with Increased Clinician Responsiveness and Improved Outcomes - - PowerPoint PPT Presentation

Reduced Alert Fatigue with Increased Clinician Responsiveness and Improved Outcomes

Steven H. Shaha, PhD, DBA

Prof, Center Public Policy & Administration Principal Outcomes Consultant, Allscripts

Steven H. Shaha, PhD, DBA

Professor, Center for Public Policy & Administration Principal Outcomes Consultant, Allscripts

Former Dir. KLAS Research/Performance Insights

125+ peer-reviewed publications, 350+ peer-reviewed presentations, 3 books Advisory and consulting work for 11 govt.s in Asia, Australia, Europe and No. America Advisory and consulting to over 50 non-healthcare organisations, among them: Disney, Ritz-Carlton, Coca-Cola, New Line Cinema, IBM, AT&T, Time Warner Employment history includes: Coca-Cola, RAND Corporation, UCLA Medical Center, Intermountain HC, Gartner Education: PhD, Research Methods & Applied Statistics DBA, Business Administration (PhD) MA, MEd, BS

Disclosures & Bio

• Prof. Steven H.

. Shaha, PhD, DBA

Sample of Peer-reviewed Journals

• Advance for Health Information Executives
• Advances in Patient Safety
• Agency for Healthcare Res & Qual (AHRQ Journal)
• American Journal of Ob & Gynecology
• American Journal of Sports Medicine
• Applied Clinical Informatics
• Archives of Otolaryngology, Head & Neck Surg
• Breast Cancer Research and Treatment
• British Medical Journal of Quality & Safety
• Epidemiology and Infection
• Health Management Technology
• Healthcare Financial Management
• Healthcare Technology Management
• Intl. Journal of Medical Informatics
• Intl. Journal of Pediatric Otorhinolaryngology
• Intl. Journal for Quality in Health Care
• Journal of Arthroscopic and Related Surgery
• Journal of Clinical Ultrasound
• Journal of Emergency Nursing
• Journal of Mat, Fetal & Neonatology Med
• Journal of Neurosurgery
• Journal of Obstetrics and Gynecology
• Journal of Orthopedic Trauma
• Journal of Pediatric Emergency Care
• Journal of Perinatal Medicine
• Journal of Perinatology
• Journal of Shoulder and Elbow Surgery
• Journal of the Am Acad of Ped Ophth & Strab
• Journal of Ultrasound in Medicine
• Journal of Ultrasound in Ob & Gynecology
• Laryngoscope
• Nurse Executive Watch
• Nurse Leader
• Nursing Economics
• Pediatric Critical Care Medicine
• Pediatric Emergency Care
• Pediatrics
• RN Magazine
• Spine
• Intl. Journal of Pediatric Otorhinolaryngology
• The Journal of Bone & Joint Surgery
• Ultrasound in Obstetrics & Gynecology

Main Concepts

• Alert Fatigue is real and harmful …

… and needs to be addressed and reduced

• Alerts are Important and “intentional” …

… and merit/require responsiveness BUT often a short cut by non-clinicians

• When Alerts are favorably impactful everyone

benefits … … right?

Signal to Noise Ratio

Notify

Improved Outcomes

Individually and collectively

For Clinicians For Responders To Systems From Devices From Systems For Clinicians For Responders To Systems

The Evolution of Health Care Information Retrospective Automated Reporting Concurrent Synchronous Care Prospective Predictive Care Retrospective Manual Reporting

Real-Time Outcomes Optimization

Best care and best quality scores optimized while the patient is still in the bed

The Power to Change Outcomes The best in advanced Clinical Decision Support

Continuous Clinical Decision Support Engine

Clinical Intelligence and Automated Assessments

Extracted from multi-disciplinary documentation

“Our” Recommended Treatments

e.g. Order sets, pathways, algorithms, evidence-based medicine, inter-disciplinary documentation templates

Providing the best care and achieving

• ptimal quality and performance

From Clinicians To Patients Patients Practitioners “Competitors” To “Competitors”

0.0% 2.0% 4.0% 6.0% 8.0% 10.0% CPOE Non-CPOE

Comparative Alert Rate: CPOE vs. Non-CPOE

Intelligent Order Sets vs. CPOE Alone 31.2% fewer alerts for Users of Intelligent Order Sets w/CDS (p<.0001)

CPOE Alone Intelligent Order Sets CPOE Alone Intelligent Order Sets

Safer Physician Practices

0.0% 5.0% 10.0% 15.0% 20.0% 25.0% CPOE Non-CPOE

Comparative Alert Response Rate: CPOE vs. Non-CPOE

Intelligent Order Sets vs. CPOE Alone

CPOE Alone Intelligent Order Sets

Safer Physician Practices

122.8% more responsive to the fewer alerts they received for Users

• f Intelligent Order Sets w/CDS

(p<.0001)

0.00% 0.20% 0.40% 0.60% 0.80% 1.00% 1.20% CPOE Non-CPOE

Errors "Reaching the Patient": CPOE vs. Non-CPOE

Intelligent Order Sets vs. CPOE Alone

CPOE Alone Intelligent Order Sets

Safer Physician Practices

59.2% fewer medication-related errors reached patients when managed on Intelligent Order Sets w/CDS (p<.0001)

Estimated £211,000 cash released annually for PICU alone

More than double the error rate (p<.001)

Patients managed through Intelligent Order Sets w/CDS experienced significantly fewer errors (p<.0001)

• 31.2% fewer alerts (p<0.001)
• 122.8% increased responsiveness to alerts (p<0.001)

double+ versus pre-computer-computed dosing recommendations

• Most importantly
• 59.2% lower medication-attributed errors and incidents (p<0.001)

Peds …

Signal to Noise Ratio

Meaningful notifications requested versus alert fatigue …

Forcing adoption versus proving value …

Improved Outcomes

Improved Outcomes

• Patients arriving sicker
• Less sick patients – other care pathways
• Trend onerous …
• What can be done?

• Constancy of 4-per-day
• Impact assessment …

Pre EPR Post

Improved Outcomes

Pre EPR Post

Impacts & Benefits with the EPR:

• Despite increased patient severity (80% more, 1.67 to 3.0)
• Controlled and decreased negative outcome

100% lower pre month vs. recent low (3 to 0) 33.3% lower pre-post year-to-year (3.0 to 2.0)

Impacts & Benefits with the EPR:

• Despite increased patient severity (105.5% more, 1.67 to 3.42)
• Controlled and decreased negative outcome

75.0% lower pre month vs. recent low (4 to 1) 25.2% lower pre-post year-to-year (3.42 to 2.56)

£116,000 Annual Cash Release

3.42 Mean to 2.56 Mean

Echo effect …

Improved Mortality Rates with MEWs

Pre EPR Post

Hosp Std Mort Ratio

Impacts & Benefits with the EPR:

• Decreased negative outcome

16.7% lower pre-post year-to-year (103.7 to 86.3)

103.7 Mean to 86.3 Mean

Falls with Injury

• 45.7% reduced year-over-year from Pre-baseline
• 11 fewer falls with injury

Pre EPR Post

Falls

• 14.1% reduced year-over-year from Pre-baseline
• 13 fewer falls

• 50.5% reduced from Pre-baseline

Grade 2

Pre EPR Post

Avg Grade 2 PUs ppd

• 78.9% reduced from Pre-baseline

Grade 3+

Pre EPR Post

Zero Grade 4

Avg Grade 3+ PUs ppd

Signal to Noise Ratio

Lessons Learnt

• EMR/EPR cannot merely be “electronified paper”
• Programmability and Adaptability
• Rigidity vs. Openness
• Local autonomy vs. dependency on vendor prioritisation
• Community Connectivity and Interoperability across Settings & “HIE”
• Local Innovation … Reflecting Broader Standardisation
• No org or facility is identical – patient pops, layout, personnel, specialties …
• Access to Clinical Data
• Ad hoc, investigational, curiosity, personal/specialty improvement and optimisation
• Test new implementations, verify impact and refinement as needed, continuous

improvement

• Routine reports formatted locally or centrally and used as needed
• Outcomes-driven
• Clinical
• Efficiency
• Cash Releasing and Cost Reducing
• Stakeholder satisfaction

24

Generalizable …

Clinicians collaborating with their own IT professionals

Remember the MEWs and Arrests?

26

Why Why does does

• Cl

Clinician inician En Engage gagement ment and and

• Lo

Local cal Pr Prog

• grammability

rammability Wor Work k Bet Bette ter? r?

Levels of Information: Needs and Perspectives

• Routine change
• Evolving figures - Improvement
• Operational perspective
• Fluid change / Live figures
• Real results – NOW or catastrophic
• Process perspective
• Slow change
• Stable figures
• Strategic perspective

Levels of Information: Needs and Perspectives

Happy Ending …

Main Concepts

• 31.2% fewer alerts (p<0.001)
• 122.8% increased responsiveness to alerts (p<0.001): double+

versus pre-computer-computed dosing recommendations

• Most importantly, 59.2% lower medication-attributed errors

and incidents (p<0.001).

Signal to Noise Ratio

33