How must healthcare organization managers evaluate quality if they wish to be as effective as managers in non-healthcare industries?
They must do so
(1) in 'real time'
based on
(2) statistics (see above and below) and
(3) experiments (see below).
(Please consider using the subtitles provided while watching the YouTube video.)
Quality Management in Radiology
Below are examples of what I can do for your radiology department:
Know when to try to reduce errors, and know whether the effort worked.
You can see that in 5/2016, we had special cause variation leading to an increased number of chest X-rays with missed nodules.
In retrospect, I identified this as being because of an increase in clinicians providing poor clinical history and radiologists reading images without a search pattern that included looking specifically for nodules each time.
Working under the service chief, I implemented a policy that required clinicians to submit to the radiologist documentation specifying whether any patient had a history of smoking, hemoptysis, weight loss, or clinical suspicion of cancer, which prompted radiologists' including nodules in their search pattern.
Starting in 6/2016, the trend in missed nodules decreased a rate mathematically improbably to be due to chance, showing that the policy did improve the nodule detection rate.
Identify the most relevant cause of errors.
A practical evidence-based approach to error reduction requires finding the most relevant causes of errors based on what the errors were and their associated peoples/places/times.
After doing so, selecting the team of individuals to further brainstorm solutions and direct policy to fix the problem will be appropriate. Selecting such a team prior to identifying the what/who/where/when is a waste of resources.
You can know whether your attempt at improvement has worked if you graph the trend in this data over time and the relative percent of the category you have tried improving decreases compared to the other categories.
Use visual dashboards.
Here is a (mostly visible) example of a control chart dashboard showing biweekly 1) patient satisfaction and 2) complication rates categorized by the Society of Interventional Radiology 2017 criteria.
Having such a dashboard better visualizes trends than looking at spreadsheet numbers.
A balanced dashboard that includes inversely proportional measures (e.g., time per patient visit and patient satisfaction) helps achieve a target middle ground.
Know where to focus your improvement effort.
In order to improve quality, managers should know:
- whether differences in performance by individual units/teams/individuals are statistically significant or not.
- whether variations in performance are because of common cause variation or special cause variation.
From this chart, what can you deduce about:
- each of these HCPs working in the interventional radiology department?
- what the manager did/should have done to improve performance?
Answer questions while avoiding unnecessary trial and error.
- Managers need to figure out which factors (i.e., variables) cause outcomes of interest.
- Both 'one-factor-at-a-time' and 'trial and error' methods of experimentation are (1) unlikely to show the best practice and (2) cost more money and time in the long run than studying the combined effects of factors of interest in a well-designed experiment.
- A scientific experiment has a control group and an experimental group. Healthcare managers can perform either (1) a true (scientific) experiment or (2) a quasi-experiment to solve real problems. The latter is more likely to be fast and keep costs low.
- When a true experiment is unfeasible, managers should use a fractional factorial design or subtype thereof (e.g., Plackett-Burman or Taguchi) to illustrate the important variables and their interactions while using the least resources, as in the real example below.
An effective manager makes decisions about what is best based on evidence, not expert opinion.
An opinion becomes expert when the opinion is highly familiar with the evidence.
Reliable evidence comes from 'experiments,' not 'experience' per se.
Exempli gratia:
- A healthcare organization has too many errors and complaints regarding radiological report communication for incidental findings.
- Rather than relying on the chief medical officer's' 'authoritative opinion,' the quality manager conducted an experiment of four categorical factors divided into two levels using only ten experimental runs to show which option was the best, two runs each for five clinical departments.
- phone call... from a radiologist's assistant to a clinician's assistant vs. from radiologist to clinician
- text message via a HIPAA-compliant software... with vs. without a return acknowledgment of receipt
- fax... with vs. without a return acknowledgment of receipt
- EMR message... with vs. without a return acknowledgment of receipt
- The manager acquired quantitative feedback from (1) five clinical department chiefs and (2) the radiologist sending the communication regarding their satisfaction with the permutations of the different communication options.
Regarding this quasi-experiment:
- Can you interpret the confounding relationship between the phone call and EMR message factors?
- Can you tell whether an additional experiment needs to be performed to confirm the validity of this trial?
In general:
- Do you know (1) which math to use for control charts for the different data types (2) different ways to express a given set of data using control charts, and (3) how different distributions of data depend on factors (1) and (2) in order to make a control chart that will answer the clinical or business question you want to answer?
- Do you think a physician who knows how to implement evidence-based quality measures will add less value to your organization than one who can only generate RVUs?
If the answer to any of these questions is "No," then consider hiring a Certified Professional in Healthcare Quality.
assertions from
Peter Drucker
about managing
quality
"The purpose of business is to create and keep a customer."
The Practice of Management (1954), p. 37.
“Efficiency is concerned with doing things right. Effectiveness is doing the right things.”
“Unless we determine what shall be measured and what the yardstick of measurement in an area will be, the area itself will not be seen.”
What organization sets standards for healthcare professionals and non-healthcare professionals alike in managing healthcare quality?
