da http://gomerville.com/2012/04/29/the-ethics-of-data-mining-in-healthcare/ un blog gestate da “a paramedic who works for an organ procurement organization in the wilds of Kentucky”
My last two posts were about data mining and the evils of it. I have also eluded to the fact that much of what I write in the future is going to come from the point of view of someone who wants to opt out of this system and maintain my privacy. I think that my privacy is a right that is being eroded by the very thing that I love, which is my electronic connection to other people. The companies and individuals who provide me with this connection also wish to make money by tracking my behavior. Many of the free services that are on the internet are not truly free. These services are provided at the cost of my privacy. For instance I could use Google Docs to write papers, take notes, and store my thoughts. Google will never charge me a dime for this service. But other companies will pay Google to make me part of a data set and demographic that they can advertise to based on the private data I have input into Google Docs. As more software and services are placed in the cloud, the temptation for companies to produce income streams from data mining will become overwhelming in this competitive market.
But technology can also be made to work for me. Technology should allow me to connect to anyone I wish and maintain the level of privacy that I desire. Since most of the people who would wish to access my information are people I actively wish not to communicate with, I think that my electronic privacy is also a right that should be protected by my government.
In my research, I have come across a few examples of people working as hard as they can to mine data from social networks in order to cure disease and save humanity. This made me pause for a moment. Here are noble people invading my privacy for a very good reason. Often times the United States government will attempt the same thing. For instance my government may wish to mine my data and invade my privacy for the purpose of thwarting terrorism. However, these measures have been largely unsuccessful and the mechanisms have often been abused for less noble purposes. But tracking epidemics and eradicating disease…this is a whole new ball game. In order to understand what’s at stake here, the reader needs to know who the big players are. If you are willing to watch a few TED Talks you can be brought up to speed on just what exactly is emerging in this field.
First up is Nicholas Christakis. He is a researcher at Harvard who maps social networks in order to detect epidemics and emerging threats. He has given two TED Talks on this subject. They are both worth your timThe work that he does, and the information he obtains from network analysis could very well create a system that could prevent the deaths of hundreds of millions of people by catching an outbreak of the flu in its early stages before it becomes a pandemic. Notice that he eludes to the fact that this technique can be used for other purposes. In both videos he mentions advertising and product trends. So it is obvious that Dr. Christakis is considering alternative income streams right out of the gate.
Larry Brilliant is another physician who is using these techniques to not only track disease outbreaks, but eradicate diseases. He was actually the winner of the TED prize in 2006, and these funds were used to help realize his vision.
Not long after this he was named Executive Director of Google.org which is the philanthropic arm of Google. So Dr. Brilliant sat atop the biggest cache of information the world has ever known and was in the perfect position to use it to detect and avert pandemics. However, as this NY Times Article suggests, Google.org failed to live up to its expectations and Dr. Brilliant was let go in 2009. Google.org still operates and there is a flu tracking tool, but it is apparently just a shadow of its previously hyped self.
The last TED Talk I have for you presents an interesting alternative to data mining. Jamie Heywood started a website called Patients Like Me meant to track the trends of patients diagnosed with various conditions.
He got the idea after his brother died of ALS. There wasn’t any way to connect with other people with ALS to find out if various treatments were prolonging life or improving the quality of life for these patients. The premise is interesting: patients sign up for this service, input relevant data and allow themselves to be tracked This information is made available with powerful statistical analysis tools that haven’t been used in the healthcare setting before. The data sets are so rich and powerful that it is meant to provide a way for someone to make a study up on the fly. If you want to sit down and find out was has happened to thousands of patients who are taking two specific drugs at once you can do this. You can proceed to drill down into the data set to see outcomes, chart trends, see who died, and what the symptoms were like. There is no need to spend the money to create a study. The data is already there.
However, this service was made by patients for patients. I have said time and time again that health care often suffers when the patient or the family (usually the least qualified people in the room) calls the shots. One could make an argument against how samples and participants are chosen and how this information is acquired. However, even if the source of this information (namely the patients themselves) is in question, it would be damned hard for a medical professional to deny that there is a lot of rich data about outcomes sitting there just waiting to be accessed. And isn’t that what healthcare is all about, patient outcomes? Ignoring this data set is like a doctor going into a room and ignoring all the complaints of the patient they have come to treat. Why else would this doctor have showed up in the first place?
I like the concept of this service because no one is collecting the information of patients without their consent. Indeed, patients input information in the hopes that other people will access it in order to learn more about their condition. In this way, you have to actively opt in to have your data mined. If you wish to opt out, simply don’t sign up for the service.
So, the examples given above bring up some sticky issues about privacy. After examining a lot of evidence I truly believe that most invasions of electronic privacy by the government and the private sector are unwarranted. Many would disagree. This is the age old argument where privacy and the rights of individuals are weighed against the safety and the convenience of the society in which those individuals live. But what if a health organization could track Facebook and Twitter comments to find out that a new strain of avian flu had broken out? What if the data of those social media sites could be mined using techniques developed by researchers such as Nicholas Christakis to not only detect an outbreak, but detect it early? What is the individual privacy/social benefit analysis of that?
If the data sets of services like Patients Like Me can be called into questions because of their source, will we ever be able to trust the comments of individual citizens using social media to pull the trigger on a multi-million dollar pandemic response? It seems to me that a system like that is one false alarm away from ridicule.
Wikipedia is a great social experiment of what can happen when the average Joe decides to collaborate on something. Yet I don’t know one teacher who would accept a Wikipedia article as a source for a paper. Should physicians dismiss reams of patient data for the same reason? Do these flaws make the invasion of privacy not worth the effort? In a world where HIPAA guarantees the privacy of every patient of the United States by federal law, can we in good conscience mine data from social media sites to obtain information to treat those same patients?