4Reliability.qxd
Improving the Reliability
of Health Care
Innovation Series 2004
4
Copyright© 2004 Institute for Healthcare Improvement
All rights reserved. No part of this paper may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopying, recording, or by any information storage
or retrieval system, without written permission from the Institute for Healthcare Improvement.
Acknowledgements:
The Institute for Healthcare Improvement (IHI) is grateful to the following individuals and groups
for their support and contributions to this work:
The members of IHI’s IMPACT Network
The members of Pursuing Perfection, a Robert Wood Johnson/IHI initiative
IHI also thanks staff members Frank Davidoff, MD, Jane Roessner, PhD, and Val Weber for their
editorial review.
For reprint requests, please contact:
Institute for Healthcare Improvement, 20 University Road, 7th Floor, Cam
idge, MA 02138
Telephone XXXXXXXXXX, or visit our website at www.ihi.org
We have developed IHI’s Innovation Series white papers to further our mission of improving the
quality and value of health care. The ideas and findings in these white papers represent innovative
work by organizations affiliated with IHI. Our white papers are designed to share with readers the
problems IHI is working to address; the ideas, changes, and methods we are developing and testing
to help organizations make
eakthrough improvements; and early results where they exist.
Improving the Reliability
of Health Care
Innovation Series 2004
Authors:
Thomas Nolan, PhD: Senior Fellow, IHI; Statistician, Associates in Process Improvement
Roger Resar, MD: Senior Fellow, IHI; Assistant Professor of Medicine, Mayo Clinic College of Medicine;
Change Agent, Luther Midelfort Mayo Health System, Eau Claire, Wisconsin, USA
Carol Haraden, PhD: Vice President, IHI
Frances A. Griffin, RRT, MPA: Director, IHI
Editor: Ann B. Gordon
Innovation Series: Improving the Reliability of Health Care1
Executive Summary
Reliability principles are used successfully in industries such as manufacturing and air travel to help
evaluate, calculate, and improve the overall reliability of complex systems. Reliability principles, used
to design systems that compensate for the limits of human ability, can improve safety and the rate at
which a system consistently produces desired outcomes.
Reliability is measured as the inverse of the system’s failure rate. Thus, a system that has a defect rate
of one in ten, or 10 percent, performs at a level of 10-1. Studies suggest that most US health care
organizations cu
ently perform at a 10-1 level of reliability.
The Institute for Healthcare Improvement (IHI) uses a three-step model for applying principles of
eliability to health care systems:
1. Prevent failure (a
eakdown in operations or functions).
2. Identify and Mitigate failure: Identify failure when it occurs and intercede before harm is caused,
or mitigate the harm caused by failures that are not detected and intercepted.
3. Redesign the process based on the critical failures identified.
Within each step of this model, specific reliability principles and change concepts can be applied
to reduce ambiguities and opportunities for e
or, and improve the reliability of the processes used
to support care.
Using the Prevent, Identify-and-Mitigate, Redesign approach, IHI has created a template for increasing
eliability of care for heart failure (HF) patients. Since a number of quality assessment and accreditation
organizations are using quality measures for heart failure care, as well as promising or considering
financial reward for those who achieve top performance, a template for improving reliability of heart
failure care is an important tool.
IHI urges hospitals to increase their efforts to improve the reliability of care by adopting or adapting
the principles of the heart failure care template presented in this paper. The template presented is
not meant to be the only or the best way to improve the reliability of heart failure care, but gives an
example of how the principles can be employed.
© 2004 Institute for Healthcare Improvement
2Institute for Healthcare Improvement Cam
idge, Massachusetts
Introduction
It is a widely held view that the American health care system does not perform nearly as well as it
should or could. Recent studies show widespread inconsistency in the delivery of high-quality care.
In particular, two studies by RAND Health found that Americans with common health problems
eceive only about 50 percent of recommended care.1,2
These studies confirm an earlier assessment of the state of US medical care by the Institute of Medicine
(IOM). In 2001, the IOM published an influential report designed to guide efforts to improve the
system. Crossing the Quality Chasm: A New Health System for the 21st Century calls for fundamental
change, organized around six aims for improvement. The IOM says health care should be:3
Safe: Patients should not be harmed by the care that is intended to help them.
Effective: Care should be based on scientific knowledge and offered to all who could benefit,
and not to those not likely to benefit.
Patient-Centered: Care should be respectful of and responsive to individual patient preferences,
needs, and values.
Timely: Waits and sometimes-harmful delays in care should be reduced both for those who
eceive care and those who give care.
Efficient: Care should be given without wasting equipment, supplies, ideas, and energy.
Equitable: Care should not vary in quality because of personal characteristics such as gender,
ethnicity, geographic location, and socio-economic status.
Many health care organizations have em
aced the challenges set forth by the IOM, and are making
progress in these six areas. However, the progress still falls far short of the goal. For example, fo
treatment of community-acquired pneumonia, improvements that increase the compliance with
evidence-based practice from 60 percent of cases to 85 percent are typical. While the relative
improvement is impressive, the fact remains that a minimum of 15 percent of patients receive
substandard care; the true figure is probably much higher.
Reliability principles—methods of evaluating, calculating, and improving the overall reliability of a
complex system—have been used effectively in industries such as manufacturing to improve both
safety and the rate at which a system consistently produces appropriate outcomes.
Can reliability principles be applied effectively to improve the consistent delivery of high-quality
health care? The Institute for Healthcare Improvement (IHI) believes that applying reliability
principles to health care has the potential to help reduce “defects” in care or care processes, increase
the consistency with which appropriate care is delivered, and improve patient outcomes.
© 2004 Institute for Healthcare Improvement
Innovation Series: Improving the Reliability of Health Care3
Background
IHI is working with a number of hospitals to apply reliability principles to care processes. This work
cu
ently focuses on improving the outcomes of five diagnoses: community-acquired pneumonia; heart
failure; acute myocardial infarction; hip and knee replacement; and coronary artery bypass graft surgery.
These five diagnoses are of particular importance because they are the focus of a three-year quality
improvement demonstration project sponsored by the Centers for Medicare & Medicaid Services
(CMS), which oversees care in the US for elderly and poor, and Premier, Inc., an alliance of hospitals
and health systems. The five diagnoses are also the source of core quality indicators used by the Joint
Commission on Accreditation of Healthcare Organizations (JCAHO), the National Quality Forum,
and the Leapfrog Group, a Washington, D.C.-based consortium of private and public health care
purchasers focused on recognizing and rewarding quality.
Although the care processes for the five diagnoses are varied, they share a reliance on multiple steps
or processes, each one of which can affect the ultimate outcome.
Reliability in Health Care
Reliability is defined as failure-free operation over time. In health care, this definition connects to
several of the IOM’s aims for the health care system, particularly effectiveness (where failure can result
from not applying evidence), timeliness (where failure results from not taking action in the required
time), and patient-centeredness (where failure results from not complying with patients’ values and
preferences).
Reliability is measured this way:
Reliability = Number of actions that achieve the intended result ÷ Total number of actions taken
It is convenient to use failure rate (calculated as 1 minus Reliability), or “unreliability,” as an index,
expressed as an order of magnitude. Thus, 10-1 means one defect per 10 attempts, 10-2 is one defect
per 100 attempts, and so on. Put in terms of health care, a process measuring 10-1 fails to be effectively
applied for one out of every 10 patients. For example, if 90 percent of surgery patients get their
prophylactic antibiotic within an hour of surgical incision, the reliability of that process as measured
y defect rate is 10-1.
These levels are measures of reliability (or unreliability), but they also serve as useful labels for design
characteristics of systems. The characteristics of systems that perform at 10-1, for instance, are different
from those that perform at 10-3, which represents one defect in 1,000 attempts. It is those design
characteristics that organizations must integrate into their systems in order to improve reliability.
© 2004 Institute for Healthcare Improvement
4Institute for Healthcare Improvement Cam
idge, Massachusetts
To help describe what these levels look like in an organization, IHI offers the following framework:
10-1 performance on process measures indicates no articulated common process, and an emphasis on
training and reminders. A range of international studies of adverse events in hospitalized patients
shows a convergence around an e
or rate of 10 percent (plus or minus 2), suggesting that this is the
level at which most health care organizations cu
ently perform.4,5,6,7 (Since this e
or rate represents
an average, clearly for some tasks and processes the rate is lower, but for some, it is higher.)
10-2 performance on process measures indicates processes intentionally designed with tools and concepts
ased on the principles of human factors engineering.
10-3 or better performance on process measures indicates a well-designed system with attention to
processes, structure, and their relationship to outcomes.
To understand these performance levels in a
oader context, consider that aviation passenger safety
is measured at 10-6. Nuclear power plants must demonstrate a design capable of operating at 10-6
efore they can be built.8
It is important, however, to note that an essential aspect of reliability is the level of performance ove
time. Thinking about health care reliability simply in terms of overall defects doesn’t differentiate
eliability from the definitions of quality that are typically used in health care. While efforts to
examine defects over time in a hospital, for example, often look across patients in time, these data
epresent the aggregate experiences of different patients flowing through the system. Our definition
of reliability—failure-free operation over time—also refers to an individual patient’s experience ove
time. This is a crucial distinction, and an aspect of health care reliability that connects effectiveness
with patient-centeredness.
The measure of operation over time is depicted in the “bathtub” curve shown in Figure 1. Whether
measuring the performance of