Cli linic nical al De Decis isio ion n Sup uppo port rt - - PowerPoint PPT Presentation

BPM Roundtable undtable Eind ndhov

• ven

en Univ iver ersit ity of Technol hnology

• gy

5/11/2012

Op Oppo portu rtunit nities ies for r Ev Evid idence nce-base based d Cli linic nical al De Decis isio ion n Sup uppo port rt Syst stems: ems: An An Ap Appl plic ication ation for r On Oncolo logy gy

• Dr. Yaron
• n Denek

nekam amp

CMIO, Hospital Division, n, Clalit Health th Service ces, Tel Aviv Facul ulty ty of M Medicine ne, Techni hnion, Haifa

Europ ropean an BPM Round undtabl table Eind ndhov

• ven

en Univ iver ersit ity of Technol hnology

• gy

5/11/2012

• Specialist in Internal Medicine
• Medical Informatics - fellowship and studies at

Harvard MIT division of Health Sciences and Technology in Boston

• Chief Medical Informatics Officer

Hospital Division, Corporate Headquarters, Clalit Health Services, Tel Aviv

• Faculty of Medicine, Technion, Haifa

• Largest health care provider in Israel
• 4 Million Insured Members – over half of the population
• Second largest organization of its type in the world
• 14 hospitals, 1400 clinics
• 36,000 employees, 10,000 physicians
• Clinical IT systems
• Community- one EMR system
• Hospitals – mainly 2 EMR systems
• Several BI systems

Clalit Health Services Group

Israel’s Leading Health Care Organization

Hospitals across Israel

• Community outreach
• National centers of excellence
• Children’s Hospital
• Cancer center
• Organ transplant center
• Rehabilitation centers

5

I.

Why do we need to support clinical decisions and processes?

II.

Types and examples of clinical decision support systems (CDSS)

III.

Opportunities and challenges of implementing CDSS

• Preventable <> side effects
• Medication – allergy, dosages, drug-drug

interactions etc.

• Misdiagnosis
• Treatment – medical and surgical

Covell study of LA Internists, Ann Intern Med 1985

2 unanswered clinical questions for every 3 pts

• Holds across PCPs and specialty care
• Holds across urban and rural

Gorman, Medical Decision Making 1995, Gorman, Medinfo 2001

• 30% of questions were pursued
• 70% of information needs were not-pursued!
• Too busy, no immediate access to resources, or office

materials (books, journals, etc.) in disarray or out-of-date

Some outcomes - Overuse

• 30% of children receive excessive antibiotics for ear infection
• 20-40% of surgical procedures unnecessary
• 50% of back pain x-rays unnecessary

Underuse

• 50% of elderly patients don’t get the necessary immunizations

in winter

• > 20,000 journals published
• 17,000 new books per year
• Size doubles every 10-15 years
• 2 Million facts needed to practice

‘‘Men are men; the best sometimes forget’’

Othello, 1605; Act II, Scene iii

Medical School Conferences, CME, Lectures Clinical Reports Medical Record Books, Journals Guidelines Administrative Issues Decisions, Decisions, Decisions Curbside Consultation

• A constellation of psychological studies converging to a

description of human decision making under uncertainty

• It was awarded the Nobel Prize of Economics (2002)
• Main points:
• People use a few simple heuristics when

making judgments under uncertainty

• These heuristics sometimes are useful and
• ther times lead to severe and systematic

errors

“Information Technology must play a central role in the redesign of the health care system…”

• Current Approach
• Professional autonomy drives variability
• Decision making is based on training and

experience

• New Rules
• Knowledge is shared and information flows

freely

• Decision making is evidence-based

Richardson, William C. IOM 2001 Crossing the Quality Chasm, pg. 71

Active knowledge systems which use two or more

items of patient data to generate case-specific advice Support <> Replace link health observations with health knowledge to influence health decisions by clinicians for improved health care processes

Evidenced Based Medicine

• High quality level studies – ranking for quality of

research, level of evidence Clinical practice guidelines

• Developed by professional organizations, HMO’s,

Ministry of health Problems –

• Physicians are not aware to that knowledge
• Time constraints to read and follow
• Compliance

Evidenced Based Medicine Clinical Actions

CDSS

• Medications – prescribing, ordering
• Treatments – most useful cost effective

alternatives

• Diagnostic – complex multi step process
• Institutional administrative processes – duplicated

testing, consultations, transferring labs specimens to the labs etc.

• Alerting - when a clinical data is abnormal
• r a clinical guideline is not followed
• Critiquing – when ordering a medication or a test
• Reminders – reminding the clinician to follow

desired practice guidelines and policies

• More challenging -
• Computer-interpretable guidelines
• Expert systems

Drug-Drug Interaction Screening Dosages Renal and Hepatic function Duplicate Therapy Checking Drug Food Interactions Drug Allergy Checking Drug Indications Drug Side Effects Drug Disease Contraindication Checking Paediatric Precautions Patient Education

Supporting cost effectiveness Redundant test example

Suporting the diagnostic process Alternate exam

Lower cost Guideline for expensive medication

Reminders

Physicians reminded to give flu shots do so twice often

Inference methods

• Algorithmic
• Statistical
• Pattern Matching
• Rule-based

(Heuristic)

• Fuzzy sets
• Neural nets
• Bayesian

Knowledge Base Inference Engine

• Still are not a success
• Problem in having the probability data like for

symptoms, signs etc.

• Awareness to the need, also from the perspective of claims and

risk management

• Platforms – EMR systems are increasingly used including CPOE

cannot be a standalone system

• Standards of clinical data and information

More standards are used for diagnoses, medications, clinical information model

• Less reluctance of healthcare workers
• More evidence that it works
• Documentation of the processes in EMR systems enables

process mining to discover improved workflows

• Workflow integration on top of the EMR system - naturally fit into the

process of care

• Representing and maintaining medical knowledge and process

models

• Complexity of modeling time oriented data
• Need for flexibility – supporting process, not replacing the

professionals

• Dealing with ambiguity
• Mechanisms to avoid the alert fatigue phenomena
• UI – simple interface directed by the user

• Improved patient safety
• reduced medication errors and adverse events
• improved medication and test ordering
• Improved quality of care
• increased application of clinical guidelines, facilitating the use
• f up-to-date clinical evidence
• Improved efficiency in health care delivery
• reductions in test duplication
• decreased adverse events
• changed patterns of drug prescribing favoring cheaper but

equally effective generic brands