Automatic Formalization of Clinical Practice Guidelines Matthew S. Gerber and Donald E. Brown Department of Systems and Information Engineering University.

Slides:



Advertisements
Similar presentations
1 Using Ontologies in Clinical Decision Support Applications Samson W. Tu Stanford Medical Informatics Stanford University.
Advertisements

January 12-13, 2006 Montpelier, VT Chronic Care Management for all Vermonters Kenneth E. Thorpe, Ph.D. Robert W. Woodruff Professor and Chair Department.
ICGP Professional Competence System How to complete the cycle.
Clinical decision support system (CDSS). Knowledge-based systems Knowledge based systems are artificial intelligent tools working in a narrow domain to.
MOLEDINA-1 CSE 5810 CSE5810: Intro to Biomedical Informatics The Role of AI in Clinical Decision Support Saahil Moledina University of Connecticut
FACTORS HINDERING ATTITUDE TO TREATMENT AMONG PATIENTS WITH TYPE-2 DIABETES MELLITUS IN THE NIGER DELTA, NIGERIA by AGOFURE OTOVWE and OYEWOLE OYEDIRAN.
Utilizing Evidence Based Practice in the Acute Care Clinical Setting Brenda P. Johnson, PhD, RN Department of Nursing Southeast Missouri State University.
Overview of Biomedical Informatics Rakesh Nagarajan.
Biomedical Informatics Some Observations on Clinical Data Representation in EHRs Christopher G. Chute, MD DrPH, Mayo Clinic Chair, ICD11 Revision, World.
Guideline implementation Types of CDSS A.Hasman. Do physicians need support? In 2.3% of the 1.3 million patients ( patients) preventable errors.
Guideline interaction scenarios  At the point of care Physicians apply marked-up guidelines, thus they Need to find an appropriate guideline in “ real.
A Multiple-Ontology Template-Based Query Interface for a Clinical Guidelines Search Engine Robert Moskovitch, Talie Lavie, Akiva Leibowitz, Yaron Denekamp.
Seven Lectures on Statistical Parsing Christopher Manning LSA Linguistic Institute 2007 LSA 354 Lecture 7.
Cohort Studies Hanna E. Bloomfield, MD, MPH Professor of Medicine Associate Chief of Staff, Research Minneapolis VA Medical Center.
APPLICATION : DIAGNOSTIC CODING 1 SIEMENS  Coding is the translation of diagnosis terms describing patients diagnosis or treatment into a coded number.
Critical Appraisal of Clinical Practice Guidelines
Decision Support for Quality Improvement
 Definitions  Goals of automation in pharmacy  Advantages/disadvantages of automation  Application of automation to the medication use process  Clinical.
For Evidence-based Practice Information Retrieval for Evidence-based Practice Fall 2001 Suzanne Bakken, RN, DNSc, FAAN School of Nursing & Department of.
Estimate of Certainty (Precision) of Treatment Effect Level of evidence of B or C does not imply that recommendation is weak. LEVEL A Multiple populations.
Healthcare Services as Collective Activity Susan Wakenshaw Xiao MA.
Learning Object Metadata Mining Masoud Makrehchi Supervisor: Prof. Mohamed Kamel.
Systolic hypertension not an isolated problem Michael Weber, MD Professor of Medicine Associate Dean Downstate College of Medicine State University of.
Computers in Healthcare Jinbo Bi Department of Computer Science and Engineering Connecticut Institute for Clinical and Translational Research University.
Systematic Reviews.
Automatic Detection of Tags for Political Blogs Khairun-nisa Hassanali and Vasileios Hatzivassiloglou Human Language Technology Research Institute The.
A Multiple Ontology, Concept-Based, Context-Sensitive Search and Retrieval Robert Moskovitch and Prof. Yuval Shahar Medical Informatics Research Center.
Using Text Mining and Natural Language Processing for Health Care Claims Processing Cihan ÜNAL
Introduction to Systematic Reviews Afshin Ostovar Bushehr University of Medical Sciences Bushehr, /9/20151.
 Text Representation & Text Classification for Intelligent Information Retrieval Ning Yu School of Library and Information Science Indiana University.
Evidence-Based Public Health Nancy Allee, MLS, MPH University of Michigan November 6, 2004.
Combining terminology resources and statistical methods for entity recognition: an evaluation Angus Roberts, Robert Gaizauskas, Mark Hepple, Yikun Guo.
Copyright 2006, Ida Sim Ida Sim, MD, PhD Associate Professor of Medicine Associate Director for Medical Informatics Program in Biological and Medical Informatics.
Finding Relevant Evidence
Plymouth Health Community NICE Guidance Implementation Group Workshop Two: Debriding agents and specialist wound care clinics. Pressure ulcer risk assessment.
Correlating Knowledge Using NLP: Relationships between the concepts of blood cancers, stem cell transplantation, and biomarkers Katy Zou and Weizhong Zhu.
1 Incorporating Data Mining Applications into Clinical Guidelines Reza Sherafat Dr. Kamran Sartipi Department of Computing and Software McMaster University,
Indirect Supervision Protocols for Learning in Natural Language Processing II. Learning by Inventing Binary Labels This work is supported by DARPA funding.
Evidence Based Practice RCS /9/05. Definitions  Rosenthal and Donald (1996) defined evidence-based medicine as a process of turning clinical problems.
Introduction to Healthcare and Public Health in the US The Evolution and Reform of Healthcare in the US Lecture b This material (Comp1_Unit9b) was developed.
From the Advanced Search page of the Cochrane Library, we have clicked on the Cochrane Reviews: By Topic hyperlink. This has displayed the Topics for Cochrane.
Medical Information Retrieval: eEvidence System By Zhao Jin Mar
This material was developed by Duke University, funded by the Department of Health and Human Services, Office of the National Coordinator for Health Information.
Systematic Approaches to Literature Reviewing Dr Tamara O’Connor Student Learning Development
Evaluating the Medical Evidence ​ A TOOLKIT FOR THE INTERPRETING THE EFFECTIVENESS OF INTERVENTIONS Niteesh Choudhy, M.D., Ph.D.
Journal Club Alcohol, Other Drugs, and Health: Current Evidence November-December 2012.
Virtual Examples for Text Classification with Support Vector Machines Manabu Sassano Proceedings of the 2003 Conference on Emprical Methods in Natural.
SAGE Nick Beard Vice President, IDX Systems Corp..
Table of Contents – Part B HINARI Resources –Clinical Evidence –Cochrane Library –EBM Guidelines –BMJ Practice –HINARI EBM Journals.
Informatics and Evidence-based Practice M8120 Fall 2001 Suzanne Bakken, RN, DNSc, FAAN School of Nursing & Department of Medical Informatics Columbia University.
The US Preventive Services Task Force: Potential Impact on Medicare Coverage Ned Calonge, MD, MPH Chair, USPSTF.
FROM ONE NOMENCLATURES TO ANOTHER… Drs. Sven Van Laere.
N VISUAL ANALYTICS FOR HEALTHCARE: BIG DATA, BIG DECISIONS David Gotz Healthcare Analytics Research Group IBM T.J. Watson Research Center.
Introduction to General Epidemiology (2) By: Dr. Khalid El Tohami.
Mobile Technology Improves Patient Outcomes JULIE POPE COLUMBUS STATE UNIVERSITY.
Conference on Medical Thinking University College London June 23, 2006 Medical Thinking: What Should We Do? Edward H. Shortliffe, MD, PhD Department of.
Introduction to Health Informatics Leon Geffen MBChB MCFP(SA)
Assessing SNOMED CT for Large Scale eHealth Deployments in the EU Workpackage 2- Building new Evidence Daniel Karlsson, Linköping University Stefan Schulz,
Using Technology to Support Evidence Based Practice
1st International Online BioMedical Conference (IOBMC 2015)
Evaluating Sepsis Guidelines and Patient Outcomes
Quality Health Care Nursing 870
Strategies to incorporate pharmacoeconomics into pharmacotherapy
Component 11/Unit 7 Implementing Clinical Decision Support
A Multiple-Ontology Template-Based Query Interface for a Clinical Guidelines Search Engine Robert Moskovitch, Talie Lavie, Akiva Leibowitz, Yaron Denekamp.
Martijn Schuemie, Peter Rijnbeek, Jenna Reps, Marc Suchard
Usability Techniques Lecture 13.
Table of Contents Why Do We Treat Hypertension? Recommendation 5
By Hossein Hematialam and Wlodek Zadrozny Presented by
Table of Contents – Part B
Presentation transcript:

Automatic Formalization of Clinical Practice Guidelines Matthew S. Gerber and Donald E. Brown Department of Systems and Information Engineering University of Virginia James H. Harrison Department of Public Health Sciences University of Virginia

Many treatment options – what to do? Clinical Practice Guidelines 2 Benefits / costs Evidence quality Randomized clinical trial: beneficial Expert opinion: might be beneficial Meta-analysis: usually beneficial Strength Recommended Should consider Might consider

Clinical Practice Guidelines Development –Expert synthesis of current evidence –Example from heart failure: 3

Clinical Practice Guidelines Expected outcomes –Evidence-based clinical decision aid –Reduction in cost and treatment/outcome variation –Improvement in patient health Challenges –A guideline for any occasion –Guidelines change periodically –Lengthy (HFSA CPG is 259 pages) 4 Total (NGC)2,269 Cardiovascular diseases486 Heart failure152

Clinical Decision Support Systems Goal: deliver CPG knowledge at point of care Alleviate burden on clinician Problem: CPGs contain minimally structured text 5 Formalization is required

Traditional CPG Formalization 6 Knowledge engineers Medical experts Knowledge representation Knowledge management software (e.g., Protégé) CDSS CPG Automatic formalization

… Retrospective analyses The Big Picture 7 Structured knowledge Medical decision support ? Endocrine Infections … Cardiovascular NLP

Data Collection Yale Guideline Recommendation Corpus –Hussain et al. (2009) –1,275 recommendations –Representative sample of domains and rec. types “Oral antiviral drugs are indicated within 5 days of the start of the episode and while new lesions are still forming.” –Simplifications Delimited recommendations No inter-recommendation dependencies Random sub-sample of YGRC (n=200) 8

Recommendation Representation SNOMED-CT –Medical concept ontology –Broad coverage 9 Keywords ? Asbru, etc. Fidelity: Low High Automation: Trivial Impossible

Recommendation Representation 10 (Sundvalls et al., 2012)

Recommendation Representation 11 PrimarySecondaryExample recommendation ACTIONEVALUATION[computed tomography CT] should be used PROCEDURE[red blood cell transfusion] is appropriate … EVIDENCESTRONG[it has been shown to reduce the occurrence of NTDs] WEAK/NONE[there is insufficient evidence] MODALITYOBLIGATORYcomputed tomography CT [should] be used NEVERoral risedronate [should not] be used OPTIONALphysician [may] choose AGENT[physician] may choose MORBIDITYprevent [preeclampsia] POPULATION[obese women with gestational diabetes mellitus] PURPOSEis used [to prevent osteoporotic fractures] TEMPORAL[initial] treatment SNOMED-CT CONCEPT:

Recommendation Annotation Task: manually identify representational elements within recommendations Example Diuretics are recommended for patients with heart failure. [ DRUG Diuretics] are recommended for [ POPULATION patients with [ MORBIDITY heart failure]]. 12

Methods Natural language processing Supervised classification Per-recommendation pipeline 1.Syntactic parsing 2.Parse node classification 3.Post-processing 13

Methods: (1) Syntactic Parsing 14 Constituency parser (Charniak and Johnson, 2005)

Methods: (2) Parse Node Classification 15 Unit of classification: node Multi-class logistic regression Example: 1 positive, 17 negative Actual –12K nodes –10 classes (primary)

Methods: (2) Parse Node Classification 16 Linguistic features –Word stems under node –Syntactic configuration of node –…

Methods: (2) Parse Node Classification 17 Learning –Forward feature selection –Per-class costs (LibLinear)

Methods: (3) Post-processing 18 Remove duplicates Other possible issues –Conflicts –Embedding

Evaluation Results 19 Element#Precision (%)Recall (%)F1 (%) ACTION MODALITY POPULATION TEMPORAL TRIGGER PURPOSE EVIDENCE AGENT MORBIDITY Overall fold cross-validation

Discussion High variance across classes Alternative strategies –Identify more informative features –Change the model formulation –Annotate more data 20

Conclusions CPGs are an important knowledge source Difficult to use within CDSS Prior CPG formalization –Manual –Automatic for specific domains / recommendations Our contributions –SNOMED-CT representation –Manually annotated recommendation sample –Statistical NLP model / evaluation 21

Future Work Refined representation Model formulation Feature engineering Controlled natural language 22

Questions? References 1.Charniak, E. & Johnson, M. Coarse-to-fine n-best parsing and MaxEnt discriminative reranking. Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics, 2005, Hussain, T.; Michel, G. & Shiffman, R. N. The Yale Guideline Recommendation Corpus: A representative sample of the knowledge content of guidelines. I. J. Medical Informatics, 2009, 78, Fan, R.-E.; Chang, K.-W.; Hsieh, C.-J.; Wang, X.-R. & Lin, C.-J. LIBLINEAR: A Library for Large Linear Classification. Journal of Machine Learning Research, 2008, 9, Sundvall, E.; Nystrom, M.; Petersson, H. & Ahlfeldt, H. Interactive visualization and navigation of complex terminology systems, exemplified by SNOMED CT. Studies in health technology and informatics, IOS Press; 1999, 2006, 124,