Dear Doctor K

By / in Healthcare Quality /

When writing training materials for a general audience it is not possible to make examples that are tailored to a specific audience. One of my friends, a physician, commented that he thought the design of experiments exercise was too abstract. I wrote the following in response:

The example is for training a general audience, so it’s necessarily an oversimplification. As far as applications in healthcare, consider the randomized clinical trial. These are simple Designed Experiments. Usually they are one-factor-at-a-time (OFAT) experiments that attempt to isolate the effect of one drug across a population. Screening experiments would allow the assessment of multiple factors simultaneously with relatively small sample sizes. Since the OFAT approach is used instead, sample sizes are normally very large in RCTs. RCT designs predate the development of designed experiment theory and healthcare hasn’t adopted this more powerful approach.

Another application of DoEs could be bacteremia  infection reduction. Different protocols for handling CVCs could be tried. For example, different sterilization chemicals, different methods of applying the sterilization, different groups inserting the catheter (e.g., physicians vs. nurses,) different wait times from sterilization to insertion, etc.

Still another application might be readmission reduction. The experiment might investigate different discharge instructions, different follow-up protocols, different types of patients, etc.

Emergency Departments present numerous opportunities for experiments. Different staff schedules or staff mixes, different paperwork processing procedures, different ED layouts, etc.

In short, the screening DoE is useful whenever there are many factors we want to investigate to get an idea of which factors require additional attention. They help us improve the process right away by showing us simple, linear relationships. And they point us toward optimization by highlighting the critical to quality characteristics that require further study. There are many such situations in healthcare.

Designed experiments are controlled prospective studies. They are the gold standard for learning about process control and improvement.

No comments yet! You be the first to comment.

Leave a Reply

Your email address will not be published. Required fields are marked *