Does Your Communication Fail Safe?

We often need to communicate important information to other people in business. As we try to prevent problems and improve reliability of our processes, it's interesting to look at what happens if our communication fails. Does it fail in a safe way? Here's an example from the medical profession...

Patients at risk of blood clots often end up being prescribed Warfarin, an anti-coagulant, commonly but inaccurately called a "blood thinner". When a patient is taking Warfarin, they need regular blood tests (called INR tests) to make sure the dosage is right, with an INR value between 2 and 3. If the INR is too low, say 0.5, there's a high risk of dangerous blood clots. If the INR is too high, say 5, there is a high risk of dangerous bleeding. So, the patient needs feedback from their doctor based on the results of these regular tests. Usually this amounts to "stay on the same dose", "reduce the dose by X mg/day", or "increase the dose by Y mg/day". And, here's the process, at least in parts of Saskatchewan...

1. The doctor recommends a schedule for regular INR blood tests, say about once a week.
2. The patient goes to a lab when it's convenient, and gets an INR blood test done.
3. The lab sends the results to the doctor.
4. When the doctor receives an INR test result, they check the INR number. If it's too low or too high, the doctor phones the patient to adjust the dosage.

Simple enough, right? If there's a problem with the test result, the doctor makes an adjustment. When there's a problem, we deal with it. When there's not a problem, we do nothing.

One patient had a blood test done (and it turned out the INR value was 4.8 putting them at a high risk of dangerous bleeding). But, the lab either failed to send the results to the doctor, or the doctor's office lost or misfiled the test results. Whatever the reason, the doctor didn't get a test result, so the doctor didn't check the results, so the doctor didn't inform the patient.

Since the patient was only expecting a call if there was a problem, the patient thought everything was OK.

Since the doctor only responded to lab tests actually received, the doctor thought everything was OK.

Since the lab thought they'd sent the results to the doctor, the lab thought everything was OK.

As designed, this communication process did not fail safely. There was no guarantee that a problem result would get attention. There wasn't ever a confirming message sent to say "the test result were received and are OK."

To fail safely, one possible strengthening of the process would be to ALWAYS notify the patient with results within 1 day of the test, (dosage OK, increase dosage by X, or decrease dosage by Y). So, if the patient had not heard back within one day, they would not assume that the dosage is OK, they would assume that something went wrong with the communication process. Many other remedies could also help this process,

The point is, we can apply this thinking to many types of reporting and communication. If we only get a "signal" when there is a problem, we never know for sure if everything is actually OK. Someone might just have failed to communicate the signal. And that's not safe.