The Model for Improvement
Every improvement is certainly a change, but every change is not always an
improvement. Making changes to the way that we do things can be timeconsuming
and can sometimes feel risky.
The Model for Improvement (Langley et al.) is a tried and tested approach to achieving successful change
1. Use of the model offers the following benefits:
• It is a simple approach that anyone can apply
• It reduces risk by starting small
• It can be used to help plan, develop and implement change
• It is highly effective
The Model for Improvement
The model for improvement was first published in 1992 by Langley, Nolan et al in ‘The Improvement Guide: A Practical Approach to Enhancing Organisational Performance’. The model provides a framework for developing, testing and implementing changes to the way that things are done that will lead to improvement. The model consists of two parts that are of equal importance:
• Three fundamental questions that are essential for guiding improvement work.
• The Plan, Do, Study, Act (PDSA) cycle to test and implement changes in real world
1. What are we trying to accomplish?
This question is intended to help you be clear about the improvements that you
would like to make, what results you would like to get and how you would like
things to be different. Having a clear vision of your aims is crucial to prevent
**Write an Aim Statement**
• Define the specific population of patients
• Clearly state the goal
Example of an Aim Statement for Improvement in Chronic Depression Management:
The clinic’s practice will be redesigned to focus care on the population of depression patients by implementing the six components of the Chronic Care Model so that at least 70% of patients being treated for depression will have a structured diagnostic assessment for major depression recorded in their medical record.
For the pilot population (all patients in North Clinic), we will achieve at least a 30% increase over baseline in the percent of patients experiencing a major improvement in depressive symptoms after 12 weeks, and a 25% increase over baseline (absolute) in the number of patients continuing depression treatment at 12 weeks after initiation.
2. How will we know that a change is an improvement?
Without measurement it is impossible to know whether you have improved. Think about how you want things to be different when you have implemented your change and agree what data you need to collect to measure it. You can do this in terms of the way in which your results or outcomes might be different, how the service that your patients receive will be better, or how your processes might change. This usually requires more than one measure. You should always include at least one balancing measure to ensure your change does not create
1) Increased use of a new patient assessment tool
Considerations for balancing measures: What if patients hate the process/tool?
What if it takes much longer than expected for the MA staff to complete these assessments? What if staff feels uncomfortable using the tool?
• Rate of no-shows for f/u visits
• Patient satisfaction with the tool
• Proportion of next appointments starting more than 10 mins after
• Staff satisfaction evaluation
2) Improvement projects targeting comprehensive restructuring of a population with a specific disease.
Considerations for balancing measures: What is happening to the rest of your patients? How is this affecting staff satisfaction? How is it affecting clinic flow?
• Check performance stability in a different population (ie – pregnant patients, or immunization registry participation, etc).
• Staff satisfaction
• Average patient value added time (1/% waiting time)
3) Tests of workflow change
Considerations for balancing measures: Changes in workflow always affect more than one person, as it is very easy to suggest ‘efficiency’ by having one person/role type do more work. If workflow is being changed to accommodate electronic information the work-arounds start with the first patient treated ‘electronically.’
• Staff satisfaction – both those directly affected and those you believe shouldn’t be affected at all.
• Average patient value added time
• Number of sticky notes generated
• Number of sticky notes used/week
• # items/patients represented in the MA “hold stack”
3 . What changes can we make that can lead to an improvement?
Change Concept: a general notion or approach to change that has been found to be useful in developing specific ideas for changes that lead to improvement.
Look at other industries/processes for ideas. How does Starbuck’s approach customization of orders? How does Fedex track information? What you and your team think is a good idea? What do others like you (competitors) do? What do your patients say they want? You can adapt ideas or be completely creative.
Concept An opportunity to create a new connection
Thought process Specific idea Be Specific idea
Remember that you know your own system best, so keep your objectives in mind and use your knowledge and experience to guide you.
After testing a change on a small scale, learning from each test, and refining the change through several PDSA cycles, the team can implement the change on a broader scale — for example, for an entire pilot population or on an entire unit.
After successful implementation of a change or package of changes for a pilot population or an entire unit, the team can spread the changes to other parts of the organization or in other organizations.
*Langley GL, Nolan KM, Nolan TW, Norman CL, Provost LP. The Improvement Guide: A
Practical Approach to Enhancing Organizational Performance.
**The Plan-Do-Study-Act cycle was developed by W. Edwards Deming (Deming WE. The New
Economics for Industry, Government, Education
The two most common problems are:
1 - Not breaking tests of change into small enough steps
2 – Not suggesting realistic measures. A realistic measure is one that can be
reliably collected during the course of your real day. If data collection is difficult or
time consuming it will be harder to engage your team and/or the team will be
tempted to proceed without measuring. If you skip measuring, the project will be
likely to fail.
Improvement is nearly always a team endeavor. Try to ensure that you involve
the right people in your work. For a small practice the team will likely be
Avoid the temptation to jump straight to solutions and never get to the root of the
problem. Using the three fundamental questions will help ensure that the project
stays on track and addresses the issue that needs to be addressed.
When planning cycles, be clear about who is doing what, where and when.
Results and satisfaction with the improvement process are dependent on how
good the plan is.
Use PDSAs consecutively to build up the information about the change being
tested and then use them to implement it systematically into your daily work.
PDSA cycles generally do not operate in isolation but should occur in a series
leading towards the goal.
Putting It All Together: Cutting Intubation Time in Half
Fort Sanders Parkwest Medical Center, in Knoxville, TN, set the following aim:
Extubate all CABG patients in less than 6 hours.
Forming the Team
The team working on reducing extubation time consisted of the following:
• Anesthesiologist: Manages anesthesia during surgery and assists with
the extubation. The anesthesiologist is a key member of any early
extubation team because early extubation is possible only with changes in
anesthetic management of the patient during surgery.
• ICU nurse: Has primary responsibility for extubating patients.
• Respiratory therapist: Works with ICU nurse to extubate patients early.
• Surgeon: Helps identify patients who could be extubated early. The
surgeon needs to be comfortable allowing patients to be extubated early
and must delegate control of that process to the ICU nurse.
• Lab personnel: Help design a process to expedite obtaining arterial
blood gas (ABG) results.
To determine if changes were leading to improvement, the team decided to track
the following measures:
• Average time from admission to the ICU to extubation (Figure 3.2)
• Reintubation rate
• Average number of ABGs per case
• Rate of pulmonary complications
• ICU length of stay
In order to get buy-in from anesthesiologists and surgeons, the team first tested
the following changes only on elective CABG patients. As the comfort level of the
physicians grew and data showed that patients could safely undergo early
extubation, the team extended the changes to include all CABG patients.
Change 1: Standardize pain management.
In order to be extubated early, patients must not be too heavily sedated. The
team began by revising the existing standards for postoperative pain
management. Instead of using the traditional high dose of morphine, the team
tested the use of smaller, more frequent doses. In this way, patients’ pain was
managed adequately, yet patients were awake enough to be extubated safely.
Change 2: Standardize anesthesia management.
Patients cannot be extubated if they are heavily sedated. The team tested having
anesthesiologists use lower doses of sedatives to prevent patients from
remaining heavily sedated long after the surgery was completed.
Change 3: Establish a nurse- and respiratory therapist-run rapid weaning and
extubation protocol. The team also developed a set of criteria that patients need
to meet in order to be extubated safely, given the changes in anesthesia and
Change 4: Reduce delays in obtaining arterial blood gas (ABG) results.
The team identified delays in obtaining ABG results and weaning parameters as
barriers to early extubation. They assigned a dedicated respiratory therapist to
obtain these results.
Change 5: Educate physicians, nurses, certified registered nurse assistants
(CRNAs), and respiratory therapists on the new goals and procedures for early
Change 6: Extend the changes from elective CABG patients to all CABG Results
Within several months, Fort Sanders Parkwest Medical Center reduced the
average extubation time for elective CABG patients from 12.2 hours to less than
5 hours. In May 1997, the changes were extended to all CABG patients, and by
October the average extubation time for all CABG patients was less than 6
Example: Use a sliding scale for insulin to decrease hypoglycemic events.
Plan: The team predicted that both the physician and nurse would find the scale
easy to use and that using the scale would not lead to any adverse drug events
involving insulin. The plan was to have one physician and one nurse test the
scale with one patient for one day.
Do: The day before the test, the team showed the sliding scale to the physician,
who agreed to the small test. The day of the test, a patient was selected and it
was agreed that all insulin doses for the day would be determined using the
• Both the nurse and the physician liked the sliding scale.
• The physician reported that the scale was easy to use and would likely
save him a lot of time.
• The nurse liked the scale but had some problems. She had to do some
of the calculations manually and found the scale difficult to follow, as the
dosages were in columns and it was hard to keep them straight.
• The patient did not experience any hypoglycemic events.
• The team decided that the nurse could highlight the column with the
correct dose range to make it easier to find.
• The nurse suggested that weight calculations for common weights could
be placed on the back of the sliding scale, for easy reference.
• The team agreed to test the sliding scale for one more day with three
patients, using the suggestion to highlight the column with the correct
dose range for each patient.
Example: Use a sliding scale for insulin to decrease hypoglycemic events.
Cycle 1: Tested a sliding scale for insulin with one physician and one nurse, for one
patient receiving insulin one day.
Cycle 2: Scale tested on three patients with the same physician and nurse for one day,
with the correct dose column highlighted
Cycle 3: The physician agrees to have all of his patients receiving insulin on the medical
unit have their doses determined by the sliding scale for one week.
Cycle 4: The physician presents the results to other physicians who frequently admit to
the medical unit, and they all agree to use it for one week.
Cycle 5: The medical unit adopts use of the sliding scale for all patients receiving insulin.
Cycle 6: Two more units begin use of the sliding scale.
Cycle 7: The medical staff approves use of the sliding scale across the hospital.
Change 1: Use a sliding scale for insulin to decrease hypoglycemic events.
Culture Change 2: Include pharmacists in daily rounds to decrease adverse medication events.
Cycle 1: One pharmacist participates in rounds on one nursing unit for one day.
Cycle 2: Test expanded to include a pharmacist in rounds on nursing unit every day.
Cycle 3: Pharmacist participation in rounds spreads to a second nursing unit.
Core Medication Processes
Change 3: Improve the medication dispensing process.
Cycle 1: Pharmacy technician checks for discontinued medications on nursing unit.
Cycle 2: Rounds tested for one day of pharmacy technician removing discontinued medications twice per day.
Cycle 3: Rounds tested for one week.
Cycle 4: Rounds for removal of discontinued medications implemented house-wide.
Cycle 5: Rounds expanded to three times per day.
Change 4: Reconcile medications at admission.
Cycle 1: Nurse in intensive care unit reconciles admission medication orders for the
next three admissions.
Cycle 2: Nurse and physician test an order form for admission orders that incorporates reconciliation on one patient.
Cycle 3: Order form is tested with same physician for one week on all admissions
Cycle 4: Two more physicians test the order form for one week.
Cycle 5: Order form is adopted by intensive care unit, for all admissions.
Cycle 6: Order form is tested on two additional patient care units for one month.
Cycle 7: Order form is approved for use throughout the organization.