Configuring CDS Rules To Improve

Care And Achieve Meaningful Use

AMIA NOW!
25/26 May 2010, Phoenix AZ

Faculty
y
Jerome A. Osheroff, MD
Robert A. Greenes, MD, PhD
Si j A
Siraj Anwar, MS
MS, MBBS
 Workshop Objectives
Describe
D ib generall strategies
t t i ffor successful
f l
CDS deployments; e.g. CDS 5 Rights.
Understand
U d t d rolel ffor materials
t i l fforthcoming
th i
from AHRQ to support CDS rule
implementation.
Demonstrate skill in configuring CDS rules
to address meaningful use and related
performance improvement targets.
Create draft CDS rule, including detailed
implementation considerations.
 Workshop Agenda
Introductions/Session Overview: 10 min
Did ti
Didactics: 30 min
i
General CDS success factors
AHRQ eRecommendation template; hands-on worksheet
In
I the
h trenches
h CMIO perspective
i on eRecommendations
R d i
Hands-on work, Part 1: 20 min
Implementation Considerations worksheet: breakout groups
Break: 15 min
Hands-on work, Part 2: 30 min
Breakout groups
g p
Full group report-out: 30 min
Share best practices/learning from exercise
 Introductions: Faculty

Jerome A. Osheroff, MD
Chief Clinical Informatics Officer, Thomson Reuters
Adjunct Assistant Professor of Medicine, University of Pennsylvania

Robert A. Greenes, MD, PhD

Ira A Fulton Chair and Professor, Department of Biomedical Informatics
Arizona State University

Siraj Anwar, MS, MBBS

CDS Clinical Informaticist
Memorial Hermann Healthcare System
University of Texas
Texas-Houston
Houston
 Introductions: Participants
Sector?
Healthcare
H lth d
delivery
li vs. other
th ((consultant,
lt t vendor)
d )

Care Setting?
Large/small practice vs. hospital/health system

Role?
Informatics (CMIO/director) vs. IT vs. other

CDS stage?
Have
H you ever iimplemented
l t d a CDS rule?
l ?
Clincial Decision Support Is A Cornerstone
Of Meaningful Use

provide
id persons involved
i l d iin care
processes with general and person-
specific information,
information intelligently filtered
and organized, at appropriate times, to
enhance
e a ce health
ea t a
and d health
ea t ca e from
care
MU NPRM

Includes/builds on current processes

NOT just rules and alerts

6
 But Not Easy To Get
Get Right
Right

9 Advanced clinical systems with CDS

BUT
o admissions at least 1 ADE; 9% serious harm
o Problems with drug dosing, selection, monitoring
Guidance On Getting CDS Right
HIMSS/AMIA Textbook
2005 HIT book of the Published by Elsevier,
year 2007
All-time HIMSS Widely used reference
bestseller
Focus on fundamentals,
Widely used by technologies, and
CMIOs/others strategies

Co-published 4/09 by
leading societies; ~100
contributors Other
2009 HIT book of the
10 commandments for CDS:
year
Bates et al. JAMIA 2003
Co-sponsors include
AHRQ, 3 CIS vendors
This is not a book
A Formula For Success:
The CDS Five Rights
To improve care outcomes with CDS you must provide:
the Right Information
Evidence-based, useful for guiding action and answering questions

to the Right Stakeholder

Both clinicians and patients

in the Right Format

Alerts, Order Sets, answers, etc.

through
through the Right Channel
Channel
Internet, mobile devices, clinical information systems

at the Right Point in the Workflow

to influence key decisions/actions
CDS IIntervention
t ti T
Types

Relevant
R l tddata
t presentation:
t ti flowsheets, surveillance

Order creation facilitators: order sentences, sets

Reference information: infobuttons, Web
Unsolicited alerts: proactive warnings
Documentation templates: patient history, visit note
Protocol support: pathways

10
 Workflow Opportunities for CDS

Rules to query
databases (e.g.
registries) to
identify patients
needing
intervention

Population-
based Encounter-based
Goals of Structured Care Recommendations for
CDS (SCRCDS) p project
oject
12-month project funded by AHRQ (Oct 09-Sept 10)
Thomson Reuters as prime, subcontractors at Arizona State, UCSD, and Intermountain Health
Goal is to develop methods to facilitate wide dissemination and uptake of CDS
Focus on US Preventive Services Task Force (USPSTF) prevention and screening
recommendations
Limit to A & B recommendations (45 in all)
Turn these into implementable rules (alerts, reminders, etc.)
Also include selected additional recommendations tied to Meaningful
g Use Quality
y
Measures
Emphasis on
Consensus on level of specificity to satisfy vendor/implementer desires for further adaptation by
themselves
Adaptable to multiple platforms
Approach
Rapid turnaround
Broad availability of results
Iterative refinement in future efforts
 USPSTF A&B Recommendations
Grade Title Grade Title

A* Aspirin to Prevent CVD: Men age 45 to 79 to prevent myocardial infarctions

B Abdominal Aortic Aneurysm: Screening -- Men 65-75, Smoker
B Alcohol Misuse: Screening and Behavioral Counseling -- Men, Women, and Pregnant Women
A* Aspirin to Prevent CVD: Women age 55 to 79 to prevent ischemic strokes
B BRCA Mutation Testing for Breast and Ovarian Cancer: Women, Increased Risk
A* Asymptomatic
y p Bacteriuria: Screening
g -- Pregnant
g Women
B Breast Cancer: Preventive Medication Discussion -- Women, Increased Risk
A Cervical Cancer: Screening -- Women who are sexually active
B Breast Cancer: Screening Mammography -- Women 40 and Older
Chlamydia: Screening -- Women Ages 24 and Younger OR Women Ages 25 and Older at Increased
A B* Breastfeeding: Primary Care Interventions to Promote -- All Pregnant Women and New Mothers
Risk
Chlamydia: Screening -- Pregnant Women Ages 24 and Younger OR Pregnant Women Ages 25 and
B
A* Colorectal Cancer: Screening -- Adults, beginning at age 50 years and continuing until age 75 years Older at Increased Risk

A* C
Congenital
it l Hypothyroidism:
H th idi S
Screening
i -- Newborns
N b
B Dental Caries: Oral Fluoride Supplementation
pp -- Preschool Children 6 Months and Older

A* Folic Acid: Supplementation -- All Women Planning or Capable of Pregnancy

B* Depression: Screening -- Adolescents, 12-18 years of age, in Clinical Practices with Systems of Care

A Gonorrhea: Preventive Medication -- Newborns B Depression: Screening -- Adults, In Clinical Practices with Systems of Care in Place

A HIV: Screening -- Adults and Adolescents at Increased Risk B Gonorrhea: Screening -- Pregnant Women and Women at Increased Risk

A HIV: Screening -- Pregnant Women

B Healthy Diet: Counseling -- Adults with Hyperlipidemia and Other Risk Factors for CVD
B* Hearing Loss in Newborns: Universal Screening -- Newborns
A* H
Hepatitis Virus: Screening
titi B Vi S i -- Pregnant
P t Women
W
Iron Deficiency Anemia: Iron Supplementation -- Asymptomatic Children 6-12 Months, Increased
A* High Blood Pressure: Screening -- Adults 18 and Over B
Risk
A Lipid Disorders in Adults: Screening -- Men 35 and Older
B Iron Deficiency Anemia: Screening -- Asymptomatic Pregnant Women
A Lipid Disorders in Adults: Screening -- Women 45 and Older, Increased risk for CHD B Lipid Disorders in Adults: Screening -- Men 20-34, Increased risk for CHD
A* Phenylketonuria (PKU): Screening -- Newborns B Lipid Disorders in Adults: Screening -- Women 20-44, Increased risk for CHD
A Rh(D) Bl
Blood
d Typing:
T i S
Screening
i -- Pregnant
P t Women,
W First
Fi t Pregnancy
P Related
R l t d Visit
Vi it B Ob it Screening
Obesity: S i and
d IIntensive
i C Counseling
li -- Obese
Ob M
Men and
dWWomen

A* Sickle Cell Disease: Screening -- Newborns Osteoporosis: Screening -- Postmenopausal Women 65 Years and Older with No Risk Factors, or 60
B
A* Syphilis: Screening - Pregnant Women Years and Older with Risk Factors

A Syphilis: Screening -- Men and Women at Increased Risk B Rh(D) Blood Typing: Screening -- Antibody Testing Unsensitized Rh (D)-Negative Pregnant Women

A* Tobacco Use: Counseling and Interventions for Adults Sexually Transmitted Infections: Behavioral Counseling -- Sexually Active Adolescents and Adults at
B*
A* T b
Tobacco Use:
U C
Counseling
li and
d IInterventions
t ti ffor P
Pregnantt Women
W Increased Risk
B* Type 2 Diabetes Mellitus: Screening Men and Women -- Sustained BP 135/80+

* denotes new grade definition B Visual Impairment: Screening -- Children Younger than 5 Years

* denotes new grade definition

 Life Cycle of Rule Refinement
Start with EBM statement (such as USPSTF recommendation)
Stage
St
1. Identify key elements and logic who, when, where, what to be done
Structured headers, unstructured content
M di ll specific
Medically ifi SCRCDS project focusing on this stage
2. Formalize definitions and logic conditions
Structured headers, structured content (terms, code sets, etc.)
Medically specific Implementation notes and
considerations as start of this stage
3. Specify adaptations for execution
Taxonomy of possible workflow scenarios and operational considerations
Selected particular workflow-
workflow and setting-
setting specific attributes for particular sites

4. Convert to target representation, platform, for particular implementation

Host language (Drools, Java, Arden Syntax, )
Host architecture: rules engine,
g , SOA,, other
Ready for execution
 eRec Template/Sample
1. Header Information
eRecommendation Name
eRecommendation ID
eRecommendation Version Date/Number
Template Version Date/Number
Related eMeasure(s)
Author
Verified by
Maintained by
Description/Purpose
Recommendation Text from Source
Setting (if specified by Source)
Recommendation classification
Rationale
Reference
Reference URL
USPSTF SCREENING FOR BREAST CANCER Recommendation Set USPSTF A and B
(B Recommendation on mammography only) Recommendations
USPSTF-MAMMO-B-REC Set ID USPSTF-A-B-RECS
Recommendation Version Date/Number 2 (revision of 2002 guidelines)
PQRI112:Preventive Care and Screening: Screening Mammography [PQRI age range40 69]
Agency for Healthcare Research and Quality (AHRQ) and Preventive Services Task Force (USPSTF)
U.S. Preventive Services Task Force (USPSTF) recommendation statement on screening for breast cancer in the general population.
Summary Statement The USPSTF recommends biennial screening mammography for women between
the ages of 50 and 74 years.
Additional Inclusion/Exclusion Criteria . . . . . . . . . This recommendation statement applies to women 40 years or older who
are not at increased risk for breast cancer by virtue of a known underlying genetic
mutation or a history of chest radiation. . . . . . .
Not specified. See implementation considerations.
Screening: primary prevention
Importance
Breast cancer is the second-leading cause of cancer death among women in the . Widespread use of screening, along with treatment
advances in recent years, has been credited with significant reductions in breast cancer mortality.
Detection
Mammography, as well as physical examination of the breasts (CBE and BSE), can detect pre-symptomatic breast cancer. Because of
its demonstrated effectiveness in randomized, controlled trials of screening, film mammography is the standard for detecting breast
cancer; in 2002, the USPSTF found convincing evidence of its adequate sensitivity and specificity.
Benefits of Detection and Early Intervention:
There is convincing evidence that screening with film mammography reduces breast cancer mortality, with a greater absolute reduction
for women aged 50 to 74 years than for women aged 40 to 49 years. The strongest evidence for the greatest benefit is among women
aged 60 to 69 years.
Clinical Guidelines: Screening for Breast Cancer: Preventive Services Task Force Recommendation Statement. U.S. Preventive
Services Task Force. Ann Intern Med 151:716-726
http://www.ahrq.gov/clinic/uspstf/uspsbrca.htm
1
5
 eRec Template/Sample (Cont.)
2.a Data definitions
Category Data Elements Relevant Notes
Eligibility/ Demographic For PQRI 112 to which this logic statement is related, age high limit = 69
Inclusion-related Target gender: F, Target age low limit: 50, Target age high limit:
data 74
Condition/ risk
[not relevant to mammography example]

Intervention Screening interval: 2 years [See Section 3. Implementation

interval Considerations below for details on operational exclusion criteria
and related logic where screening interval is used ]
Exclusion High risk patients
criteria-related <Value set: History of chest radiation >
data Quality data type: Procedure Result, Code set: (CPT 4, ICD9,
SNOMED), Code list: {list of relevant codes relating to Hx of chest
radiation}.
<Value set: Known genetic mutation, BRCA1, BRCA2, [possibly
others]>
Quality
Quality data type: Laboratory test result, Code set: (LOINC,
SNOMED), Code list: {list of relevant codes for genetic tests}
<Value set: mammogram results documented within 2 years >
Quality data type: Diagnostic study result , Code set:
(CPT,LOINC, SNOMED), Code list: {list of relevant codes}
Other exclusion-related data
[not relevant to mammography example]
High risk patients may require a different screening protocol. The USPSTF recommendation
states that a known genetic mutation or a history of chest radiation puts a woman at an increased
risk for breast cancer and excludes this group from the screening recommendation. The
recommendation implies that a different screening/treatment recommendation/protocol applies to
this high risk group, although it does not make explicit such a recommendation/protocol.
Therefore, it might be appropriate for implementers to consider if there is a
recommendation/protocol for the screening/treatment of the given high risk group in place in the
system:
If there is a protocol, and if there is evidence that a high risk patient is already on such a
protocol, exclude this patient from the recommendation.
If there is a protocol, and a high risk patient is not on it, recommend that the patient be put on
the protocol
If there is no protocol, or if there is evidence that the patient is on such a protocol elsewhere
(e.g., having had BRCA1/2 testing), exclude this patient.
Otherwise, do not exclude this high risk patient.

Operational [Will depend on implementation considerations/choices: See Optional element: implementer may define and use operational exclusion criteria pertinent to local
exclusion Section 3, Implementation Considerations for examples] needs and constraints. For example, if the intervention recommended is addressed/pending, or if
criteria-related patient has condition being screened and is already undergoing treatment, etc. then
data implementers may wish to suppress the intervention recommendation to minimize false positive
notifications. See Implementation
p Consideration section for further details and examples.
p

Action related <Value set: Bilateral mammogram>

data Quality data type: Diagnostic Study Order , Code set:
(CPT,LOINC, SNOMED), Code list: {list of relevant codes for
screening mammography tests}
16
 eRec Template/Sample (Cont.)
(C t )

2.b Logic Statement

Category Logic Elements Relevant Note
<Eligibility/inclusion Patient gender = Target gender <Evidence of condition/risk> statement is a template placeholder for other
criteria> AND: rule types: not pertinent to this breast cancer screening sample
<Patient age >= Target age low limit>
AND
<Patient age <= Target age high limit>
AND:
<Evidence of condition/risk = non-null ]>

<Exclusion criteria> <Patients for whom a different intervention protocol may be warranted> See section 3, subsection on Optimizing Rule Specificity for further details on
<Value set: History of chest radiation > = non-null operational exclusion criteria, e.g., related to pertinent pending interventions,
OR: <Value set: Known genetic mutation > = non-null etc.
<Patients that have already received intervention within recommended
interval>
<Value set: mammogram
g results documented within 2 yyears > = non-null

<Operational exclusion [Will depend on implementation considerations/choices: See Section 3,

criteria> Implementation Considerations for examples]

<Action> <Recommended action: perform Intervention:

procedure/test/medication/counseling/etc.>
<Bilateral mammogram>
o Quality data type: Diagnostic Study Order>
o <Code set: (CPT,LOINC, SNOMED)
o Code list: {list of relevant codes for screening mammography tests}

17
 eRec
e ec Template/Sample
e p ate/Sa p e (Co
(Cont.)
t)

3. Implementation Considerations
OPTIMIZING RULE SPECIFICITY:
Operational data
o Notification fired : Provider, date;
Operational exclusion criteria data
o Tests for diagnosis or problem in process or done within specified screening interval: Mammogram completed within past 2 years: Record of the patient having
received a mammogram in the previous 2 years (by history or by stored data); By history: Mammogram externally as per patient history or need for such
request
q to be asked in CDS;;
o Pre-existing condition diagnosis or problem: Patient has condition being screened (thus being managed, not in primary prevention mode); ;Indirect evidence
of diagnosis or problem already made: Recurrent tests or procedures implying diagnosis; <Value set: Pathology diagnoses, cytology, etc.;
o Rule having fired within specified alerting interval;
Operational exclusion criteria logic
- AND NOT: Tests for diagnosis or problem in process; AND NOT: Pre-existing condition diagnosis or problem;
- ELSE AND NOT: Rule having fired within specified alerting interval; OR NOT: Reason noted for not following rule recorded within specified alerting interval
DETERMINING RULE TRIGGERING:
Is operation interactive/real time? Batch mode, e.g., through clinic/practice administration? Can information be obtained from patient at time of rule firing?
DEFINING NOTIFICATION APPROACH:
User notification: Is it desirable to set an indicator that a notification has been delivered, e.g., to avoid redundant firing? Notification Acknowledgment: Is it
desirable to document notification response
response, ee.g.,
g for rejection of recommended action?
OBTAINING KEY DATA:
What minimum data are needed to fire a useful rule for this recommendation in your organization?
ACCOMODATING LOCAL CLINICAL POLICIES:
Target age high limit;

18
The CDS Implementer Dilemma

Opportunities >>> Resources !

What To Do?

Critical role of internal CDS

governance for prioritization
Effective internal processes for
design, build, test, communication
Use
U effective
ff ti available
il bl ttemplates
l t
for starting points
Governance: Role of CDS
Oversight Committee

Co-Chaired by CMIO and System Exec,

Pharmacy
(Primarily led by CDS Clinical Informaticist!!)
Multi-disciplinary
Multi disciplinary team:
Physicians, nursing, pharmacy, quality, case
management,
g , risk management,
g , informatics,, ISD
Reports to (1) System Quality and (2) Clinical
Systems Workgroup
Comprehensive
p Approach
pp to
Implementation

CDS used to be departmental focused

rules
rules
Now requires more attention on workflow
analysis for data capture
capture, rules triggers
triggers, alert
notifications
Evolution from installing/implementing clinical
systems to optimizing systems to achieve
quality
q y and safety
y
Questions?

Thank You!
Si j Anwar:
Siraj A
Anwar.MohammadSirajuddin2@memorialhermann.org

Bob Greenes:
greenes@asu edu
greenes@asu.edu

Jerry Osheroff: jerry.osheroff@thomsonreuters.com

Workshop Exercise

Simulates tasks done by CDS experts to

successfully implement template
recommendations
Exercise Goal

Take proposed AHRQ eRecommendation for

Breast Cancer Screening and develop
implementation considerations for sample
organizations
Worksheet 1: Obtaining Key Data
Worksheet 2: Determining g Rule Triggers
gg and
Defining Notification Approach
Logistics
Tables:
Small group practice (1-5 practices)
Medium-sized multispecialty clinic (5-100 providers)
Large academic medical center (>100 providers)
Assign table spokesperson
Use templates to structure your analysis
Time is artificially condensed heretry to focus
quickly on most valuable approaches (Worksheet 2)
Read out of pearls from hands-on
Workshop Takeaways

CDS 5 Rights as a success strategy

Potential role for eRecs/related shared tools
Pearls for implementing rules for MU, etc.
Draft rule on breast cancer
Value from collaboration (AMIA/SI/AMDIS)