Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
A Prototype Knowledge Base and
    SMART App to Facilitate
     Organization of Patient
Medications by Clinical Problems

...
Summarization Aims
• Develop methodologies that:
  – Model and summarize complex, chronically-ill
    patients’ EHR data
 ...
Research Focus
• Develop a knowledge base for linking
  medications and problems
• Evaluate the knowledge base with real
 ...
Ontology-Based Summarization
• The Unified Medical Language System
  Metathesaurus (UMLS)
   – RxNorm, SNOMED CT, NDF-RT

...
Ontology-Based Summarization
• Generated over 7 million problem–
  medication links
RxCUI                     Medication  ...
Evaluation Setting
• Large, multi-specialty ambulatory
  academic practice for adults, adolescents
  and children
• Electr...
Evaluation Population
• 5000 randomly selected patients
  • At least one outpatient encounter during
    July 1, 2010-Dece...
Knowledge Base Utility
• Knowledge base linked 5,251 medications
  to problems
  – 10,738 total links with manual provider...
Knowledge Base Accuracy
• Expert review of 25 patients with ≥ 3
  problems and ≥ 5 medications
• 82.1% positive predictive...
SMART Architecture
• Substitutable Medical Apps, Reusable
  Technologies
• Allows “iPhone-like” substitutability for
  med...
SMART App Development
• SMART Javascript Library
• Developed and tested in the SMART
  Reference EMR
  – 50 test patients
...
SMART App Demonstration




                          12
13
14
Future Directions
• Expand knowledge base to include medications
  with “isa” relationship
• Improve local mapping to stan...
Summary
• Ontology-based summarization knowledge
  base can assist problem-medication
  linking
• SMART app effectively ut...
Acknowledgments
• Funding
   – National Center for Cognitive Informatics and Decision
     Making in Healthcare SHARP Prog...
Thank You

Email: allison.b.mccoy@uth.tmc.edu

      Twitter: allisonbmccoy




                                     19
A Prototype Knowledge Base and SMART App to Facilitate Organization of Patient Medications by Clinical Problems
You’ve finished this document.
Upcoming SlideShare
Nuts and Bolts of Content Marketing for Plastic Surgeons
Next
Upcoming SlideShare
Nuts and Bolts of Content Marketing for Plastic Surgeons
Next

0

Share

A Prototype Knowledge Base and SMART App to Facilitate Organization of Patient Medications by Clinical Problems

Increasing use of electronic health records requires comprehensive patient-centered views of clinical data. We describe a prototype knowledge base and SMART app that facilitates organization of patient medications by clinical problems, comprising a preliminary step in building such patient-centered views. The knowledge base includes 7,164,444 distinct problem-medication links, generated from RxNorm, SNOMED CT, and NDF-RT within the UMLS Metathesaurus. In an evaluation of the knowledge base applied to 5000 de-identified patient records, 22.4% of medications linked to an entry in the patient’s active problem list, compared to 32.6% of medications manually linked by providers; 46.5% of total links were unique to the knowledge base, not added by providers. Expert review of a random patient subset estimated a sensitivity of 37.1% and specificity of 98.9%. The SMART API successfully utilized the knowledge base to generate problem-medication links for test patients. Future work is necessary to improve knowledge base sensitivity and efficiency.

Related Books

Free with a 30 day trial from Scribd

See all

Related Audiobooks

Free with a 30 day trial from Scribd

See all
  • Be the first to like this

A Prototype Knowledge Base and SMART App to Facilitate Organization of Patient Medications by Clinical Problems

  1. 1. A Prototype Knowledge Base and SMART App to Facilitate Organization of Patient Medications by Clinical Problems Allison B. McCoy, PhD Adam Wright, PhD, Archana Laxmisan, MD, MA Hardeep Singh, MD, MPH, Dean F. Sittig, PhD
  2. 2. Summarization Aims • Develop methodologies that: – Model and summarize complex, chronically-ill patients’ EHR data – Enhance decision making with context- appropriate, evidence-based recommendations • To improve clinician decision-making under information overload and time pressure 2
  3. 3. Research Focus • Develop a knowledge base for linking medications and problems • Evaluate the knowledge base with real patient data • Implement a prototype app that utilizes the knowledge base to summarize patient data 3
  4. 4. Ontology-Based Summarization • The Unified Medical Language System Metathesaurus (UMLS) – RxNorm, SNOMED CT, NDF-RT NDF-RT “may_treat” NDF-RT SNOMED CT RxNorm CUI Preparation Disease Concept “isa” SNOMED CT Concept 4
  5. 5. Ontology-Based Summarization • Generated over 7 million problem– medication links RxCUI Medication SNOMED Code Problem 198211 SIMVASTATIN 40 MG Oral (systemic) tablet 13644009 Hypercholesterolaemia 198211 SIMVASTATIN 40 MG Oral (systemic) tablet 302870006 Hypertriglyceridaemia 198211 SIMVASTATIN 40 MG Oral (systemic) tablet 3744001 Hyperlipoproteinaemia 198211 SIMVASTATIN 40 MG Oral (systemic) tablet 129589009 Endogenous hyperlipaemia … … … … 200345 SIMVASTATIN 80 MG Oral (systemic) tablet 13644009 Hypercholesterolaemia 200345 SIMVASTATIN 80 MG Oral (systemic) tablet 302870006 Hypertriglyceridaemia 200345 SIMVASTATIN 80 MG Oral (systemic) tablet 3744001 Hyperlipoproteinaemia 200345 SIMVASTATIN 80 MG Oral (systemic) tablet 129589009 Endogenous hyperlipaemia … … … … 5
  6. 6. Evaluation Setting • Large, multi-specialty ambulatory academic practice for adults, adolescents and children • Electronic health record utilization • Manual links enabled between a medication and a diagnosis within the patient’s clinical problem list 6
  7. 7. Evaluation Population • 5000 randomly selected patients • At least one outpatient encounter during July 1, 2010-December 31, 2010 – At least one active, coded clinical problem and medication – Included only problems and medications mapped to SNOMED CT and RxNorm 7
  8. 8. Knowledge Base Utility • Knowledge base linked 5,251 medications to problems – 10,738 total links with manual provider links – 4,763 medications not previously linked by providers (47% of total links) 8
  9. 9. Knowledge Base Accuracy • Expert review of 25 patients with ≥ 3 problems and ≥ 5 medications • 82.1% positive predictive value – Ex. Linking Prednisone 20 MG Tablet to Hypertrophic Scar – false due to wrong route • 37.1% sensitivity – Ex. Difference in precision of medication and problem entries in EHR compared to KB – Ex. Relationships within UMLS are incomplete – Ex. Incomplete medication and problem mappings in EHR to standards 9
  10. 10. SMART Architecture • Substitutable Medical Apps, Reusable Technologies • Allows “iPhone-like” substitutability for medical apps 10
  11. 11. SMART App Development • SMART Javascript Library • Developed and tested in the SMART Reference EMR – 50 test patients – Medications mapped to RxNorm – Problems mapped to SNOMED CT 11
  12. 12. SMART App Demonstration 12
  13. 13. 13
  14. 14. 14
  15. 15. Future Directions • Expand knowledge base to include medications with “isa” relationship • Improve local mapping to standardized terminologies • Compare accuracy with knowledge bases developed using data mining and other methods • Expand knowledge base to include other data types (e.g., lab values) • Implement knowledge base into live EHR • Evaluate use of summarization and knowledge base on clinical practice 16
  16. 16. Summary • Ontology-based summarization knowledge base can assist problem-medication linking • SMART app effectively utilizes knowledge base to display problem-oriented medication list 17
  17. 17. Acknowledgments • Funding – National Center for Cognitive Informatics and Decision Making in Healthcare SHARP Program Award (Grant No. 10510949) – Houston VA HSR&D Center of Excellence (HFP90-020) – NCRR Grant (3UL1RR024148) – UT Houston-Memorial Hermann Center for Healthcare Quality and Safety • Collaborators – SHARP-C Project 3 Members – UTHealth CDW Team – UTHealth SBMI IT Support – SMART Team 18
  18. 18. Thank You Email: allison.b.mccoy@uth.tmc.edu Twitter: allisonbmccoy 19

Increasing use of electronic health records requires comprehensive patient-centered views of clinical data. We describe a prototype knowledge base and SMART app that facilitates organization of patient medications by clinical problems, comprising a preliminary step in building such patient-centered views. The knowledge base includes 7,164,444 distinct problem-medication links, generated from RxNorm, SNOMED CT, and NDF-RT within the UMLS Metathesaurus. In an evaluation of the knowledge base applied to 5000 de-identified patient records, 22.4% of medications linked to an entry in the patient’s active problem list, compared to 32.6% of medications manually linked by providers; 46.5% of total links were unique to the knowledge base, not added by providers. Expert review of a random patient subset estimated a sensitivity of 37.1% and specificity of 98.9%. The SMART API successfully utilized the knowledge base to generate problem-medication links for test patients. Future work is necessary to improve knowledge base sensitivity and efficiency.

Views

Total views

700

On Slideshare

0

From embeds

0

Number of embeds

4

Actions

Downloads

0

Shares

0

Comments

0

Likes

0

×