Some say that "harnessing the wealth of information in EHRs and other troves of real-world data" will help us "get new cures."
Others may build on this claim adding this supposedly overrules all the concerns about inefficiency and documented harm to patients by the data collection process itself, especially in the emergency room and in critical situations like anesthesia.
Assuming there were none of those risks, what of the ultimate claim itself, that having the data somehow inherently means having a cure?
That is a hypothesis. Where is the testing of this hypothesis? Where has this been observed? Where is the data itself on this supposed value of massive data collection? How is this any more scientifically sound than saying, "Now that we can see cancer, we know how to cure cancer"?
With no basis for this claim, it is only a hypothesis, and a cynical one at that which plays on people's hopes, fears, and vulnerabilities. After all, what sick or dying patient or family member will turn away any sign of hope for extending the life of someone they love?
It's helpful to ask if this makes sense, even if it sounds logical. It does sound logical, too: If the disease is in the body, and if we had data on the disease, then we should be able to figure out what's going on and fix it, right?
About 100 years ago, not long after radioactivity was discovered, there were some in the medical community who thought there was radioactivity in the human brain. Absurd, yes? How will it sound in 50 years that some people claimed just because mankind could create certain data structures for holding information about the condition of the human body and genome, and fill those data points, that we would thus be able to cure more diseases?
There are some who say that big data algorithms can "see things in the data that people cannot see." A couple analogies illuminate this claim. Have you seen how algorithms do automated trading? They look like a machine with very regimented and regular trades. Another example would be the scene in The Hunt for Red October where he speeds up the sound of what they thought were whales and when he plays the rapid and continuous pounding sound it was obvious that they had heard a man-made machine instead. The human body, and diseases for that matter, are living and are not man-made machines. Man-made algorithms will see algorithm-like things in the data.
To say nothing of the medical and political risks involved with this massive enterprise on many levels, the results may be that we have nothing more than more data than ever on human beings, the condition of their bodies, their behaviors, their ailments, and still be little closer to these elusive new cures.
We have more data on the weather now than ever before. Does this mean we can solve hurricanes and tornadoes? Short of that, are we any better at forecasting the weather more than a few days out? Or have we simply learned that the weather and all the factors that play into it are more complicated than we had ever imagined?
Here's an analogy for lawyers. Imagine someone created a system for electronic legal records (ELRs) to collect all data about all legal cases ever into a massive interoperable database on legal cases. Attorneys would only be paid by the government, and they would be incentivized for reporting data correctly and getting clients to correct the government's records, and penalized in payments for not complying. All this is done so that the data could be run through algorithms that can see things in the data that people can't see. Would that mean the end of litigation? Or would this be a formula for opening the floodgates of litigation?
Without any scientific basis for the belief that data and algorithms simply lead to cures, we may be on the most expensive path ever to find out something that could already be plainly evident.
Knowledge has greatly increased through science with the testing of hypotheses and comparing them to collected data. There is an established pattern and precedent for these things and the value they produce. Data, in the right context, with the right constraints on how it is gathered, with an appropriate amount of space to let the process work, indeed has its uses.
For those who would claim there is only value in massive comprehensive data sets, and claiming it is essential to cast aside privacy concerns, it's appropriate to check for conflicts of interest. Most everything of value starts small and grows. If someone claims there is only value in bigness and not growth, that is highly suspect.
“Physicians practice medicine. Physicians do not practice data.” - Tell CMS today. Stop #MACRA before it stops you. https://t.co/9tm4y6ySz8— Texas Medical Assoc. (@texmed) June 24, 2016