Sandi Gulbransen University of Utah Health Care Frank A. Drews University of Utah Center for Human Factors in Patient Safety VA Salt Lake City Health Care.

Sandi GulbransenUniversity of Utah Health Care

Frank A. DrewsUniversity of Utah

Center for Human Factors in Patient Safety VA Salt Lake City Health Care System

Informatics, Decision-Enhancement, and Surveillance Center

Elements of an Agile Safety Culture in Health Care

Socio-Technical Systems Safety

Culture of Safety Components

Safety Space Perspectives on improvement

Principles of process improvement Task analysis and improvement in task performance

Take away

Socio-Technical Systems

Socio-Technical System (STS) Core Idea: Systems have technical and social components

Technical system Machinery, processes, procedures

Social system People and their habitual attitudes, values, behavioral styles and

relationships

Socio-Technical Systems

How to strengthen a socio-technical system? Resilience Engineering

Resilience “The ability of a system to adjust its functioning prior, during, and past

changes and disturbances to maintain operation” Hollnagel, 2011

Assumptions Performance conditions are underspecified; constant adjustment to

changing conditions is required Safety and productivity are not independent

Safety and Safety Culture

What is safety? Safety as the absence of accidents, incidents, injuries,

illnesses? Problem: What causes these events?

What is within our control, and what is not? (e.g. I can control how I drive but I cannot control how others drive).

Definition by International Organization for Standardization (ISO): “Safety is freedom from unacceptable risk.”

Issue of risk vs. uncertainty

Safety culture / safety climate Result of Chernobyl nuclear power accident (1986)

Rule violations and poor culture of safety as contributors Usually there is no intention to create unnecessary risk

But: to get the job done safety is eroded (violation) Routine violations reducing safety margins

Safety and Safety Culture

Safety Culture

Commitment Competence Cognizance

Safety and Safety Culture

Commitment Motivation to stay safe even under management changes Resources with regard to quantity and quality

Competence The technical competence to improve safety Safety information system

Cognizance Correct awareness of the threats to the organization

Safety and Safety Culture

Safety space Continuum from resilient to brittle organizations

resilient brittle

Safety and Safety Culture

Over time organizations move in safety space Position in the safety space is a function of the number of

negative outcomes Resilient organizations suffer fewer negative outcomes

Drifting towards the brittle region increases the likelihood of accidents

Public and/or regulatory pressures result in improvements of safety Moving towards the resilient region has also contrary forces

Safety initiatives may run out of steam There is a diminishing return for safety improvements

Safety and Safety Culture

How to increase resilience? Learning

Knowing what has happened Ability to address the factual

Responding Knowing what to do Responding to regular and irregular disruptions Ability to address the actual

Anticipating How to anticipate threats, developments and opportunities Ability to address the potential

Monitoring Knowing what to look for, i.e., what can become a threat in the future Focus on what happens in the environment, but also what happens in

the system Ability to address the critical

Learning(factual)

Monitoring(critical)

Anticipating (potential)

Responding(actual)

Abilities required for resilience

Examples from two perspectives Macro perspective

Principles of process improvement Example: Joint replacement

Micro perspective Task Analysis / improvement of task performance

Examples: Kit development using Adherence Engineering; ICU Display Design

Value Management System (VMS):Toward a Learning Health System

Sandi GulbransenUniversity of Utah Health Care

March 20, 2014

Outline

Value Management System (VMS) Principles of Lean and ISO 9001:2008 Use case VMS as Virtuous Cycle

Data

Process

Clinical

Patient

Today

Make the best decisions with the most complete information available – but our best information has gaps.

Tomorrow

Make the best decisions with a better understanding of our key processes and how they are performing.

Lean/Value Improvement ISO 9001:2008

Customer defined value Customer focused

Leader as teacher Leadership

Everyone solves problems Involvement of people

Understand the value stream Process approach

System optimization:reduce waste System approach to management

Pursue perfection Continual improvement

Make problem visible Factual approach to decision making

1. Document

What we do

2. Record

What we did

3. Audit

How well we do it

4. Non conformity

What didn’t work

5. Correct or Prevent

The fix

Joint Replacement:Value Driven Care Process

Physician Lead: Chris Pelt, MDSponsor: Charles Saltzman, MD

Multidisciplinary Team:NursingPhysical TherapyAmbulatory ClinicCase ManagementValue EngineeringDecision SupportEDWQuality & Patient Safety

PoliciesProceduresGuidelinesBylaws

+=Daily Activity

Chart Activity

DAILY WORK

DOCUMENTWhat we do

RECORDWhat we did

Joint Replacement Care ProcessPatient Information

Early MobilizationIncomplete Discharge Orders

+=

>

DAILY WORK AUDITHow well we do it

INTERNAL AUDITS

EXTERNAL AUDITS

No specific audits

= +

>

IDENTIFYNON-

CONFORMITYWhat didn’t work

INTERNAL AUDITS

EXTERNAL AUDITS

INTERNAL AUDITS X X X X

AUDITHow well we do it

Opportunity for Improvement

Early Mobilization Prioritized – Change of PT ShiftsUpdated discharge order setPatient Selection pre-op re: post op

SNF, Rehab, HH

= +

>

REPORT

INTERNAL AUDITS

EXTERNAL AUDITS

CORRECTThe fix

= +

>

PREVENT PREVENTHow we keep itFrom happening

INTERNAL AUDITS

EXTERNAL AUDITS

CORRECTThe fix

Perfect Care Metric - VMS

Perfect Care Composite: Joint Replacement

National Metrics Local Metrics

30 day readmission Early mobility

8 SCIP measures ED visit within 90 days

35 HAC/PSI metrics Discharge unit

Anesthesia technique

ResultsPatient Care and Average Cost

Interpret

Assemble

Analyze

Change

Feedback

Virtuous Cycles1

InterpretIdentify nonconformityor preventive action

AssembleIdentify change for sustained improvement

AnalyzeMonitor process performanceMetrics and drivers

ChangeBest practices – internal and external

FeedbackFind opportunityMake it easy to do the right thing

Agility

Understanding of key processes Metrics that will drive change Actionable information at POC Extensible to other organizations

Reactive Proactive

ProjectsPDSA cycles

Lean ManagementVirtuous cycles

Journey to Learning Systems

LHS Considerations1 VMS Elements

Evolutionary approach (less is more)

Management review (existing infrastructure)

Design for adaptation Actionable information at POC

Embrace the fractal

Attend to all parts of virtuous cycle Framework for structured approach

Attend to all domains Understanding local system as reference for extending

Engage all stakeholders

1Friedman, C. Informatics for the Nationwide Learning Health System, 1/27/2014

There is a way to do it better – find it.

-Thomas Edison

Frank A. DrewsUniversity of Utah

Center for Human Factors in Patient Safety VA Salt Lake City Health Care System

Informatics, Decision-Enhancement, and Surveillance Center

WithAaron Angelovic, Jonathan Bakdash,

Alexa Doig, Brittany Mallin

Task analysis and

improvement in task performance

Adherence and Violations

Procedure violations Common problem in many industries Routine violations

Person perceives an alternative, more efficient way to perform task Lack of feedback External (social) pressures reinforce routine violations

Violation producing conditions Perceived low likelihood of detection Inconvenience Time pressure

Design for Adherence

Central line maintenance: A trivial task? Maintenance requires more than 25 steps

Breakdowns in maintenance can result in central line associated bloodstream infection (CLABSI)

Design for Adherence

Adherence Engineering to reduce Violations

A procedure: Central line maintenance Status quo

Current equipment does not support clinicians Opportunity to redesigning the task / equipment based on

Human Factors

Design for Adherence

Building an alternative Integrating checklist into equipment to support adherence to

best practices Applying AE principles

Multi step approach Involvement of Infusion Team Members / Physicians Involvement of Manufacturer Involvement of HF Engineers Iterative design and evaluation process

Virtuous cycle

Design for Adherence

New Kit

Non-Sterile Portion

Sterile PortionSmall size

Results

Clinical data CLABSI rates

Pre-intervention CLABSI rate: 3.23 / 1000 patient line days

Post-intervention CLABSI rate: 0 / 1000 patient line days

Incident rate ratio = 0 (95% CI:0-0.63; p<.01)

Results

Best Practice Odds Ratio / 95% CI Significance

Hand Sanitization 4.86 2.45-9.62

p < .0001

Chlorhexidine Scrub Duration

7.62.26-25.59

p < .0001

Anti-Microbial Bandage Application

0.70.14-3.57

p = .69

Catheter Hub Disinfection

7.854.14-14.9

p < .0001

Pre-intervention n = 107, Post-intervention n = 85

Design for Adherence

Discussion Clear improvement in adherence to best practices Fewer item omissions / errors Reduction in CLABSI

Development of a

Patient Monitor for Critical Care

ICU display

Two step approach Semi-structured interviews with ICU nurses

Goal: Understanding the limitations of current displays

Design Involvement of nurses, physicians, cognitive

psychologists Iterative design process

Evaluation study

Interviews to inform design

Interview Focus on experience with current displays

Confusing variables Missing information Error Relevance of trend information Patient variability Usability

Results (emerging themes)

Themes Example statement

Data acquisition / processing

Data processing leads to frequent false alarms

Data / event integration Marking events as explanation for changes in vital signsOnly contextual information allows for detection of artifacts

Data interpretation Applying meaning to variables

Monitoring trends in numerical data

Trend functions are not routinely used, not accessibleDefaulting to memory for trend assessmentNeed to visualize interrelationships between intervention and

physiologic variables

Usability issues Small font sizeColor coding of variables is not consistentCables should be color coded for ease of use and troubleshooting

Interviews to inform design

Discussion Current monitor equipment does not support

integrated patient assessment Slow, piecemeal-wise processing Increases cognitive load

Information needs are not met Trend information not immediately available High information access costs

Interviews to inform design

Design process Focus on most commonly monitored patient

variables Trending information Configural approach (patient centered variability)

Septic shock

Design of the display

Study Design IV:

Display (configural vs. traditional display) 4 scenarios (Septic shock, pulmonary embolism, early sepsis,

normal)

DV: Time for nursing diagnosis Percentage of correct diagnoses Percentage of trend data being accessed in traditional display

condition

Participants 40 ICU nurses (25 female)

Evaluating the display

Results (Time to diagnosis) Significant

differences for all scenarios but early sepsis

Evaluating the display

Results (Percentage correct diagnoses) Significant

differences for for septic shock and pulmonary embolism; trend for stable

Evaluating the display

Discussion Configural display leads to

Improvement in time for diagnosis Improvement in quality of diagnosis

Up to 24 % improvement in correct diagnoses

Nurses appear not to use trend information if not readily available

Evaluating the display

Summary

To facilitate the development of a agile safety culture there is a need for both perspectives Macro perspective

Operations perspective Micro perspective

Perspective on human performance

Sustainable safety improvement only with both in tandem Learning is only possible if we allow for it

Continuous effort, continuous change Responsive