Implementing Change

“Inadequate systems produce inadequate results.”

“Every system is perfectly designed to get the results that it gets.”

A model of systems analysis was developed by British psychologist James Reason to understand the nature of preventable adverse events. By studying industrial accidents across diverse fields, Reason determined that adverse events are almost never caused by individual errors in isolation. Conversely, he recognized that the majority of serious errors were caused by embedded flaws in a system or environment. The “Swiss Cheese” Model of Harm describes this, depicting flaws in the environment as holes in the cheese, which when lined up with human error, can lead to disastrous consequences.

As seen in the image of the model below, in poorly designed systems, inadequacies line up and build on each other, with the end result being harm or loss.

Impact:

By improving the quality and reliability at each step in the process we were able to see dramatically improved results. The chart to the right shows the steady decrease in rates of medication errors on both the pilot unit, as well as the spread units from 2006 through 2009.

"Unpacking" the components of “change"

1. How we did it/ How you can do it

Focus on reliable systems & processes
- Medication errors
- Sepsis
Followed evidence-based practice
Decisions based on hard data

2. Why it works

Goal-directed with specific aims
Institution-wide approach
Identifies sub-optimal systems (cross departmental)
Real-world tests (PDSA cycles) identify viable fixes (small steps, measured and validated)
Iterative process enables constant refinement

3. What it looks like in practice

Teamwork (Core Resource Teams / Unit-based Deployment Teams)
Front-line leadership
Executive & MD Sponsorship

The INLP formula for change relies on three primary components to support a sustainable improvement: sponsorship and support from the executive level and doctors; developing frontline leadership; and implementing evidence-based practices.

A Systematic Process for Managing Sepsis:

Based on our success with reducing medication errors, we asked, "Can the Same Approach Be Applied to Sepsis?" Clinical Evidence suggested that it could, and that sepsis diagnosis & treatment could be managed as a repeatable, standardized process. Over the following 2 years we learned that creating reliable processes could dramatically improve results, and Early Goal-Directed Therapy (EGDT) works.

Using the Swiss Cheese Model of Harm, we were able to map out the different processes that, when lined up, could result in potentially avoidable patient harm (in this case, sepsis mortality).

The INLP Sepsis project goals included:

15% reduction in the mortality rate from severe sepsis through improved early recognition and treatment within 18 months
Use of screening tool 95% of the time
Use of “mini” bundle section of SSB/EGDT 85% of the time

An interdisciplinary steering committee of nurses, pharmacists, administrators, physicians, and laboratory directors was chosen from each hospital. This committee was responsible for oversight of the work done at the unit level, and driving the program hospital-wide.

Four core strategies were used to reduce sepsis mortality:

Each hospital attended a series of leadership training seminars.
Each unit was trained on the importance of identifying severe sepsis and septic shock. Sepsis screening of all patients was implemented, along with diagnostic testing according to protocol.
Teams implemented the “sepsis bundle” on appropriate patients. Timely treatment was based on key elements of Early Goal-Directed Therapy (EGDT), including:
- Obtaining a lactic acid level, and if elevated, or if the patient is hypotensive (or believed by the medical staff to be severely septic), then obtaining blood cultures
- Obtaining a lactic acid level, and if elevated, or if the patient is hypotensive (or believed by the medical staff to be severely septic), then obtaining blood cultures
- Administering fluids and antibiotics
- Inserting a central line to measure patient response to therapies. (optional measure for this grant)
Ongoing data review: coded data was captured and submitted by each participating unit.
- Final program measurement: April 2011

Errors come from bad teamwork and a toxic hospital culture, something that is endemic to the entire health care system…Checklists are useful, but they’re not Harry Potter’s wand.

-Peter Pronovost, MD, PhD

Evidence-based Practice:

The best available clinical practice
Supported by research
Proven effective
Systematic & rigorous
Standardized protocols
NOT ad hoc
Measurable

Frontline Leadership:

Process Redesign

Identifying successful innovations, Making process changes, Developing tools, Identifying work flow changes needed, Identifying tests of change, Rapid cycle testing

Data Management

Proper coding for data collection, Data collection processes, Statistical analysis, Data reporting, Using data to validate tests of change

Communications

Goal-based communications, Branding, Promotion & Awareness, Communications Program, Development, Presentation training, Stakeholder analysis

Change Management

Organizational influence, Executive engagement, Policies & procedures, Team building, Peer sharing and adoption of innovation, Implementing workflow changes

Executive & MD Sponsorship:

Expertise & Clinical Knowledge
Keepers of protocols & procedures
Heads of departments
Encouragement & Support
Advocacy for positive change
Collaboration

"

Focus on reliable systems & processes

While it may seem like the long-route, focusing on process instead of just the end result is important. When there are multiple steps within a certain process, the probability of successful implementation is reduced significantly. As you can see from the chart below, mistakes add up with increasingly complex processes.

The six safety processes included:

Compare medication to the medication administration record (MAR).
Keep medication labeled throughout.
Check two patient identifications (IDs).
Explain drug to patient (if applicable).
Chart immediately after administration.
Protect process from distractions and interruptions.

Barriers to Improved Sepsis Process:

Perception Barrier:

Need to upgrade sepsis to critical event (e.g., AMI, trauma, stroke)

Institutional barriers:

Interdepartmental communication (MD, RN)
Department collaboration (Lab, Pharm)
Silo mentality between the ED, ICU and other floors

Professional barriers:

Variation in expertise with HCPs
Lack of expertise acknowledgment
Cross coordination between disciplines
Professional Role challenged

Time lag:

Between care elements & between departments

Data:

Real-time and useful information not available

Amended from Rivers & Ahrens

because you can’t get better results without putting a better system in place.

We focus on process improvement and reliability across the entire system...

reliable systems

Med Errors

Sepsis

A critical piece of the INLP Sepsis improvement process was that it involved ALL Departments, including:

Nursing
RRT
MDs
Units(ED, MICU, MedSurg, etc.)
Senior Leaders
Pharmacy
Labs

Only by working cross-departmentally were we able to identify each step of the process, and create a standardized, and reliable system for sustained, improved care for severely septic patient care.

A Systematic Process for Reducing Medication Errors:

The Medication Administration project goals included:

Improve Administrative Medication Errors by 50% from baseline
Improve process measures by 80%
Improve outcome measures by 50% (harm measures)

Project teams each worked on six safety processes to improve the accuracy of medication administration. These processes were chosen based on their endorsement by the California Nurse Outcome Coalition (CalNOC), a collaborative alliance for nursing outcomes. Each of the processes were derived from evidence-based literature.