, Gadabu OJ, Gillespie EM, Douglas GP, Crowley RS

The caBIG(R) Life Science Business Architecture Model

Boyd LB, Hunike-Smith SP, Stafford GA, Freund ET, Elhman M, Chandran U, Dennis R, Fernandez AT, Goldstein S, Steffen D, Tycko B, Klemm JD. The caBIG(R) Life Science Business Architecture Model. Bioinformatics. 2011 May 15;27(10):1429:35. doi: 10.1093/bioinformatics/btr141. Epub 2011 Mar 29. PMID: 21450709. PMCID: PMC3087952.

Motivation: Business Architecture Models (BAMs) describe what a business does, who performs the activities, where and when activities are performed, how activities are accomplished, and which data are present. The purpose of a BAM is to provide a common resource for understanding business functions and requirements and to guide software development. The cancer Biomedical Informatics Grid (caBIG®) Life Science BAM (LS BAM) provides a shared understanding of the vocabulary, goals and processes that are common in the business of LS research. Results: Results: LS BAM 1.1 includes 90 goals and 61 people and groups within Use Case and Activity Unified Modeling Language (UML) Diagrams. Here we report on the model's current release, LS BAM 1.1, its utility and usage, and plans for future use and continuing development for future releases.

Publication Year: 
Faculty Author: 
Publication Credits: 
Boyd LB, Hunike-Smith SP, Stafford GA, Freund ET, Elhman M, Chandran U, Dennis R, Fernandez AT, Goldstein S, Steffen D, Tycko B, Klemm JD
PDF icon ThecaBIG.pdf379.36 KB

The FaceBase Consortium: A Comprehensive Program to Facilitate Craniofacial Research

The FaceBase Consortium: A Comprehensive Program to Facilitate Craniofacial Research. Harry Hochheiser, Bruce J. Aronow, Kristin Artinger, Terri H.Beaty, James F. Brinkley, Yang Chai, David Clouthier, Michael L. Cunningham, Michael Dixon, Leah Rae Donahue, Scott E. Fraser, Junichi Iwata, Mary L. Marazita11, Jeffrey C. Murray, Stephen Murray, John Postlethwait, Steven Potter, Linda Shapiro, Richard Spritz, Axel Visel, Seth M. Weinberg and Paul A. Trainor*, for the FaceBase Consortium.

Publication Year: 
Faculty Author: 
PDF icon DBIO-11-44R1.pdf164.11 KB

Translational Bioinformatics

Translational bioinformatics is an emerging area in informatics focused on “the development of storage, analytic, and interpretive methods to optimize the transformation of increasingly voluminous biomedical data, and genomic data in particular, into proactive, predictive, preventive, and participatory health” (from the American Medical Informatics Association Strategic Plan http://www.amia.org/inside/stratplan/). Within DBMI,  Xinghua Lu, MD, PhDVanathi Gopalakr

Data-Driven Modeling of Usual Clinical Care

Clinical care is complex and often fast paced. Preventable medical errors can and do occur, as has been well documented in recent years. Clinical guidelines and rule-based alerts provide clinical decision support that is intended to reduce medical errors. These methods are driven by expert knowledge. As such, they tend to focus on high impact areas in which medical errors are either prevalent, serious, or both. However, the coverage of such methods is relatively narrow.

Clinical Predictive Modeling

Clinical care involves making many predictions under uncertainty, including risk assessment, diagnosis, prognosis and therapeutic management. The better those predictions can be made, the better clinical care is likely to be. The increasing availability and richness of electronic health records (EHRs) are increasing the opportunities for developing and deploying computer-based clinical prediction methods. Such methods can serve as key components of computer-based decision support systems. The data in EHRs can be used to construct prediction models using machine learning methods.

Health Informatics for the Underserved

Gerald Douglas, PhD is leading research efforts that will contribute to the elimination of health disparities, improve health care quality, encourage the adoption of personal wellness strategies, and provide support for the development and advancement of underserved communities. Dr. Douglas implemented several innovative approaches to using technology in health care, including the use of a low-power, robust and inexpensive touch screen workstation for clinicians at the point of care in Malawi, Africa.


Over many years, Roger S. Day, ScD has studied how computational and modeling tools could help people create better biological understanding, then apply it to better individual treatment decisions.