By: Steve Meurer PhD, MBA, MHS
I teach Quality and Safety in Healthcare to Masters in Health Administration (MHA) students. Most of the students come to the MHA fresh off of their undergraduate studies, so my first class is typically met with confusion on why they are learning about what they believe to be the purview of clinicians, and why it is being taught by a non-clinician. I, therefore, always begin with definitions. Safety is the absence of harm broadly to include not only complications and clinical mistakes, but also financial, operational and experience harm. For example, a patient who receives 2 or more chest x-rays in a 4-day stay is considered harm even if there was no physical complication.
As for quality, I define it with one word – improvement. While many health systems have complicated improvement efforts with multiple improvement departments, those that continually outpace other health systems in outcomes have centralized the support of improvement in finance, operations, quality, safety and satisfaction. Unfortunately, most health systems will divvy improvement in metrics to different departments – mortality to the quality department, length of stay to the performance improvement department, satisfaction to the experience department and so on. Because all of these metrics are connected, it creates confusion and duplicate efforts.
I then turn to Avedis Donabedian who developed and detailed the structure, process, outcome model, which has been the prevailing model of quality/improvement in healthcare since the 1960’s. Essentially, the model suggests that if you want to change an outcome, you will need to make process changes but those will not be successful without the appropriate structure. While clinicians provide the care to patients, developing a structure that supports the clinicians, and motivating necessary change are the responsibility of health administrators.
I further delve into the non-clinician’s role in quality and safety by distinguishing two types of improvement work in the provision of healthcare. The first is conducting a study to determine whether one treatment works better than another. The second involves driving improvement in the health system’s balanced scorecard or what I call ‘moving the metrics’. The metrics are typically the same and include risk adjusted mortality, length of stay, cost, readmissions and complications as well as patient satisfaction. This first type of improvement work is conducted by clinicians and researchers and the second is led by health system leadership and supported by those who are schooled in 5 concepts: the healthcare ecosystem, curiosity, variation, problem solving and storytelling.
Note that clinical or statistical knowledge are not included. Those are absolutely necessary for the first type of improvement work, but can sometimes be a barrier to improvement in moving the metrics. The easier of the two to explain why is the statistician because they are concerned with strong p-values which demands large sample sizes. The health system improvement leader will never have the sample size to achieve statistical significance and therefore needs to understand Walter Shewart’s work around variation.
There are two reasons why a non-clinician can be better for moving the metrics. First, the MHA degree provides the skills, knowledge and tools in those 5 concepts that are needed to drive improvement in a health system much more than the curriculums that clinicians take. The second is in understanding the difference between information and insights. The process for moving the metrics starts with the development and the distribution of the health system’s balanced dashboard on a monthly or quarterly basis. The balanced scorecard is information that tells health system leadership that they are different, not why they are different. Health systems who spend their time using business intelligence tools to create an interactive, visually pleasing dashboard only to hand it directly over to leadership are missing the important step of transforming the information into insights, which tell leadership why a metric shows an opportunity. Particularly for quality and safety metrics, many would say that a non-clinician cannot provide the why. However, insights come in 6 flavors: clinical quality, harm, data quality (documentation and coding), patient/hospital/community characteristics, throughput, and definition interpretations. Clinicians tend to focus on the first two which directly leads to the first type of improvement (literature review and / or conducting a study). However, particularly with metrics developed using administrative coded data, the difference between one health system and another can be found in the last 5 flavors.
For example, a hospital wanted to improve in hospital sepsis mortality. Without drilling into the metric to understand why they were different from other hospitals, the clinicians focused on improving antibiotic timing. Over an 8-month period, the hospital’s appropriate antibiotic timing improved from 82% to 88%. After a statistician said it was a statistically significant difference, the hospital celebrated. Unfortunately, in hospital sepsis mortality over the same time increased. Drilling into this metric revealed the following: the clinician’s interpreted ‘shock’ differently from other hospitals and therefore were not capturing as a comorbid condition which lowered expected mortality, and the LOS of sepsis patients that died in the hospital was 3 days longer than the average from a group of similar hospitals. While we never want to stop the clinical quality improvements (antibiotic timing), the changes needed to improve risk adjusted sepsis in hospital mortality are in data quality, definition interpretations and throughput.
It is important to note that changes that do not involve clinical quality or harm are many times derided by clinicians. In a recent conversation, two physicians claimed a top performing hospital in mortality was purposely over capturing fluid and electrolyte disorders as a comorbid condition and keeping dying patients in the ED so they would not be counted as an inpatient death. They were calling this poor care. The top ranked hospital, however, responded by saying that they have made it easy for physicians to document mortality predictors on admission so that they have more time to spend with patients, and that they do keep near death patients in the ED because they have a space where they have ramped up palliative services and that it is a better death for the patient, the family and the inpatient clinicians who do not want to be admitting and caring for a patient only to have them die within hours of the admission.
Both types of improvement have to exist together. The theory is that hospitals with a greater number of clinicians trained to do clinical studies will be more apt to engage in moving the metrics. Conversely, you cannot move the metrics without engaged clinicians to help understand the why, and to participate in the change. The American Health System is not experiencing widespread improvements in metrics commonly found on balanced dashboards. There are many reasons for this, but I try to instill into my MHA students that it is their role and responsibility to be leaders that can move the metrics.
About the Author:
Steve Meurer, PhD, MBA, MHS, is the Director at the Center for Healthcare Analytics & Improvement, in the Armstrong Institute for Patient Safety & Quality, as well as the Director for the Masters in Health Administration Program at Johns Hopkins Medicine, and Johns Hopkins Bloomberg School of Public Health respectively.
The views and opinions expressed in this post are those of the author(s) and do not necessarily reflect the official policy or position of The University of Ottawa. For more details on our site disclaimers, please see our ‘About’ page