1 The Importance of Unambiguous Medical Unambiguous Medical Terminology in Patient Terminology in Patient Care and Research Care and Research Or, why doctors and healthcare administrators Or, why doctors and healthcare administrators shouldn’t glaze over when informatics is discussed shouldn’t glaze over when informatics is discussed Robert M Califf MD Robert M Califf MD Vice Chancellor for Clinical Research Vice Chancellor for Clinical Research Duke University Duke University
44
Embed
1 The Importance of Unambiguous Medical Terminology in Patient Care and Research Or, why doctors and healthcare administrators shouldnt glaze over when.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
The Importance of Unambiguous The Importance of Unambiguous Medical Terminology in Patient Medical Terminology in Patient
Care and ResearchCare and ResearchOr, why doctors and healthcare administrators Or, why doctors and healthcare administrators
shouldn’t glaze over when informatics is shouldn’t glaze over when informatics is discusseddiscussed
Robert M Califf MDRobert M Califf MD
Vice Chancellor for Clinical ResearchVice Chancellor for Clinical Research
Duke UniversityDuke University
2
The Information SituationThe Information Situation
We are increasingly able to assimilate We are increasingly able to assimilate information about the health of people when information about the health of people when measurements are made by machinesmeasurements are made by machines
Lab dataLab data
ImagesImages
Test results (ECG, PFTs, etc)Test results (ECG, PFTs, etc)
Genomics, proteomics, metabolomics, etc.Genomics, proteomics, metabolomics, etc.
What are we missing?What are we missing?
The synthetic terms that tie the raw data into The synthetic terms that tie the raw data into actionable constructs about a personactionable constructs about a person
3
Clinical TerminologyClinical Terminology
We have excellent compliance with terms when We have excellent compliance with terms when they are required for billingthey are required for billing
Unfortunately, these terms for billing are not Unfortunately, these terms for billing are not the same as the preferred terms for clinical the same as the preferred terms for clinical quality or research assessmentquality or research assessment
If billing, patient care and research terminology If billing, patient care and research terminology come together, we can make monumental come together, we can make monumental strides in clinical quality at all levels (patient, strides in clinical quality at all levels (patient, practice, system, ? Population)practice, system, ? Population)
4
5
People are dying People are dying because we don’t use because we don’t use the same names for the the same names for the same things!same things!
6
A PatientA Patient
60 yo woman admitted to the ED with “chest 60 yo woman admitted to the ED with “chest pain”pain”
HR 100, sinus rhythm, BP 100/70, exam HR 100, sinus rhythm, BP 100/70, exam unremarkableunremarkable
ECG: sinus rhythm, ST segments abnormalECG: sinus rhythm, ST segments abnormal
Labs: K 4.2, creatinine 1.5, LDL 130, troponin WNLLabs: K 4.2, creatinine 1.5, LDL 130, troponin WNL
CXR: no abnormalities apparent in CV, lung, bone or CXR: no abnormalities apparent in CV, lung, bone or tissue structurestissue structures
7
Possible Clinical SituationsPossible Clinical Situations
Mild throat tightness relieved with MylantaMild throat tightness relieved with Mylanta
Ripping pain going down the backRipping pain going down the back
Midsternal chest pain, relieved after 2Midsternal chest pain, relieved after 2ndnd NTG NTG
Pleuritic chest pain and extreme shortness of Pleuritic chest pain and extreme shortness of breathbreath
Stabbing pain that lasts a few seconds and Stabbing pain that lasts a few seconds and then goes awaythen goes away
Great Baltimore Fire of 1904Great Baltimore Fire of 1904
One reason for the fire's duration was the lack of One reason for the fire's duration was the lack of national national standardsstandards in fire-fighting equipment. Fire crews in fire-fighting equipment. Fire crews fire enginesfire engines came from as far away as came from as far away as PhiladelphiaPhiladelphia and and Washington that day (units from Washington that day (units from New York CityNew York City were on were on the way, but were blocked by a train accident; they the way, but were blocked by a train accident; they arrived the next day). The crews brought their own arrived the next day). The crews brought their own equipment. Most could only watch helplessly when they equipment. Most could only watch helplessly when they discovered that their hoses could not fit Baltimore's discovered that their hoses could not fit Baltimore's hydrants. High winds and freezing temperatures added hydrants. High winds and freezing temperatures added to the difficulty for firefighters and further contributed to to the difficulty for firefighters and further contributed to the severity of the fire. As a result, the fire burned over the severity of the fire. As a result, the fire burned over 30 hours, destroying 1,545 buildings spanning 70 city 30 hours, destroying 1,545 buildings spanning 70 city blocks — amounting to over 140 acres.blocks — amounting to over 140 acres.
While Baltimore was criticized for its hydrants, this was a While Baltimore was criticized for its hydrants, this was a problem that was not unique to Baltimore. During the problem that was not unique to Baltimore. During the time of the Great Fire "American cities had more than six time of the Great Fire "American cities had more than six hundred different sizes and variations of fire hose hundred different sizes and variations of fire hose couplings." It is known that as outside fire fighters couplings." It is known that as outside fire fighters returned to their home cities they gave interviews to returned to their home cities they gave interviews to newspapers that condemned Baltimore and talked up newspapers that condemned Baltimore and talked up their own actions during the crisis. In addition, many their own actions during the crisis. In addition, many newspapers were guilty of taking for truth the word of newspapers were guilty of taking for truth the word of travelers who, in actuality, had only seen the fire as their travelers who, in actuality, had only seen the fire as their trains passed through the area. All of this aside the trains passed through the area. All of this aside the responding agencies and their equipment did prove responding agencies and their equipment did prove useful as their hoses only represented a small part of the useful as their hoses only represented a small part of the equipment brought with them. One benefit to this equipment brought with them. One benefit to this tragedy was the standardization of hydrants nationwidetragedy was the standardization of hydrants nationwide
Wikipedia 2009Wikipedia 2009
12
The Learning Health System at All LevelsThe Learning Health System at All Levels
Individual health care transactionsIndividual health care transactions
ProviderProvider
ConsumerConsumer
Clinic and health system qualityClinic and health system quality
Libby P. Circulation 2001;104:365, Hamm CW, Bertrand M, Braunwald E, Lancet 2001; 358:1533-1538; Davies MJ. Heart 2000; 83:361-366.Anderson JL, et al. J Am Coll Cardiol. 2007;50:e1-e157, Figure 1. Reprinted with permission.
6 Medical Therapies Proven to Reduce Death6 Medical Therapies Proven to Reduce Death Reduction in deaths:Reduction in deaths:
Goals for CRUSADE:Goals for CRUSADE: Improve Adherence to ACC/AHA Guidelines for Improve Adherence to ACC/AHA Guidelines for Patients with Unstable Angina/Non-STEMIPatients with Unstable Angina/Non-STEMI
Evaluating the Process of CareEvaluating the Process of Care
An adherence score is applied to each patient. An adherence score is applied to each patient. incorporating the components of process of care.incorporating the components of process of care.
The score from each patient then combined for all The score from each patient then combined for all patients at each hospital. patients at each hospital. Typical scores ranged from 50 to 95%.Typical scores ranged from 50 to 95%.
All 400 hospital adherence scores then ranked in All 400 hospital adherence scores then ranked in quartiles — best to worst.quartiles — best to worst.
Evaluating the Process of CareEvaluating the Process of Care
An adherence score is applied to each patient. An adherence score is applied to each patient. incorporating the components of process of care.incorporating the components of process of care.
The score from each patient then combined for all The score from each patient then combined for all patients at each hospital. patients at each hospital. Typical scores ranged from 50 to 95%.Typical scores ranged from 50 to 95%.
All 400 hospital adherence scores then ranked in All 400 hospital adherence scores then ranked in quartiles — best to worst.quartiles — best to worst.
CRUSADE: CRUSADE: Overall Adherence Score Trends Over TimeOverall Adherence Score Trends Over Time
68.1%69.6%
71.0%72.3%
73.0% 73.6%75.2%
77.9% 78.0%79.3%
60%
70%
80%
Q1'02
Q4'02
Q3'03
Q2'04
Peterson et al, ACC 2004Peterson et al, ACC 2004
CRUSADE: Link Between Overall ACC/AHA CRUSADE: Link Between Overall ACC/AHA Guidelines Adherence and MortalityGuidelines Adherence and Mortality
0
2
4
6
8
<=25% 25–50% 50–75% >=75%
Hospital Composite Quality Quartiles
% In
-ho
spit
al M
ort
alit
y Adjusted Unadjusted
Every 10% in guidelines adherence 11% in mortality
Impact of Quality Impact of Quality Improvement on Outcomes Improvement on Outcomes
in ACSin ACS
Trilogy in American Heart JournalTrilogy in American Heart Journal
January 2009January 2009
Treatment of STEMI PatientsTreatment of STEMI Patients
19901990 20062006
FibrinolysisFibrinolysis 52.5%52.5% 27.6%27.6%
Primary PCIPrimary PCI 2.6%2.6% 43.2%43.2%
D2N time*D2N time* 59 min59 min 29 min29 min
In-hosp mortalityIn-hosp mortality 7%7% 6%6%
D2B time**D2B time** 111 min111 min 79 min79 min
*Fibrinolysis-eligible pts who rec’d fibrinolysis**Non-transfer pts who rec’d primary PCI since 1994
All people in developed nations will have —All people in developed nations will have —
An electronic health recordAn electronic health record
Biological samplesBiological samples
Digitized imagesDigitized images
Healthcare will be personalized using an individual’s Healthcare will be personalized using an individual’s images, samples and clinical data.images, samples and clinical data.
The health of a community will be monitored using The health of a community will be monitored using aggregate records.aggregate records.
** All R&D costs (basic research and preclinical development) prior to initiation of clinical testing** All R&D costs (basic research and preclinical development) prior to initiation of clinical testing*** Based on a 5-year shift and prior growth rates for the preclinical and clinical periods*** Based on a 5-year shift and prior growth rates for the preclinical and clinical periods
DiMasi et al. 2003DiMasi et al. 2003
Comparative Pre-Approval Capitalized Comparative Pre-Approval Capitalized Costs per Approved New MoleculeCosts per Approved New Molecule
Innovation Gap Getting WiderInnovation Gap Getting Wider
26 2522
28
53
39
30
35
2724
1721
16
11
$39.40
$11.50
0
20
40
60
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
NMEs (New Drug Approvals)
PhRMA Member R&D Spending
Burrill & CompanyBurrill & Company
Pharma Innovation
Gap
Pharma Innovation
Gap
““Real” Clinical Trials—Done in the Setting of Real” Clinical Trials—Done in the Setting of Health Care DeliveryHealth Care Delivery
Tremendous cost of training for 3 different Tremendous cost of training for 3 different vocabulariesvocabularies
Redundant personnel costs of collecting same Redundant personnel costs of collecting same data in different ways is massivedata in different ways is massive
Problem List VocabulariesProblem List Vocabularies
Dr. Kim Wah FungDr. Kim Wah Fung
National Library of MedicineNational Library of Medicine
The problem listThe problem list
The problem list is a powerful way to organize and The problem list is a powerful way to organize and communicate clinical data and reasoning - recommended communicate clinical data and reasoning - recommended as an essential feature of an electronic medical record as an essential feature of an electronic medical record (EMR)(EMR)
Often the first (if not the only) part of clinical narration in Often the first (if not the only) part of clinical narration in an EMR that uses a controlled vocabularyan EMR that uses a controlled vocabulary
Most institutions develop and use their own problem list Most institutions develop and use their own problem list vocabulariesvocabularies Often linked to ICD codes for billing or reportingOften linked to ICD codes for billing or reporting Some are mapped to SNOMED CTSome are mapped to SNOMED CT
37
Goals of researchGoals of research
To study the problem list vocabularies of large health care To study the problem list vocabularies of large health care institutions - size, pattern of use and the extent to which institutions - size, pattern of use and the extent to which they overlap with (or differ from) each otherthey overlap with (or differ from) each other
To identify a CORE (To identify a CORE (CClinical linical OObservations bservations RRecording and ecording and EEncoding) set of terms that are of high usage in most ncoding) set of terms that are of high usage in most problem listsproblem lists
38
The CORE subsetThe CORE subset
The set of concepts that often appear in problem list The set of concepts that often appear in problem list vocabularies and are frequently usedvocabularies and are frequently used
Ways to use this subsetWays to use this subset As a ‘starter set’ to build local problem list vocabularies. If As a ‘starter set’ to build local problem list vocabularies. If
subsequent local extensions can be added in a standardized way, subsequent local extensions can be added in a standardized way, the divergence of these vocabularies can be minimizedthe divergence of these vocabularies can be minimized
Existing problem list vocabularies can be mapped to the CORE Existing problem list vocabularies can be mapped to the CORE conceptsconcepts
BenefitsBenefits Reduce variability of problem list vocabulariesReduce variability of problem list vocabularies Facilitate sharing of problem list dataFacilitate sharing of problem list data
39
Desirable features of the CORE subsetDesirable features of the CORE subset
High coverage of usageHigh coverage of usage Small number of conceptsSmall number of concepts Linkable to standard terminologiesLinkable to standard terminologies Supports reasoningSupports reasoning Supports a standard mechanism for adding local Supports a standard mechanism for adding local
extensionsextensions
40
41
42
Effective Methods of Getting the Attention of Effective Methods of Getting the Attention of Doctors and Health System AdministratorsDoctors and Health System Administrators
Appeal to humanitarian instinctAppeal to humanitarian instinct
Publicity for doing goodPublicity for doing good
Shame for doing badShame for doing bad
Distribute $34 Billion!Distribute $34 Billion!
43
It will be shameful is some portion of that $34 It will be shameful is some portion of that $34 billion allocation is not devoted to finalizing a core billion allocation is not devoted to finalizing a core terminology that is agreed to by all sectorsterminology that is agreed to by all sectors
Payors Payors
Government and privateGovernment and private
Provider groupsProvider groups
Primary care and specialtiesPrimary care and specialties
Research regulatorsResearch regulators
FDA, NIH, CMS, VA, DODFDA, NIH, CMS, VA, DOD
Pharma, DevicesPharma, Devices
With international harmonizationWith international harmonization
44
How do we resolve the “Tower of Babel” of data from How do we resolve the “Tower of Babel” of data from EHRs, PHRs, registries, databases, literature, and EHRs, PHRs, registries, databases, literature, and clinical trials?clinical trials?