What open source can teach medical practice

What open source can teach medical practice

Summary: We are collecting the data. We need to unlock it. We are writing the code. We need to share it. This is what open source can teach the practice of medicine.


It is frustrating whenever personal medical questions become political questions.

This happens all the time. Example A, ripped from the headlines, is the kerfluffle over mammograms.

Women, doctors and politicians who hate having to go through the procedure are now screaming bloody murder over the non-existent "threat" to halt access to it.

Let's go back to where this started. It was a population study, done on behalf of the U.S. Preventive Services Task Force,  which concluded that the risk of unnecessary treatment exceeds the risk of death from annual screenings for the disease.

It was a science paper, and it was complicated, couched in the words scientists use to describe their work with precision.

Every profession has its version of this language. Engineers do, lawyers do, even software developers do. Learning the language gives a journalist entree into these professional worlds, but it's not an exercise most of us will, or should, go through.

Thus this study, like so many, was filtered through the lens of journalism. A flood of words failed to answer the questions women wanted answered

  • What are the chances I will die from a late diagnosis?
  • What are the chances I will suffer from over-diagnosis and over-treatment?

Calculate the threat to me so I can make a rational choice.

This is where technology comes into play. An Electronic Health Record (EHRs) can show you your own history, we can take your family history, we can analyze your genetic history, and we can estimate based on that.

If, that is, we have access to the larger pool of data. That's our baseline. It's what we need to compare your own data to before we can give you the answer you seek.

In scientific studies like this one we don't have access. The data is locked away somewhere. Mass adoption of EHRs is going to unleash a firehose of data, and the question should occur, very soon, what to do with it.

I'm not talking here about your record. I'm talking about the gross data, this warehouse of numbers describing everyone's condition, what is being done for us, and what the results are.

An open source attitude toward that data, within the realm of science and throughout the medical community, can help patients gain access to the benefits of that data and answer the question they ask -- what should I do?

Unfortunately medicine, health IT, and medical data all suffer from a proprietary attitude born of paranoia, the fear that you may be identified in this data mountain, that your needle will appear in this haystack, and that giving everyone access to data means giving them access to you.

A database, stripped of personal information, consisting of millions of records, is safe for use by software code. The data, and the code, are what we need to provide real answers.

We are collecting the data. We need to unlock it. We are writing the code. We need to share it.

This is what open source can teach the practice of medicine.

Topics: IT Employment, CXO, Health, Open Source

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Great article, Dana.

    You did in such a short piece what CNN, MSNBC, FOX and other media journalists failed to do: tell the story as it really is. You didn't take what was news worthy and sensationalized it, nor played up a negative reading of it to frighten those possibly affected, the way so called journalists did. I remember a time when a journalist could be trusted to relay the news without slanting it one way or the other. Keep up the good work. Too bad you're not on one of the major cable networks. You'd show them how the news ought to be reported.
  • RE: What open source can teach medical practice

    The Preventative Task Force has the rule of law under the new Health Care Bill. Even CNN agrees that they have absolute power and will fit the description of "Death Panels".

    Golly! Palin was correct!
    • Palin correct????

      Of course there is no "Death Panel." People like Palin make that claim because they are so dumb, they often react to things they can't comprehend with outlandish accusations, with no basis in fact.
    • Please stop the lies

      No it doesn't and you know it.

      But suppose that it did. All it would mean would
      that the most cost effective formularies would
      be the first to be tried, and the ones paid for,
      with your tax dollars. And with your insurance

      You want efficiency, you want lower costs, but
      you cry "death panels" when it's even suggested.

      • Go take your issue with CNN.

        Or take a warm bath in Huffington-land.
        The Flintstones
  • 2 issues...

    1- Current growth in Medicare and healthcare expenditures is unsustainable. At some point in the near future they will have to deny care to old and terminal patients. Label this anyway you wish. It will happen.

    2- Every single EHR & EMR fails to capture the data in a structured manner. Despite what they will tell you, the level of research on text data is very limited. Tagging the data with SNOMED is not enough. The solution is capturing each encounter in a true structured manner.
    • Very good point.

      Data that's not structured cannot be used for analysis. So this whole thing is not going to take off period.

      Other reasons why it is a non-starter:
      - Who's qualified to analyze it? What legitimacy will be given to an analysis uploaded by nick 'baseroomdweller567'?
      - Risk of abuse 1: insurance companies will be all too happy to pick and choose any 'evidence' they to deny you your claim.
      - Risk of abuse 2: what to do if some politically incorrect findings come out, like that white people of a certain age group are healthier than diabetes-bugged immigrant Asians? Racism! Bad bad whites! Tax them. ffff. You can just wait for it to happen.

      The Flintstones
    • You're right

      Without reform, we're going to have to make
      harder choices. Even harder than those we're
      making now, which is that poor people die. Soon
      it will be lower middle class people. Insurers
      and employers can't bear the burden of this
      industry's "growth.

      I don't entirely disagree on the difficulties
      with many EHR solutions. But this is a technical
      problem for which there will be technical
      solutions. It is an enormous opportunity. I hope
      you can make money from it.
      • "poor people die. Soon it will be lower middle class people."

        Yeah, right.
        The Flintstones
        • It's already happening

          Seen the Harvard study? 44,000 people died last
          year due to a lack of health insurance.

          Health reform is not about poor people. Poor
          people have Medicaid. It's about lower middle
          class people, who are losing their insurance and
          dieing right now.

          How's Barney, by the way? If a dino-crane drops
          a rock on his foot Mr. Slate won't cover him. He
          can't afford to.
          • Many don't care that the uninsured are dying.

            They just care about themselves.
          • So, do you care? What did you do about it?

            Or... was this just a feel good catch-all blurb.
            The Flintstones
          • Support the public option.

            I believe in universal health care. What's good for me is good for others.
          • Universal health care - a good thing indeed.

            That I can agree on.
            The Flintstones
          • Who gets to live?

            quote: Many don't care that the uninsured are dying. They just care about themselves.

            I've heard that false accusation too many times. And my response is the same. Covering the presently uninsured under any of the current proposals guarantees my children a staggering debt burden. More staggering than the one they are already inheriting from the generations that refuse to pay for what they are receiving.

            So if I get to choose between providing government subsidized medical care for those who don't have it now, and making a better future for my children and others of their generation - I would pick the latter every time.

            And consider that decision in light of the fact some (many?) of the currently uninsured are unwilling to make the sacrifices necessary to gain decent medical care coverage - sacrifices I do make on behalf of my family. Some are able to pay the cost, but choose not to. Some could change to an employer who offers a reasonable medical plan. It might mean less income, working where one would rather not work, or doing tasks one would rather not perform. These are individual choices now, but will be made mandatory with the costs being paid by debt (my children and their contemporaries) and those who are less in favor with the political ruling class.

            I'm actually OK with the European single payer model if everybody pays their share. The consequence of that model will eventually be a lesser level of health care - but at least everybody gets treated the same. And they seem to be on a pay-as-you-go basis (at very high tax rates) instead of mortgaging our children's future.

            But let's get back to the topic of the blog - true health care cost reductions brought about by standardized, structured, searchable electronic health records.
            thinking about consequences
          • Options not the same for everyone.

            Some people don't get the options you mention. Some work but can't find a job that provides health insurance.

            Nobody wants to saddle their children with burdens. However, when the choice is between health or wealth, I'd choose health every time. Money is of no use, if you're not healthy enough to spend it.
          • Ok I looked up the Harvard study.

            It talks of a "40% higher risk of dying". Actually, we all have a 100% risk of dying.

            What they probably mean is that they found a relationship. Not necessary a causal relationship. Ok I'm being critical here, but with statistics you can prove anyting, and subsequently use it for your personal goal of the day.

            There is also a relationship between women who have had an abortion and women who have experienced - or are in - a mental depression. Oops - that's politically incorrect - taboo even (unless of course you can blame republicans for that, then it's ok again).

            Anyway, you get the picture.

            But back to the unused data sets. I say: a non-starter.
            The Flintstones
          • No, you can't just do anything with data

            Peer-reviewed studies like the one from Harvard, which showed that uninsured people in ERs have a 40% higher risk of death, from the same causes as the insured, aren't to be dismissed so lightly.

            I find this attitude "you can prove anything with data" as excusing the idea that "everything is political and all I have to do is shout you down to create my own reality."

            It doesn't work that way.
          • Some more critical thinking please.

            Your 'attitude' response shows exactly why I made the point of (a)buse of statistics.

            I'm afraid I'll have to make it again, and explain a bit more. The fact that someone waves a percentage in front of your face, isn't sufficient reason to turn off your brain. Let's talk about those 40%, and let's assume that it's an accurate percentage (which could actually be argued against but I won't here). Was the cause of death a *lack* of insurance, or was the cause of death due to another factor. Did they investigate that at all? Is the health of the demographic 'insured' the same as the demographic 'uninsured'? I say it's quite reasonable to assume that people who for whatever reason do not have health insurance are in a demographic that are poor(er), eat worse, and exercise less. How about that. A bold statement, but worth discussing.

            Just goes to say that I don't buy the fuzzy reasoning 'no insurance = death' (and unsaid: therefore you must vote for my universal care you nazi')

            And yes, I'm all for universal healthcare. But abuse of statistics should not have a place in any argument.
            The Flintstones
  • RE: What open source can teach medical practice

    It wouldn't necessarily have to be "open source", just if the format could be standardized. For many, using the phrase "open source" might frighten them into thinking there PHI (Personal Health Information) is being left open, which health professions cannot do under HIPPA (Health Information Privacy & Portability Act). The number of different systems, most proprietary,
    in use in medical care, all fields, is staggering. The need for a standard format has been in place for decades. In the present system, every clinic, every pharmacy, every hospital and every insurance company potentially has their own format for storing data. The task of eliminating PHI from the resulting data pool would be monumental at best, virtually impossible at worst. A filter tool that worked on one format could render another totally useless, resulting in the need for another filter, and another, ad infinitum.
    What we need is standardization...this alone could have the potential to save billions of health care dollars.