In a few days, I will be releasing the most controversial healthcare project I have ever worked on. But you do not need to take my word for it. I will be releasing a completely new healthcare data set. That data set, which will remain a “Mystery Data Set” until its release to the healthcare data scientists attending Strata RX, should completely revolutionize the way we think about healthcare delivery in the United States.
This mystery data set is the first real outcome of the Patient Skunkworks project. Patient Skunkworks is a new way for me to try and create high-impact but low-profit software projects. This is part of a new Not Only For Profit software development model that I have been working on. The new company forming to do this work will be called Not Only Development.
I will be releasing this data during the last keynote on the first morning (Oct 16) of the 2012 Strata RX conference. There is simply no way, in a single keynote, to even begin understanding all of the ways that this data set will be leveraged to improve healthcare. More importantly, there is really no way to adequately explain why I would choose to give away such a valuable and dangerous data set.
To help people digest the implications of this data set, I will be writing two articles about the data set. This one, before the release which helps to explain the underlying motivation behind the release, and another one after the release explaining what the data set is, and how I think it can be leveraged.
I am releasing this dataset because I believe that the only way to solve the problems in healthcare is to embrace a radical openness with health data. Healthcare data, with the exception of patient identity data, belongs in the open, in the sunlight. When used correctly, I believe that healthcare data should make patients feel empowered, and everyone else in the healthcare industry uncomfortable. I believe that patients deserve deep, dangerous and real access to data. I think when we start talking about how data might actually be dangerous for patients, its just a sign that we are “doing it right”. I call this concept Radical Access to Data (and yes, that recursively spells “RAD”).
Why am I doing this? I am a Hacktivist. Not a Hacktivist like those anonymous folks. I aspire to use technology to change social issues in the way that Martin Luther King used preaching to bring about social change. The anonymous folks aspire, from what I can tell, to be something like digital graffiti artists. Graffiti can have a certain nobility and it sometimes has a place, but this is not my aim. “Hacker” is a term that is misused to mean breaking into other peoples computers. Originally the term hacker means a person who creates a clever solution to a previously unsolved problem. So when I say ‘Hacktivist’ I means that I am building things that are intended to change the world. Hacktivist is a term I use without any sense of exclusion: I think that anyone working in healthcare data, as long as you are considering the ethical implications of your job, probably deserves this same title. I think there are countless other data and computer scientists that probably deserve to be called Hacktivists. I am hardly the only or best practitioner of Hacktvism, but I certainly enjoy working on tools that really make a difference in healthcare.
Ultimately, I do not get paid to be a “Hacktivist”. I get paid to build healthcare tools. I have started acting proactively, as a Hacktivist, because I feel that it is the most fun way to discharge the ethical responsibility that I have as a tool-maker. Every tool-maker in healthcare shares these ethical obligations. Please permit me to discuss, carefully, both the wrong path and the right path for a healthcare toolmaker to take.
Permit me an exception to Godwins law. For those of you who are familiar with this law, it says that any Internet discussion in which the Nazi’s are mentioned has become pointless and unproductive. But, as you can see from this link, there is an exception to this rule for discussions regarding things like racism and patient consent, where the ethics of the Nazi’s specifically broke down. Under this exception I should be able to discuss the Nazi’s in the context of medical ethics without my all of my “true geek” readers abandoning me.
Most discussions of the ethics in healthcare begin when a greek philosopher proposed that doctors had a primary responsibility not to hurt their patients but only a secondary responsibility to acquire knowledge about medicine. But this was before the advent of the modern scientific process and culture. After a few centuries it became clear that there was a problem with the cultural values of medical science.
After the conclusion of WWII, the Allies were horrified to discover that in addition to being the target of systematic genocide, Jews, especially identical twins, were subject to horrific experiments at the hands of Nazi doctors. No one has ever criticized the doctors who conducted these experiments for sloppy science. Quite the opposite, history remains in awe of how precise these doctors were. Without sacrificing precision these doctors had placed the goals of science ahead of the welfare of their patients. The Nazi doctors took great liberties with basic civil rights in order to get the data that they wanted. Later, when they were on trial for their actions, the doctors who conducted these brutal experiments argued that there were no specific international agreements for the conduct of medical experiments to which they could be held accountable. It was obvious that the mandate from ancient Greece was not enough. Society needed a well defined balance between the civil rights of patients, and the desire to advance science.
In response the international community created a series of standards for ethics in medical experiments culminating with the Declaration of Helsinki in 1964. Two years later in 1966 a government researcher, Peter Buxtun, would file a formal complaint with his superiors regarding a syphilis experiment being conducted in Alabama. After being overruled, twice, Peter went to a reporter and in 1972 exposed these experiments to the public.
The clinicians in Alabama and Germany who had acted to harm patients in the name of science all argued that they had done so “under orders”. In both cases the clinicians claimed that they were operating within the ethical bounds of science to the benefit of humanity.
In 1961 a Yale psychologist showed experimentally that typical people were willing to participate in deeply unethical and violent ways under the auspices of scientific endeavors. It was apparent that scientists, in modern society, held a role of authority. A role previously held primarily by religious and political leaders. When medical doctors and other scientists spoke they did so with authority. The respect afforded scientist and doctors is one of the few places where the class system that has otherwise fallen out of favor continues to hold sway.
In 1975, on the heels of the syphilis experiments and with the results of the Yale studies in mind, the Declaration of Helsinki was revised for the first time. This revision nearly doubled in length of the document and introduced the concept of formal peer review for experiments, a system that we now call “IRB” for “Institutional Review Board”.
The international community agreed that there needed to be a mechanism for scientific oversight, so that medical experiments on people would be conducted with oversight.
We can summarize this history of the ethics of medical experiments:
- Greek ethic: Doctors do science to patients, but must be nice.
- Nazi ethic: Doctors do science to patients.
- Current ethic: Doctors do science to patients, but need approval from a committee
But there is a new ethos in medical science. It does not matter whether you label this ethos as epatient, quantified self, personal genomics or n=1. The basic ethos is the same:
- New Ethos: Doctors and patients do science as equals.
All of the previous installments of the ethical systems are essentially elitist. A group of “those who study” get together and decide whether there is a value proposition in scientific research. There is a requirement to have a “non-scientific” person participate in an IRB, but there is no requirement for consultation with a relevant patient or caregiver. The IRB process is effective at preventing abuses, but it serves to perpetuate the class relationship between doctors and patients.
This elitist attitude presumes the doctor as an authority figure. It presumes that the patient cannot understand the context of a scientific experiment. It presumes that “informed consent” is unidirectional. It presumes that patients cannot “handle” data about their own healthcare. It presumes that patients cannot participate in the process of determining if a given experiment poses risks that outweigh its benefits to those patients. It presumes that it is the privilege of doctors to study patients rather than assuming that patients should be able to study doctors.
In all of these ways, it presumes that the doctor needs to take a parental role with patients.
There is no justification for these assumptions. There is considerable evidence that patients who refuse to be treated like children, patients who are fully engaged in their healthcare, receive better care. But even more troubling, there is evidence that doctors, despite their best intentions to the contrary, wish that patients would just “do what they say“.
Sadly, this paternalism is frequently built into the software tools that we use to run healthcare. There are notable and welcome exceptions, but most of software that holds healthcare data in this country is designed to provide insights to “experts”, without regard for how this impacts patients. Patient access is usually read-only if it exists at all, and frequently patients are shut out of their healthcare data process to a deeply troubling degree.
There is a movement to encourage patients to take an engaged role, and to rid medicine of its parental assumptions. This movement, the “engaged patient” (or epatient) movement centers around a non-profit called the Society for Participatory Medicine, and the whitepaper that first articulated the term “epatient”. This movement has attracted a small but growing number of health information professionals who are devoted to building tools that truly empower patients. While the standard case is dismal, it is important to highlight those projects and ideas that buck the trend.
Robert Wood Johnson Foundation has sponsored an effort called Open Notes, where patients get real time access to the notes that their doctors make about them. This project, which doctors are frequently uncomfortable with, has recently released very impressive results. I believe that this project is the most important work that RWJF will sponsor in this decade.
The Medical Commons provides an IRB approved mechanism for patients to donate massive amounts of their own deidentified data to the scientific community.
The Direct Project is one of the protocols endorsed by the ONC for the purposes of doctor to doctor and doctor to patient communication. Direct is just a very specific configuration of SMTP and S/MIME, that enable a seperate and secure clinical messaging network for patients and doctors. This project is supporting a use case that we in the patient community call “cc the patient”: Whenever the data in a patient’s electronic health record changes, the patient will be able to get a new copy sent to them digitally. Using Direct patients will eventually be able to subscribe to their own medical record the same way they subscribe to Groupon now. More importantly patients could also allow for applications, like HealthVault or perhaps something more subversive like Qpid.me, to be the recipients of that message.
The Blue Button movement has always been fundamentally about getting patients access to digital data. But initially, Blue Button was a pretty crappy data standard. Now Blue Button and its successor the Automatic Blue Button efforts have moved to from being a crappy data format with pared with a notion of radical patient access to data, to the use of good data formats paired with the notion of radical patient access.
Some personal health records are little more than windows into EHR systems. But some PHR applications are taking an aggressive stance that the patient should be able to use the PHR to actually control their own healthcare data. The most radical of these, by far, are my friends at the Indivo project, but the guys over at Microsoft HealthVault are also pursuing an approach that counts as radically empowering. VA does a fairly good job empowering veterans with MyHealtheVet and Kaiser Permenente has done impressive things with KP.org. There are other good PHR companies and products out there, but these should be noted as taking an approach that has served to empower their users.
The Obama administration has committed to radical openness in government. Todd Park and Aman Bahndari as well as the team from HHS that will be presenting at Strata RX have delivered, in spades, on that promise regarding government healthcare data. Healthdata.gov is an amazing resource that contains substantial and deep resources that have only begun to be fully leveraged in a way that will impact patient care.
For each of these efforts, there is a sliver of this core idea: That by giving patients radically deep access to their own health data, we can change the power dynamic between providers and patients, and the result will be better healthcare.
In the past, medicine was something that was done to patients, in the future medicine will be done with patients.
In the past, doctors wanted patients to be compliant, in the future doctors will want patients to be engaged.
In the past, informed consent meant that an experiment could be done to a patient, in the future patients will conduct scientifically valid experiments on themselves and present the results to their doctors.
In the past, hospitals would horde patient data, in the future patients will choose to donate their medical records and genomes to create the most massive healthcare data that that the world has ever known.
Or at least, this is a future that we can build together. What right thinking person believes that the Big Data efforts of banks and Wall Street are typically used for the good of the average person? They are at best, merely neutral and at worst abusive. Big Data is power, and as Jesus and Stan Lee teach us, with great power, comes great responsibility. This power, for the time being, is in the hands of people like me: the toolmakers.
Big Data in healthcare will make someone more powerful. The data scientists who will be attending Strata RX will make a thousand tiny decisions that will determine who gets that new power. They can choose to make the insurance companies, or the government or the doctors or the hospitals or the drug companies more powerful. Or we can choose to use Big Data to empower patients, to put a permanent end to paternalism in medicine. There is a brief moment here, when it is the ethics of the Big Data toolmakers that will determine how society will step forward in the delivery healthcare. Our decisions now will echo for decades at least.
Alistair is wrong.
Big Data is not a civil rights issue.
All civil rights issues are Big Data issues.
For every imbalance in human society, there is an opportunity for data analysis to become a control mechanism. For many ethical issues, like the fairness of health insurance rates, Big Data was an ethical issue even before we started calling it “Big Data”.
Using Big Data, we have the opportunity to expose, classify and correct civil rights issues inside and outside healthcare. Or we can allow the insights that we gain from Big Data in healthcare to continue to exist as mere profit engines. Healthcare is merely the first and most notable fashion that we can turn the power of Big Data directly on ourselves. Modern society has made a moral choice by rejecting the implicit argument, made by the Nazi doctors, that we can pierce the skin for the sake of scientific discovery. Now we have a new tool, that will finally show us things that have been hidden since that decision. How will we use this new virtual scalpel?
We can allow Big Data to be tool that is used against patients or a tool that is used to help patients. Over the next few years, we will see efforts to do both.
Patients have the right to do science to the same degree that any clinician does. The “best” data sets in healthcare are those data sets that are used to empower patients and doctors to work together to improve healthcare. The “best” software in healthcare empowers patients and doctors at the same time.
I am about the release a data set, that can be used for the good and the detriment of patients. If you are a data scientist, or data journalist and you are considering using the data set, please consider honoring these requests:
- Try to see how this can be used to help patients
- Try to see how damage to patients can be reduced
- Try to discuss, openly, the ethical problems in this and other big data sets in healthcare
- Don’t make assertions from this data set that you cannot back up. Don’t make me call Susannah Fox, because I will…