Reducing Patient Harm with Lean Systems

Project details

  • Date

     September 19, 2019

Reducing Patient Harm with Lean Systems

The Problem

According to the Institute of Healthcare Improvement (IHI), preventable harm in health care is a public health crisis, with estimates placing it as a leading cause of death in the United States. Let that settle in for a second. For healthcare administrators across the country, this crisis translates into an overwhelming amount of malpractice costs and erodes already razor-thin margins.

The National Patient Safety Foundation (NPSF), now formally a part of IHI, has introduced an evidence-based approach that, according to them, identifies effective, replicable interventions for effective propagation across the health care system.

Many healthcare systems have effectively implemented many of the components of this framework; primarily, they set goals, measure/monitor, and identify causes and interventions. However, where they fall short is in creating systemic interventions (with a focus beyond just the clinical elements) and making those interventions stick. At W3 Group, we have identified three main root-causes as to why this occurs:

  1. Interventions do not coordinate all of the workers involved in patient care, including the patient
  2. The focus on measure and monitor is on outcome measures
  3. Traditional education and training are ineffective

Now, let’s explore how our evidence-based management system addresses these.

The Solution

To understand how to address these root-causes, let’s use a specific and recent client example.

Managing Patients with Diabetes. More and more people are developing diabetes which presents a considerable challenge to healthcare systems. Primarily because patients with diabetes present a higher risk and, thus, must be better managed. For instance, in one healthcare study, wrong dosage or delayed/omitted dosage of Insulin accounted for 46% of harm to diabetic patients. At one of our current clients, we conducted a current state analysis of how healthcare organizations typically manage their diabetic patients. Here is what we found.

Current State of Diabetic Patient Care. Our client’s goal was to reduce incidents of harm and measure A1c levels in their diabetic patients. To accomplish this, they implemented a policy to administer Insulin within 30 minutes of a finger-stick to ensure the insulin dosage would be correct, and there would be no issue with omittance. In addition to this process, they coordinated patient meal delivery to ensure the patient would have a proper meal to eat per their individualized protocol.

Failure Modes
  1. Interventions do not coordinate all the workers involved in patient care, including the patient.
    Upon closer observation, our team found that Technicians had a difficult time finding the two glucometers to administer the finger stick because there was no common knowledge of who was using them. Additionally, the two glucometers that were used to monitor the A1c levels had issues uploading data, which caused further delays in the process. Our teams also observed complexity in the administration of the Insulin. The Administration of Insulin required two nurses, but the primary nurse was often left to look for the second nurse, and this time looking for the second nurse often led to missing the 30-minute window. Because of the delay in the administration of the Insulin, meals were often left by the nutrition team and would be cold by the time they were ready to be delivered, leading to poor customer satisfaction.
  2. The focus on measure and monitor is on outcome measures.
    Although there were policies in place to administer Insulin within 30 minutes of a finger stick, a process that directly impacts the patients’ A1c, this process was not being tracked and led to unnecessary incidents of harm. We discovered that, when the data was pulled, the teams responsible for the patients’ Insulin injections were successful at administering within the 30 min window, only 40% of the time!
  3. Traditional education and training are ineffective.
    In the dissection of the current state process, we also found that the Technicians were not administering finger sticks consistently and were even performing the work in a way that compromised the readings. In observation of the nurses, we found the same inconsistencies in the way Insulin was administered. This lack of standard was directly traced back to the discrepancies found in the education and training of each team member, as you’ll see in our future state successes.
Future State

In the agreed-upon future state, we addressed each one of these failure modes as follows:

  1. Interventions do not coordinate all the workers involved in patient care, including the patient.
    To address the inconsistencies found in the coordination of resources, morning team huddles were used to communicate the new target pattern for the administration of Insulin. This consistent communication mechanism allowed the technicians to coordinate the use of glucometers so they would know where it is throughout the day. To further facilitate this new pattern, we deployed a lean tool called 5S which helped ensure the glucometers were maintained, functional, and organized in a common area at all times. This new morning huddle also enabled nurse pairs and the nutrition team to proactively coordinate the delivery of Insulin and food to ensure the 30-minute window was consistently met, and patients were kept happy and well taken care of.
  2. The focus on measure and monitor is on outcome measures.
    To combat the issue with insulin administration, our team worked with the nurses and technicians to implement and add a new process metric for the percentage of time insulin was delivered on time (within 30 minutes of finger stick) to the unit visual management board. This new process metric would now become a part of their daily team huddles so they could transparently iterate towards improvements. After all, teams cannot manage what they are not regularly measuring.In addition to modifying their huddle board, visual management was added throughout the unit to assist each team member in the execution of the new standard and established work pattern. Specifically, we equipped each patient room with a three categories visual indicator: red (for finger stick), yellow (for insulin delivery), and green (for food delivery). When Technicians administered the finger stick, a small timer was placed in the red. At 20 minutes, it would go off to alert the two nurses that they had 10 minutes to administer the finger stick. At 30 minutes, the alarm would sound, signaling the finger stick would have to be re-administered if the timer was still in the red zone. This form of error-proofing guaranteed that Insulin would never be administered after the 30 min time window. When the nurse pair delivered Insulin, they would move the timer to the yellow zone. Finally, when food was delivered, nutrition would move the timer to the green.
  3. Traditional education and training are ineffective.
    To address the incongruent training each team member received, our coaches did several things. First, they installed Job Instruction Training, also known as Training Within Industry. This tool was first developed and used during World War II to address the need to train people and increase productivity quickly. They found after a short time, this technique increased productivity levels by 80% and helped workers execute the work they were assigned correctly, safely, and conscientiously. To accurately determine where each team member stood regarding their ability to perform the job successfully, a skills matrix was developed and used to track all the technicians, nurses and nutrition staff that could perform the process to the standard.Additionally, a complementary tool called Leader Standard Work was deployed so that the charge nurse and unit manager had protected time to observe the process. This non-negotiable protected time was a vital step because it enabled leadership to regularly evaluate the process to ensure work was being performed correctly, safely, and conscientiously. If team members were not doing work according to the agreed-upon standard, they were retrained using the Job Instruction protocol. When team members and leadership observed obstacles in the process, they were captured using Pareto Analysis, and the unit manager would take ownership of implementing experiments to overcome those obstacles using PDSA (Plan Do Study Act) cycles of learning and innovating.
The Results

Upon implementing this solution, the process adherence resulted in the percentage of Insulin delivered within 30 minutes of a finger stick to increase from an average of 40% to a high of 100%. Immediate improvements are exciting; however, long-term sustainable improvements are what we strive towards. In this particular example, we just hit 100% after a month and a half of iterations and incremental improvements. Over the last six weeks, we have tremendously improved, and within the next couple of weeks, we should be hitting our target of 85%.

Image: The following image depicts the improvements made to the process of completing the administration of Insulin to the patient within 30 minutes of a finger stick.


As you can see, here at W3, our evidence-based management systems go beyond surface level solutions. Our evidence-based management system involves a series of tools that comes together to create an evolved system that ensures delivery of care, decreased incidents of harm, and improved margins. Please give us a call if you would like to implement such systems that can extend beyond the management of patients with diabetes to any healthcare problem.