Back Again

Hello, everyone. I have been away from this website for over a year due to some tragic circumstances. During the past year, my beloved wife contracted a rare form of an affliction that resulted in paralysis for several months and her eventual death. My mother, who suffered from dementia, also passed away. Finally, the family dog Molly, who reached the ripe old age of seventeen, also passed away from internal organ failure. 

I don’t bring this up to ask for pity, as I am acutely aware there are clearly those who had much more traumatic experiences than I. My position, for better or worse, is contained in the book of Job: The Lord gave, and the Lord has taken away. Blessed be the name of the Lord (Job 1:21).

 I do think that there was clearly a learning moment here for nonprofit organizations. My wife, Susan, was originally diagnosed with an aggressive form of melanoma, resulting in two operations that failed to curtail the spread of the disease. As a result, Susan underwent immunotherapy, which is today thought to be the best approach to handling the melanoma.  Chemotherapy would be the second line of defense. Several months into the treatment, Susan began to feel ill. Within five days, she could no longer get out of bed and was functionally paralyzed. The doctors were perplexed, having no explanation or treatment for what was happening. 

They began by treating the most likely causes of the paralysis, taking the most appropriate course of action with the highest potential for success. This approach failed. The doctors then began holding daily conference calls to brainstorm the situation. This was also an appropriate course of action, but produced no better treatment option. Finally, a friend who was also an oncologist thought outside of the box and found that similar symptoms were reported from treating various forms of cancer associated with female organs with immunotherapy. There were only two other reported cases where the immunotherapy resulted in this paralysis for patients being treated for melanoma. The upshot was that it was considered such a remote possibility that it was initially set aside. After all other possible diagnoses were ruled out, the correct diagnosis was made. The delay of about two weeks had devastating consequences. The proper treatment was finally given, but it was too late.  

The question that arises is whether the diagnosis and decision process was flawed or appropriate. From a decision analysis perspective, probably not. The most common possibilities were treated first. As the situation worsened, doctors began consulting with one another to pool knowledge. Where did the process break down? Perhaps the doctors did not think outside the box and did not quickly consider other options with very low probabilities. Most strikingly, there was no discussion of using AI to attempt to diagnose the malady. Perhaps a proper questioning architecture could have cut days off the diagnosis process and possibly saved my wife’s life. It is simply not possible to know now. The point I am making is that decision protocols should now include consultation with AI when an expert system or a proprietary large language model is available. Bayesian analysis might have helped as well. 

Note: The doctors were professional and worked long and hard to save my wife. I have no complaints. You will also forgive me if the medical details I have outlined here are not entirely correct. I am a simple CPA who studies decision methodologies and misses Susan dearly.