Uncategorized

Medicine’s Missing Foundation for Health Care Reform

Preface by Michael Millenson: Lawrence L. Weed published a seminal article in the Archives of Internal Medicine on using the medical record to improve patient care back in January, 1971.

To give you an idea of how glacially the health care system changes, that same issue contained an article entitled, “Universal Health Insurance is the Wave of the Future,” by New York Gov. Nelson Rockefeller, and another, “What Possible Use Can Computers Be to Medicine?” by a Duke University physician that began this way: “The physician's attitude toward computing machinery has changed greatly in the last ten years. A bright future is predicted for its application within medicine.”

In an era when the autonomy of the individual physicians was nearly unchallenged, Weed boldly asserted that “modern data acquisition and retrieval systems” could help doctors make more accurate diagnoses and provide “proper care” more effectively. Weed has continued that same fight ever since, later joined by his son, attorney Lincoln Weed. In the process he has acquired neither fame, popularity or riches — merely become legendary to a small segment of us familiar with his work.

ABSTRACT:  Medical practice lacks a foundation in scientific behavior corresponding to its foundation in scientific knowledge.  The missing foundation involves standards of care to govern how practitioners manage clinical information.  These standards of care, roughly analogous to accounting standards for managing financial information, are essential to exploit the enormous potential of health information technology. Moreover, without these standards and corresponding information tools, evidence-based medicine in its current form is unworkable.  Medical practice has failed to adopt the necessary standards and tools, because its historical development has diverged from the paths taken in the domains of science and commerce. The culture of medicine tolerates unnecessary dependence on the personal intellects of practitioners.  This dependence has blocked the use of potent information tools, and isolated medicine from forces of feedback and accountability, that operate in the domains of science and commerce.  If the necessary standards and tools are adopted, health care cost and quality could become an arena of continuous improvement, rather than a quagmire of intractable dilemmas.

I. Introduction

    Medicine is built on a foundation laid by scientific knowledge.  Medical practice, however, lacks a corresponding foundation in scientific behavior.  This foundation cannot be found in "evidence-based medicine" and "health information technology."  As currently conceived, those developments are crippled by practitioner behaviors that the domain of science rejected centuries ago.  Medicine needs new standards of care to govern how practitioners behave in managing clinical information.  These standards of care in turn must govern the design and use of information technology.

    The necessary standards and tools have in large part already been developed, and have been used in patient care by some practitioners for many years.1  Yet, these standards and tools are largely absent from mainstream medical practice and medical education.  Indeed, their absence is not even recognized, because they are incompatible with behaviors and beliefs instilled by medical education.

    A simple example will illustrate the difference between mainstream practice and the rigorous scientific behaviors that medicine needs.  Consider the ritual that occurs when a patient sees a physician about an unexplained symptom to diagnose—chest pain, for example.  Physicians make educated guesses about the diagnostic possibilities that might account for the symptom.  They proceed with a history, physical examination and laboratory tests, making more guesses about what initial data might be useful for investigating the diagnostic possibilities that come to mind.  As they collect data, they make more guesses about what the data mean, and those guesses shape their judgments about what further data are needed.   These highly educated guesses vary from one physician to another, and are not explicit—physicians do not carefully record their thinking or the information they take into account.

   We use the term "guesses" because these initial judgments are made on the fly, during the patient encounter.  The judgments are based on whatever enters the physician's mind when conducting the history and physical examination.  What enters the physician's mind may be highly intelligent and sophisticated, but inevitably it reflects that physician's personal knowledge and experience.  A more rigorous, truly scientific approach would not be so limited.  Instead, diagnostic investigation should take into account everything that medical science has learned about what initial data are most productive for diagnosing the symptom in question.  Once collected, the data should be matched objectively with corresponding medical science, not judged subjectively based on the physician's limited personal knowledge of medical science.

   Take the chest pain example.  Careful review of the literature shows that investigating this symptom needs to take into account approximately 100 diagnostic possibilities (involving most medical specialties).  Each diagnostic possibility can be defined as a combination of simple, inexpensive findings from the history, physical and basic laboratory tests.  Making all of these findings for all of the diagnostic possibilities results in approximately 440 findings on each patient. Each positive finding suggests one or more diagnostic possibilities.  Each patient's particular combination of positive findings can be matched against all of the combinations of findings representing the diagnostic possibilities for chest pain.  The output of this matching process is a set of diagnostic possibilities, plus the patient's positive and negative findings for each.  These findings constitute initial evidence for and against each possibility, and the total set of possibilities represents the diagnoses worth considering for that patient. (Those possibilities for which no positive finding is made are not worth considering and can safely be ignored for that patient.)

    This matching process is feasible only when software tools are used.  Physicians do not have command of all the medical knowledge involved.  Nor do physicians have the time to carry out the intricate matching of hundreds of findings on the patient with the medical knowledge relevant to interpreting those findings.  Physicians thus resort to a shortcut process of educated guesswork.  Euphemistically termed "clinical judgment," this behavior causes a fatal voltage drop in transmitting complex knowledge and applying it to patient data.

    Digital information technology, because it makes possible an alternative to this behavior, represents a turning point in the history of medicine.  The alternative has been available for half a century. Yet, medicine has failed to embrace it.  In sharp contrast to other areas of science, medical practitioners are permitted to rely on personal intellect rather than tools external to the mind for basic information processing.  Even now, after medical error is recognized as epidemic and "health information technology" is increasingly employed, no standards of behavior have been generally accepted, much less enforced, to require the use of external tools for applying medical knowledge to patient data.

    The above discussion describes physician behavior at the outset of the diagnostic process.  That is just one example of how scientific rigor is lacking throughout the complex processes of care, both diagnostic and therapeutic.   Tolerating the variable judgments and habits and skills of autonomous physicians deprives the entire health care edifice of a secure foundation.  The outcome is to block the forces of feedback, competition and accountability that would otherwise transform the status quo.

    Other sources2 present in detail the standards of care and external tools essential to this transformation. What follows explores two questions.  First, how did medical practice diverge from science in defining the behaviors expected of practitioners?  Second, given that scientists are sheltered from the difficult conditions of patient care, is it feasible to bring the rigorous discipline of science to medical practice, where variables are uncontrolled and practitioners must cope with whatever problems patients present?  The first question takes us back 400 years to Francis Bacon, the first thinker who systematically examined the intellectual behaviors on which modern science depends.  The second question takes us from the domain of science to the domain of commerce, where scientific knowledge and technology are applied far more reliably and economically than has ever occurred in the domain of medical practice. 

> Part 2 – Medicine and the Development of Science

15 replies »

  1. I think that the health care should be reformed. For instance there are a lot of new medicine which take care of the new issues and problems that women for instance have. Eating disorderly is a disease and now they have invented the treatment so I think it should be included into the reform.

  2. We believe that a patient, once given the tools to empower self-management (diabetes monitoring, integration with sensor devices, health rewards, coaching, messaging) will be a better health care consumer. We believe a solution, tool, or strategic approach will only be successful if the provider and end-user physician will see benefit in adopting the platform and promoting it among their patients.
    Any health care reform should include an entrepreneur and a representative from a small business. These are the people that drive the ingenuity and are the end consumers of the health care world. Combined with the capital and IT infrastructure that large organizations can supply, our health care system will be in a much better place in the next four years.

  3. Peter N., I found your comment interesting and share some of your thoughts.
    “Lets face it, in the current environment doctors simply prescribe too many tests and make too many specialist referrals that are not justified by the clinical findings. This leads to ever growing medical costs.” – true, but it is to a great part the patients and a dysfunctional culture that demands tests, tests, tests.
    And the problem is: this culture is self reinforcing. Who did not hear a story about pat. X who was not diagnosed because the right test wasn’t done. But who, outside the medical community, heard stories about people being sent on diagnostic odysseys for doubtful findings, findings that were noted on tests that were not indicated to begin with? Or the patient who was hurt by unnecessary procedures/surgery?
    Moreover, certain tests with a low pretest probability “reward” the ordering physician by showing incidental findings – take a brain MRI odered for no good reason: you still may find nonspecific white matter lesions (that guarantee you a trip to the neurologist if extensive enough and dep. on the radiologist’s wording), you may find a silent infarction, a meningioma … some of the findings may change a patient’s management (although the benefit may be marginal). Once in a blue moon, there may be something truly worth addressing, such as a large aneurysm (but what about the small and/or borderline aneurysms where the cure may be more risky than the problem itself?).
    Re. multiple tests incl. brain imaging for headaches and screening for aneurysms, there are clear guidelines … and yet only few neurologists (and rarely ever any PCPs) follow those.
    Yes, managed “lean” care (who brough up this term? I read it a few times on THCB, but it does not seem to have any theoretical basis that I could google) is the solution. But not only is there reason to doubt that the doctors are ready … but the patients aren’t ready either.
    And neither are our courts/juries. They still analyze cases with hindsight bias … and as long as this doesn’t change, the simple cost benefit analysis for a physician is: I go order the tests … I don’t have to pay for it.

  4. I have a bias against Doctors already but I find this interesting. I’ve been to doctors who have done nothing for my Eustachian tube dysfunction, I had it for 2 years spent a lot of money in doctor bills and not only did they not fix the problem but never would send me to an ear specialist. Apparently they don’t understand what lots of extra pressure in your ear and head feels like. It was finally chiropractic care that treated me like a human being and helped me to feel better. I made a blog about some of the things that helped me to feel happy and healthy at my blog: http://nutritionalfitness.blogspot.com, I actually made a post about this very topic recently called Do Doctors Really Help: http://nutritionalfitness.blogspot.com/2009/04/do-doctors-really-help.html

  5. MD as HELL wrote: “It is not about data. It is about problem-solving.”
    Yes … and No!
    Yes, it all boils down to problem-solving and decision-making. And yes, useless data is, well, useless, and no one needs it. HOWEVER, what IS needed is comprehensive cross-disciplinary health data that is transformed into useful information and presented in ways that support clinical decisions without impeding the care process.
    For other than simple/basic health problems, the most valid and reliable decisions require understanding such things as the diagnostic criteria, the causes (unless idiopathic), the treatment options (including conventional allopathic and CAM), the benefits and risks of those options, the patient’s idiosyncrasies, possible influences of the mind-body connection and genetics, drug interactions and medication side-effects, and the patient’s awareness and knowledge, etc. Of course, not all these knowledge domains are necessary for every person in every situation, but then you’ve got to know what knowledge is essential for understanding a patient well enough to make valid decisions about the person’s care.
    In addition, we’ve got to know how best to educate, instruct, and motivate patients for better self-maintenance. And we need better ways to enable clinicians to share lessons learned and to assist researchers in studying clinical outcomes for the creation and evolution of evidence-based guidelines.
    I believe the vast majority of clinicians do their honest best to help their patients. Unfortunately (a) we have to practice in a broken healthcare system and (b) since the human mind is limited in what it can retain (and this is an immutable fact), and since the time constraints of current day practice force physicians into make rapid decisions, assisting clinicians in problem-solving and decision-making would likely make things much better. But what has frustrated me for past 30 years relates to Dr. Kibbe’s statement: “Unfortunately, I don’t think the mainstream of the medical and surgical professions is listening. It’s going to take a significant shock to the system to put these reforms into place.”
    Well, it’s shock time … things simply cannot continue the way they have been! Radical changes are needed. One aspect of such radical reform is to being developing and using computer-assisted, patient-centered cognitive support tools; see http://curinghealthcare.blogspot.com/2009/04/defining-meaningful-use-of-health-it.html. It is what the healthcare industry ought to start focusing on since ever-increasing care effectiveness and efficiency are in our nation’s (and the world’s) best interest.

  6. This is an interesting discussion about the mental process of diagnosis but it does nothing to address the real issue confronting medical care and health care reform. Lets face it, in the current environment doctors simply prescribe too many tests and make too many specialist referrals that are not justified by the clinical findings. This leads to ever growing medical costs.
    Dr. Weed notes, “Each diagnostic possibility can be defined as a combination of simple, inexpensive findings from the history, physical and basic laboratory tests.” We need to reform our health care system to ensure that doctors are practicing lean medical care. Software that allows management of the entire system is essential but not sufficient. Rather, the willing compliance of physicians to lean medical care is the necessary and sufficient condition to gaining control of medical costs.
    This reform will not come from the bottom up. Physicians,for a variety of reasons, will not suddenly adopt lean medical care principles. I think that Dr. Weed would do better to address his thoughtful assessments to a system that binds together a universal network of cooperative physicians managed by a team of care coordinators who first and formost care for patient interests but also care equally for the interests of the physicians who prescribe the care for which every health care dollar is spent. Unless physicians modify their behavior, health care reform is a pipe dream of politicians who somehow think shuffling the old cards will get health care a new deal.
    Managed lean medical care will come eventually because there is no alternative except rationing. Ultimately doctors must change their behavior but wouldn’t it be astonishing if the entire medical community could somehow band together to bring about this change voluntarily. I suppose that’s my pipe dream.

  7. What spectacular arrogance. What beautiful dilusion. What a crock. Patients lie and deny. The doctor is the interface between the patient and the truth. Getting the game in the right ballpark is a skill that is harder than is commonly thought. Someone still has to decide what to do. Probability is not a standard for anything. A wise neurologist once remarked to us in a review course that a differential diagnosis is a list of wrong answers. Why generate a list of long and useless answers? Someone still has to make a decision. That is impossible for many people taking care of a patient. Do you really think more data and choices wll help? Either the patient has diagnosis x or she doesn’t. The trick is to decide with limited data, for that is the usual case. have you never played “Mastermind”? Have you never solved a Sudoku puzzle? It is not about data. It is about problem-solving.

  8. This is an important writing, and an essential discussion. Unfortunately, I don’t think the mainstream of the medical and surgical professions is listening. It’s going to take a significant shock to the system to put these reforms into place.
    Regards, dCK

  9. I’ve followed Dr. Weed’s work and writings for several decades; he’s a wise man and remarkable innovator! He was one of the people who helped me realize that it is foolish to expect the unaided human mind to know all that needs to be known to make valid and reliable diagnostic and treatment decisions in many situations.
    Consider the following, which I wrote at http://wellness.wikispaces.com/The+Knowledge+Void; it includes quotes from Dr. Weed’s earlier writings:
    The volume of clinical information expands exponentially with more than 150,000 medical articles published each month in more than 20,000 biomedical journals. Healthcare providers struggle to stay current with the clinical information, but inevitably become overloaded. This may contribute to the slow adoption of evidence-based research in clinical practice. There is just too much clinical information being generated for providers to incorporate into their internal base of knowledge.
    As a group, healthcare providers care about patients and take pride in doing an excellent job in caring for their well-being. Nevertheless, the “…task of knowing every detail is way beyond the [ability of] human mind… For example…diabetes care ought to take into account any of 120 management options and 380 possible patient conditions associated with the disease. …the unaided mind cannot reliably recall all the causes or management options that should be considered for each patient, nor can it recall all the findings in the patient needed to discriminate among those options, nor can it reliably match findings to options under the time constraints of practice.”
    In addition, “…most physicians are able to take into account only a portion of the diagnostic and management options potentially relevant to their patients and only a fraction of the evidence needed for choosing among those options. …Physicians do little better with the usual aids to medical decision making, such as practice guidelines and use of Internet resources. Those aids provide general knowledge, but do not ensure that the physician will recall all the data or successfully link it with a particular patient’s specific problem.” Problems that cross specialty boundaries and require multiple specialists, yet the current healthcare system does a poor job at supporting communication between providers and assuring continuity of care. And primary care physicians are not equipped with the information tools necessary to grapple with the information overload, nor do they have a system for coordinated care within which to function.
    “Because physician time is expensive and scarce, their initial workups can be meager [as they] …act according to their own preconceived notions about what history, physical, and laboratory findings are worth checking. Equally idiosyncratic are the conclusions they draw from whatever data they select. Both selection and analysis of data are influenced heavily by their medical education, prior clinical experience, specialty orientation, contradictory clinical guidelines, financial concerns, cultural background, personal biases, and day-to-day time constraints, all of which vary enormously among individual practitioners.”
    With this in mind, consider the recent report by the National Research Council of the National Academies, which concluded that a serious gap in health information technology (HIT) today is patient-centered cognitive support. According to the report, we need: “computer-based tools [that] examine raw data relevant to a specific patient and suggest their clinical implications given the context of the models and abstractions. Computers can then provide decision support—that is, tools that help clinicians decide on a course of action in response to an understanding of the patient’s status. At any time, clinicians have the ability to access the raw data as needed if they wish to explore the presented interpretations and abstractions in greater depth…The decision support systems would explicitly incorporate patient utilities, values, and resource constraints…They would support holistic plans and would allow users to simulate interventions on the virtual patient before doing them for real.
    This means that the next generation of HIT should focus on making the dissemination and implementation of useful scientific knowledge a top priority or we will continue to have serious quality problems and escalating costs due to ignorance, mistakes, and inefficiencies.
    Steve Beller, PhD

  10. It is a very innovative article. The challenge is not the proposal but the state of analytical modeling. Medcine is not a complete science yet..it is mixture of art and science.
    Second problem, as I think alluded in the article, is the inverse of the cause and effect relationship in diagnostic. It is easier to find a symptom for a given cause. It is more diffult to find a cause for a given system….because for one symptom, there are hundreds of causes. Now, eliminating all takes expensive diagnostics or an intelligent guess by a doctor.
    rgds
    ravi
    http://www.biproinc.com
    http://blogs.biproinc.com/healthcare

  11. Excellent essay on a crucial fundamental point.
    The practice of medicine for the most part relies on the limited brainpower and inherently biased experience of single individuals. Individuals can only consider two or three alternatives at one time. If there are 40 zebras, they are discarded and you only look for the few horses that you know.
    Attempts to introduce the scientific method to medical practice are denounced as ‘cookbook medicine’ and infringement on ‘freedom of choice’. (The freedom to make bad decisions about patient care.)
    These anti-scientific attitudes prevent medicine from taking advantage of scientific knowledge and information technology. We will not make any progress until these attitudes change and I fear that it will literally not take place until the current generation of practitioners retires. Unfortunately, their replacements are not being trained in the scientific practice of medicine either.

  12. Larry Weed-MD by the way, is one of the great thinkers and doers in progressive, innovative healthcare delivery. Over 30 years ago, he developed the concept of SOAP notes (subjective, objective, assessment, plan) used extensively by physicians as a framework to document patient encounters.
    His automated problem knowledge couplers PKCs are robust decision tools that bring the best medical research and thinking to the bedside and office but alas they’ve been under-utilized largely due to a prevailing culture that relies instead on recalling vast quantities of memorized medical research for decision making as discussed herein.
    As was discussed by several panelists at last week’s Health 2.0 conference, as well as Larry and Lincoln Weed, the future will hopefully see more direct patient management and even delivery of their own care, including lively comparison and discusion about key health metrics. This could include completing simple lab tests like cholesterol at home with kits from Target for $12, vs. $60-80 at the doctor’s office. Put it all on a PHR-personal health record where it belongs for patients to own and manage-only about 50% of recommended preventive care is provided now anyhow in outpatient settings according to the data!

  13. I wish that the authors had submitted a concise summary.
    I think the authors have a good understanding about the diagnostic process in complex diagnostic situations. One has to consider that medical practice is, most of the time, routine work (both diagnostically and therapeutically). However, I agree that diagnostic support for the remaining complex patients is paramount.
    As soon as it works and can be done efficiently. I toyed a little with Dxplain, and I am highly skeptical that it comes close to the maligned “clinical judgment” of a trained specialty physician. If there is a medical record out there that can process symptoms and suggest a differential diagnosis, it might be of great help to the primary doctor, and sometimes to the specialist as well.
    Like so many THCB posts, this one again seems to greatly overrate the blessings of participatory medicine (PM). Patients asking for inappropriate tests are the rule, while the correct self diagnosis or beneficial treatment decision is the exception. Not that PM should be discouraged, but physician guidance is needed, and patient skepticism about the results of lay googling and lay digestion of medical ressources needs to be reinforced. Even doctors misinterpret clinical info at times – how could patients be expected to get it right more often than not?

  14. Information out is only as good as information in. The judgments are based on whatever enters the physician’s mind when conducting the history and physical examination. The same can be said of computers and algorithms. You would think that we as physicians are missing the majority of diagnosis and treatments. I treat the same diseases over and over again 90% of the time. The only difference is that the information in is different from each patient.