Chapter 1. Beginning at the Beginning

Several years ago, my cardiologist told me I needed an implanted cardioverter defibrillator (ICD) installed in my chest. I have a condition known as ventricular tachycardia, which means my heart occasionally beats too fast. This can lead to fibrillation, which is usually fatal; the deck-of-cards- sized ICD monitors heart rate and triggers an electric shock if it detects an aberrant rhythm that persists too long. The shock is about the same as you'd get from those well-known electric paddles in the emergency room.

The installation of an ICD is not something you take lightly. So, as a veteran project manager, I quickly did a risk analysis. The question was binary: install, or not install? If not, my best hope was that there would never be a fibrillation, in which case not installing would be the right solution. But if a fibrillation occurred, not having the ICD would most likely lead to a wrong and fatal "solution." This branch of the decision tree favored going ahead with the surgery.

But there was more to consider. While the risk in the surgical procedure itself was fairly small, there was still the possibility that the device could failbut that turns out to be no worse than not having the device at all! The more interesting question was, what happens if I got a classic "false-positive?" That is, the device fires when there is no imminent fibrillation.

One thing was certain: When the device fires, the recipient experiences what my doctor described as "inconvenience." Certainly, dying without a necessary shock would be more inconvenient, but you can imagine scenarios where an unnecessary firing of the device could put you at risk. I needed to think about this branch of the decision tree.

What could go wrong? The device is pretty simple. It has a long-life battery, a big capacitor, and a software-driven microprocessor. The battery and capacitor are stock hardware items that can be subjected to the usual quality control metrics; the same can be said for the microprocessor. But what about the software?

Most of my career has been in software development project management. Knowing perhaps too much about how software is developed, I was concerned about trusting my life 24 x 7 to the performance of a piece of software. For the rest of my existence, electrodes implanted in my heart muscle could deliver a significant shock based on the commands processed by the software monitoring its cadence.

I went ahead with the procedure and, so far, the device and I have successfully coexisted for seven years. The data we can read from the device's memory shows that it has gotten ready to fire several times, but only once has the software logic made the decision to pull the trigger. Yes, it was "inconvenient," but not overly so. And I'm glad it worked when I needed it.