computer-assisted decision making in the twenty-first century james j. cimino, m.d. departments of...

33
Computer-Assisted Decision Making in the Twenty-First Century James J. Cimino, M.D. Departments of Medical Informatics and Medicine Columbia University

Post on 20-Dec-2015

217 views

Category:

Documents


0 download

TRANSCRIPT

Computer-Assisted Decision Making in the Twenty-First Century

James J. Cimino, M.D.

Departments of Medical Informatics and Medicine

Columbia University

Overview

• Analyzing medical errors

• Evidence-based practice (EBP)

• Using computers to support EBP

• Challenges and impediments to achieving EBP

Analyzing Medical Errors

Leape, LL. Error in Medicine. JAMA 1995; 272(23):1851-7.

Errors

Slips: errors of action

Mistakes: errors of conscious thought

Solution: Monitoring Solution: Information

Analyzing Medical Errors

• Reduced reliance on memory

• Increased vigilance

• Improved information access

• Error proofing (“forcing functions”)

• Training emphasis error prevention

• Patient education

• Standardization of practice patterns

Evidence-Based Practice (EBP)

• Decisions based on clinical evidence

• Spectrum of evidence quality

• Skills needed to:– Access literature– Summarize findings– Apply conclusions

Sackett DL, et al. Evidence based medicine: what it is and what it isn’t. BMJ 1996; 312(7023):71-2

Application of EBP

• Etiology

• Prevention

• Diagnosis

• Therapy

Computer Support of EBP

• EBP and:– access to literature– guidelines– diagnostic aids– order checking

• Identify:– applications– problems– solutions

Literature

• Applications– Medline– Textbooks

• Problems– Search expertise– Time constraints

• Solutions– Infobuttons– Palm-based access

Guidelines• Applications

– Computer-based text guidelines

• Problems– Finding applicable guidelines– Navigation– Applying to specific cases

• Solutions– Indexing text guidelines– Customizing guidelines– Automating guidelines

Diagnostic Decision Support

• Applications– Interpretation of tests– Expert systems

• Problems– Need human intervention– Terminology translation

• Solutions– Identify where aids are needed– Translate data to clinical terms– Automate data transfer

Order Checking

• Applications– Drug-interaction programs– Alerting systems

• Problems– Don’t know whole patient– May be inappropriate

• Solutions– Integration with clinical record– Open-loop

Alerts Problems: Terminology

• One day, an apparent epidemic of positive results…

…but lab showed “No Growth to Date”

• Alert checked Result not equal “No Growth”

• “No Growth to Date” “No Growth”

Alert Problems: No Human Review

• Alert checks for trends in creatinine level

• MD receives alert for patient’s Creat=1.7

• MD calls patient to come to ER

• Patient risks storm of decade to come to ER

• Creatinines are 1.1, 1.3, 1.8, 1.6, 1.3, 1.7

Challenges

• Identifying context-specific information needs

• Modeling patients and the care process

• Integration of systems

• Terminology translation

• User education

Conclusions• Potential areas for errors:

– Diagnostic testing– Interpretation of results– Therapeutic interventions– Monitoring

• Computers and EBP can be brought bear:– Literature at the point of care– Facilitating use of guidelines– Expert systems– Alerting– Patient decision support

• Enhance, not replace, human decision-making