Unreliable processes can have serious consequences for any organization, but in health care, the results can be grave. That’s why quality improvement work in health care is an ongoing, unrelenting effort to continually increase the probability that our efforts will deliver the desired outcomes to our patients.
In light of this, it’s helpful to know that there are only two ways to improve a process: reduce the number of steps involved or improve the reliability of individual steps.
A process is nothing more than a collection of individual steps strung together to achieve a desired outcome. And the probability that the process will deliver its desired outcome is actually the collection of individual probabilities that each of these steps will happen when and how they are designed.
So, let’s take for example the steps that a parent might lead children through every weekday morning to get them off to school: wake them up, get them dressed, eat breakfast, make lunches, load the backpacks, brush teeth, walk to school, etc.
Each of these steps has a reliability factor associated with it. That is, the probability that this step will occur when and how it needs to occur. This reliability, of course, varies considerably with each child and within each household. If the system is not consistently delivering its desired outcome (i.e., getting the child to school on time), then it must be improved. So, given our two methods of improvement from above — removing steps or making some steps more reliable — we have options. What about, say, switching to velcro-strap shoes for the toddler? That would eliminate the arduous lacing step for one child. Or, lunch money for the cafeteria; that would remove the sandwich-fixing step. How about buying an alarm clock for the kids’ room? That might make the wake-up step a lot more reliable.
To see the opportunities another way, let’s assume we have a hypothetical process comprising six steps, each with the following reliability rating:
| Step |
Reliability |
| Step 1 |
99% |
| Step 2 |
95% |
| Step 3 |
99% |
| Step 4 |
98% |
| Step 5 |
90% |
| Step 6 |
95% |
This entire six-step process has a probability of succeeding only 78% of the time (0.99 x 0.95 x 0.99 x 0.98 x 0.90 x 0.95). Or, stated another way, the process fails (is defective) 22% of the time.
Using our first improvement method, we could make the system better by eliminating unneeded steps. If we determine that step #5 can be eliminated, the new process would look like this:
| Step |
Reliability |
| Step 1 |
99% |
| Step 2 |
95% |
| Step 3 |
99% |
| Step 4 |
98% |
| Step 6 |
95% |
Now the process has a probability of succeeding 87% of the time (0.99 x 0.95 x 0.99 x 0.98 x 0.95), much improved from the original 78%.
We could also improve the probability of success by improving the reliability of individual steps. So if we could improve the quality of steps #2 and #6, bringing both up to 98% reliability (from 95%), we would see this:
| Step |
Reliability |
| Step 1 |
99% |
| Step 2 |
98% |
| Step 3 |
99% |
| Step 4 |
98% |
| Step 5 |
90% |
| Step 6 |
98% |
The overall process would now perform at 83% reliability (0.99 x 0.98 x 0.99 x 0.98 x 0.90 x 0.98), again much improved from the original 78%.
Of course, if we can combine the two improvement methods — remove unnecessary steps AND enhance the reliability of individual steps — all the better. So eliminating step #5 while also enhancing steps #2 and #6 will result in:
| Step |
Reliability |
| Step 1 |
99% |
| Step 2 |
98% |
| Step 3 |
99% |
| Step 4 |
98% |
| Step 6 |
98% |
Now the system succeeds 92% of the time, with defects only 8% of the time. This is a quite a leap from our original performance of 78%.
Health quality champions are increasingly taking cues from counterparts in other industries who understand that reaching for perfection is not naïve but vital to success. For a company or institution to thrive in a complex, competitive, or high-risk environment — be that manufacturing of cars, buildings or computer chips, navigating a plane, raising a child or running a medical center — it must continually hone its performance. Innovators know that, while a 95% score may be outstanding for a student, that level of function in industry would be crippling. In fact, even 99.9% may be reckless. Here are some examples of what would happen “if 99.9% were good enough”:
- There would be a major plane crash every three days.
- 16,000 pieces of mail would be lost by the U.S. Postal Service every hour.
- 12 babies would be given to the wrong parents every day.
- 37,000 ATM errors would occur hourly.
- 20,000 incorrect prescriptions would be written annually.
- The IRS would lose two million documents this year.
- 107 erroneous medical procedures would be performed every day.
[Source: InSight, Syncrude Canada Ltd, Communications Division]
As we’ve just seen, the overall soundness of each of these systems, or those in any organization or family, depends on each of its parts. So if a system has 100 independently occurring elements, each of which functions properly only 99.9% of the time, the overall system will function correctly only 90% of the time (0.999 to the 100th power).
Given the implications of tolerating unreliable system performance, organizations embracing quality improvement should encourage their staff at every level to look at basic everyday functions in place and ask: “What steps could be removed?” and “How could each step be made more reliable?” These simple questions are paramount in the drive toward perfection.
