What is the probability of a patient visiting the emergency department not encountering an error of process?
Registration functions perfectly.
Labs are ordered correctly and return on time.
Vital signs, medications, transports, x-rays, exams…all performed without error.
Each step of each of these processes has its possibility of error. Every handoff of a task to another person has it's own risk of mistake.
Take the example of ordering and getting the results of an ankle xray.
Assume that getting an ankle xray takes 20 steps (it likely takes many more).
See the patient…
write the order for the xray on the correct ankle…
put the order in the clerks box…
have the clerk enter it correctly…
etc…
Let's say the accuracy of each of these steps is 98%.
In addition, the process of getting an ankle xray involves 5 different handoffs of information. The clerk, the radiology tech, the nurse, the radiologist, and the emergency physician are all exchanging information.
Let's assume that each of these handoffs has an accuracy rate of 95%.
Rolled throughput yield measures the predicted probability that an ankle xray will be ordered and completed error free.
Calculating the rolled throughput yield shows us that the probability of an error free ankle xray process is:
(0.98)20 x (0.95)5 = only 52%
What happens if we reduce steps in the process?
We decrease the process from 20 steps to 8 steps.
Instead of 5 handoffs of information we have 3.
Now the probability of an error free ankle xray process is:
(0.98)8 x (0.95)3 = 73%
Every step and every handoff of information decreases the probability of an error free process.
An error the majority of the time does not result in a bad outcome for the patient. (eg "Can you re-fax that order?") However, it increases the time it takes to get through the system. It increases the frustration experienced by all involved.
Simplify the process.
