Loose Strife

We got a call today. The person was upset to here that Laura was not fairing well and wanted to talk about whether everything that could be done to cure her had been tried or considered. I want to say definitively yes. But the truth about dying from cancer is more nuanced. Certainly we want to be absolutely certain that Laura is suffering from cancer progression and not from some other cause — another undiagnosed condition, a complication of treatment, or whatever. To that point there is no doubt in my mind. But has she done everything possible to reverse her disease or to stave off its progression? No, not exactly.
I know I’ve trod through this territory several times already, but it’s one of those nagging, frightening thoughts that there might be something out there — some rock unturned — that would break this case wide open. Nobody wants to be guilty of death. And everyone would love for some sudden fix to a seemingly intractable problem. Magic pill or magic herb or magic touch or whatever it is, we can’t stop searching. We hear about people who are “fighters” and “beat” cancer. The myth is pervasive and understandable because everyone wants to feel empowered. We read about Lance Armstrong who didn’t give up, but most people don’t realize that his non-metastatic testicular cancer is one of the most successfully treated types of cancers. The truth is cancer has no sympathy and can be relentless. But even in the face of foreboding odds, we want to do whatever we can. Right?
What makes this complicated is the reality that treatments almost all have benefits and disadvantages — some well known and others very uncertain. In Computer Science there is a subfield called “Decision-making under uncertainty” — which is pretty much just what it sounds like. Researchers try to come up with the most effective technique for making a series of decisions towards achieving some goal (e.g. trying to guide a robot across terrain or optimally investing in a hedge fund) within a context that is not fully known (e.g. the robot can’t see what other obstacles might be on the other side of the hill and the automatic investor doesn’t know whether the hurricane will strike a major population center… and how that uncertain weather might have a cascading effect in numerous uncertain ways on the financial markets).
We can frame cancer treatment similarly: the goal perhaps is to maximize survival within the complex context of thousands of unknowns. Many of these unknowns are literally unknowable at this time in human history. Things like “does Laura have micro-tumors in her brain?” or “does her individual variety of breast cancer respond to a particular chemotherapy drug?” (Why are they unknowable? Because we can’t see such micro-tumors with current, non-invasive imaging technology. And nobody knows enough about the biology of breast cancer, yet, to know precisely how Laura’s particular cancer cells might react in the presence of some small pharmaceutical compound — even if the DNA from some of Laura’s cancer cells were sequenced. Oh, and by the way, cancer DNA can vary dramatically even within one person from cell to cell!) Even so, it is theoretically possible to put numbers on most everything — say a probability on the presence of micro-tumors and another number describing our confidence in the probability! Computer Scientists do this sort of thing all the time.
But that doesn’t fully solve the problem because our goal is wrong. The goal is not just maximizing survival, it is staying alive while maintaining a good quality of life. If you’re not alive, then your quality of life is not so good. So really we just want to maximize quality of life. OK. We could still create some kind of new goal that was some numeric measure of quality of life that cleverly combined different indicators of the good life. Biostatisticians do this sort of thing all the time! And what about all that uncertainty? We could try to reduce it down as much as possible by subjecting the patient to countless tests over time to look for indicators and track changes. We read about how Ted Kennedy invited a multi-disciplinary team from around the country to recommend appropriate treatments. Think how much better our computer program would be if I had that same access to expertise?
And so continues the thinking from the ivory tower. Of course this is all doomed from the start because what’s fundamentally missing from such an analytic exercise is the recognition of individual, distinctive, emotional human nature. Even if there were a computer program that could take in all the facts and spit out recommended actions, even if we could formulate the major interactions, parameterize the system, and estimate to the best of our ability all the uncertainties therein, it would still be unsatisfactory. To one person, a predicted quality of life score of 5.6 (I’m making this up!) for another 6 weeks is just the long shot she was hoping for. To another, it’s a miserable death march. (OK, OK, I know someone out there is thinking about how a personality test could determine a patient’s attitude about quality of life in order to properly calibrate that score. This is the mind of the modeler!) To one person, a commitment to not stop fighting is what keeps her motivated to get out of bed. To another, insufferable fighting is what drives her to bed.
This is not to say that reason should play no role in end of life decision making. Most of us probably want to be sure that all treatment options are known and considered, but there is no precise threshold for acceptance or rejection. Such decisions are intensely personal, idiosyncratic, and impossible to fully quantify.
Now I’ve been very abstract up until now and will try to bring it back down to reality. Laura could try another chemotherapy. There are actually numerous very long shot options. Laura could try acupuncture to relieve her symptoms. She could binge on anti-oxidants. She could flush out her body in the desert. She could pray for a miracle. She could try another round of radiation. She could attend a healing service. She could fly to the best cancer center in the country (it’s in Texas) for a second or third opinion. She could listen to recordings of subliminal healing meditations. She could get her shunt checked again. She could get another set of PET/CT scans. She could say, “enough”. She hasn’t. But she could. And that’s OK.