Thursday, October 1, 2009

Social Work in Silico

A study released in the BMJ describes a technique for using case histories and medical records to identify victims of domestic abuse. By finding a characteristic pattern that precedes reports of abuse, this method can determine whether a patient is the victim of domestic abuse, on average, 10-30 months before they would otherwise be detected. Most importantly, the method was determined by a computer analysis. With the increasing reliance on digital medical records, it will become standard practice to have records routinely scanned for a variety of ailments, and human doctors will increasingly rely on the recommendations of advanced statistical analyses.

The day of the doctor as the emissary of diagnostic technology may seem alarming, but this is certainly not a recent development. As medicine has become increasingly science based in the last two hundred years, it has A: become effective, B: changed the role of the physician. We can luckily exclude from our list of perfectly random chances the likleihood of surviving routine surgery (thank Lister!). We now find the concept of having lost even one parent while still a child uniquely terrible; as opposed to commonplace (prior to scientific medicine, by the age of ten a child had a 50% chance of this being the case). No longer used is the blessing, "may you live to see your children's children," in fact, we have come to consider the 2.5x increase in life expectancy so normal that we are more likely to interpret this saying as a threat.

A physician relating to a patient in a world before the great advances we now take for granted would have a very different job indeed. As life has been extended, we have come to take the luxuries of scientific medicine as a right; most people today would be rather upset if given the news that they would only live to be 40 years old (a 33% improvement on the pre-1872 worldwide average). The expectation that medicine can save us from most ailments, and that a doctor can be expected to actually effect that cure, has forced the physician to become more diplomatic, with every sense that the word engenders; sensitive, well-informed, compassionate, and self-protecting.

Imagine now the position of a doctor given a positive result after running the above mentioned software (or perhaps the software is programmed into whatever medical database system is employed, so that this is automatically flagged). Of course there is the fear of the patient reacting negatively, but this is only to be expected in 16% of the cases where a physician offers the opportunity to speak to a counselor. When considering that some of the largest diagnostic indicators of domestic abuse are alcoholism, anxiety, and perhaps most alarmingly, poisoning, the physician clearly must act.

What will this physician say? Will they admit that they themselves did not notice the pattern of abuse, that in fact the patient's advocate is actually a non-conscious piece of software? Will they approach the patient by saying, "Not once in your visits to the emergency room did anyone have enough suspicion that you were being abused to say something. In fact, if it weren't for the patterns of electron movement set by researchers you've never met, representing statistical methods developed by a Presbyterian minister who lived three hundred years ago, coursing through hardware made cheap enough that we can actually afford to have it for uses such as this due to a 50 year old error in the making of a crystal that yielded the semiconductor, which has... you'd probably have another one to three years of abuse before we caught it."

Of course not the latter, unlikely the former, which is not a censure of the physician. This exercise simply forces us to recognize the physician as the emissary of a long list of developments, built upon each other and brought together in ways unexpected and in many ways unintended. The ability to diagnose with statistics and mathematics is already commonplace; body-mass indices and indices of abdominal adiposity are so routinely used we begin to forget what they actually are: substitutes for human diagnoses. We are so reliant on them that physicians use them to convince patients of their need for change, "you need to lose weight," has been substituted with, "your BMI is here; it should be here."

What is not commonplace is the ability to diagnose not only complicated illnesses such as being the victim of abuse, but the ability to diagnose someone who need not ever visit the doctor to be caught, in this case the actual abuser. This is a profound level of technological subtlety, with huge implications.

Welcome to the age of the Silicon Savior.

1 comment:

  1. !!!

    Seriously... and think about it on even smaller levels, like being able to diagnose a child with not getting enough attention, or someone with not reading enough...