Diagnostic Error

I attended the meeting of the Society for Improved Diagnosis in Medicine this past week in Baltimore. This is a new society, being formally constituted within the past year, although this was the fifth meeting. Previously it had met with other groups, most recently the Society of Medical Decision making.

The major focus was one developing tools and approaches to define the scope of the problem which likely huge but difficult to define and quantify. It was a fascinating group of people which included physicians from a host of specialties. However, it was not limited to physicians and I would venture to guess that nearly half (and perhaps more than half) of the attendees were non-physicians, including many patients.

Even that characterization is not entirely accurate since every one of the attendees are currently recipients of some form of medical care, some more intensively than others. This is of note in that all inquiries into the sources diagnostic error and approaches to addressing diagnostic error interested with the patient communities. Their stories of harm were compelling and at times overwhelming. Their stories were not an add-on to the meeting but in fact central to it.

I was struck by the gulf that still separates patients from physicians and the health care delivery system, a gulf that is maintained by perceptions, processes, and desires to maintain power. However, as I heard the stories of patients, many of whom were very smart and sophisticated, I realized that the elements which maintain the separateness are the same elements which are a source of diagnostic error in the first place.

My perspective as a physician is shaped by the perceptions of those I trained with and those I work with. We view patient histories with great skepticism. What we hear from patients may not be accurate, relevant, or complete. That may be because they do not give us complete stories, not surprising since we demand they provide them under the most time constrained conditions where they suffer from significant performance anxiety. It may also because we fail to hear what they are trying to say and fail to give them sufficient time to reflect upon what the feedback we give them.

My residents often complain that I am able to elicit a history different from what they obtain. I found this very curious until I had a realization as to why this happens. Resident interviews act as a facilitator for patients to reflect upon what we as physicians want to know. Patients come into visits as poorly prepared as we do an questions offered by the residents surprise them. Answers obtained under those circumstances are often incomplete or wrong. However, in the time supplied after the resident leaves and I arrive, patients reflect and rethink their answers, having time to get their stories straight. No guile involved; just the opportunity for reflection improves the quality of the information delivered.

There were lots of interesting aspects of the conference and many interesting speakers. However, one specific session made a particular impression. It was very focused on the role of patients in addressing diagnostic error. The roles proposed proposed for patients raised in my mind the thought that addressing the problem of diagnostic error will require a fundamental change in how we view the role of patients in the diagnostic process. In each of the patient stories of missed diagnosis the opportunities for improvement and avoidance of similar events appeared to require an engaged and activated patient providing a feedback loop to the health care provider to avoid diagnostic pitfalls.

Once the patient is viewed as part of the diagnostic team, it creates the need to define boundaries and ownership. It was absolutely fascinating to hear the divergence of opinions between the physician and patient communities on where these boundaries should be and there was great concern that the unintended consequences of unfiltered information sharing could be dire. However, it was also pointed out that we survived having family access to delivery rooms and ICUs and that patient and family involvement has turned out to be an enhancement, not a detriment to healthcare delivery.

There is no question in my mind that the old model of care delivery has a short half-life. Patients want and deserve better .However, one particular attendee who was a primary care MD spoke up. He was a bit annoyed the primary care physicians were often pilloried as central characters in stories of harm. He urged patients to ask why their primary care physicians cut them short, ignored their complaints, and appeared to be unsympathetic. Providing thoughtful and reflective care using shared decision making takes time and effort. That fall squarely in the uncompensated or poorly compensated quadrant of physician efforts.

Most of the stories of harm and missed diagnosis could be linked back to insufficient time to hear or assimilate information from patients and to integrate all of the pieces. Circumstances and models where this does not happen now generally involve heroic efforts of extraordinary people and non-scalable or sustainable models of care. Patients may want and deserve better, but patient leaders need to understand they are dealing with an economic problem. The current administrative pricing structure in health care values physician activities required for reflective and shared care at essentially nothing. The price signals linked to these activities are telling doctors that there is no value in these activities and against this relentless tide of pressure, no one but the most heroic of physicians continues to resist.

Economic laws are almost as immutable as the law of gravity. It takes much energy to keep things from falling. Items valued as nothing tend to disappear over time. It is not surprising we have found ourselves with a scarcity of reflection and time spent with those who patients believe should be their advocates and counselors. Our own moral compasses as physicians tell us when we have time to reflect that we should provide this time to patients, whether we are paid to do so or not. However, that is simply unrealistic. As things stand now, we are the engines that support not only ourselves but entire teams of people who are almost completely dependent upon the revenue generating capacity of physicians. Thus, we make compromises and rationalize these decisions on the basis that overall we do overwhelmingly more good than harm and it is important to keep the lights on and the doors open. We do our own little bits of God's work and recognize that we cannot fix all of the world's ills.

The payment system that fails to incentivize physicians to deliver what is needed to address diagnostic error continues and, for the most part, few recognize that the scarcity has been created by the dysfunctional administrative pricing system. While patient engagement is an essential piece to address diagnostic error, it is necessary but not sufficient. Patient engagement creates activated patients. However, this is of limited utility without activated clinicians. Some clinicians will become activated because of their moral compass but for the broader provider audience, activation will require financial rewards or at least the lack of financial deterrents. The sooner the activated patients realize this, the better off they will be.

Turning the EHR into a project management tool

I have spent a few hours this morning finishing my notes from my clinic yesterday. It occurred to me that for each patient when my note was completed, I had a series of tasks which needed to be done in the future. However, the EHR I use is really poorly equipped to aid me in tracking my future tasks.

EHR's still tend to be used as documentation tools as opposed to something that adds value by helping our brains remember what needs to be done, what has been done, and helping us decide what should be done. There are circumstances where the EHR needs to function simply as a note generator, but that really barely touches upon where it can provide value.

For each patient who I follow long term, I need the equivalent of a project management tool since in many respect, the care of each patient is a long term project. For each patient I need to define what is on my to do list, what am I in the midst of doing, what specific action items are on my plate, and what have I delegated to someone else. At this point in time, the EHR I use does none of this well and many of the tasks it simply has no functionality to use.

Hard to believe that we continue to use such dysfunctional management tools.

Medical Reformation

500 years ago, Martin Luther nailed his 95 thesis on the door of the local cathedral.

What this event marked was a dramatic change in the relationship of people with the most important human institution in western Europe at the time; the catholic church. Luther's actions was preceded by key technological advances of the day, most notably the development of the printing press by Gutenburg.

Prior to the printing press, the church through the clergy controlled access to the printed word, primarily in the form of the bible. After deployment of the printing press, the holy bible, and all of the wisdom contained, was no longer was controlled by the priesthood. Then as now, he who controlled information controlled power.


Roll the clock forward to the last century. The status of physicians in the 20th century underwent a profound change. In the previous century, physicians held a status only slightly higher than barbers. With notable exceptions, most physicians has limited useful knowledge and few tools to deploy when confronted with serious medical situations. The developments of anesthesia, antisepsis, and antibiotics changed all that and by the mid portion of the 20th century, physicians held the status of gods. As Arthur C. Clark once noted, “When technology becomes sufficiently advanced, it becomes indistinguishable from magic”, and medical magic for most of the 20th century held the public is complete awe.

The continuation of awe was also facilitated by the relatively proprietary nature of the knowledge base behind which the medical magicians practiced. We held our cards close and as a rule the public was sufficiently intimidated to inquire further or question physician status.

The internet has begun to change all of this.

Like the printing press, it has taken cloistered texts and other writings out of the hands of the clergy, in this case the medical clergy, and made it available to all those who can search and read. The impact is profound and likely to become even more profound. Prior to the reformation, the catholic church firmly held to the belief that the bible and the information it contained was best controlled by the Church. So much was at stake, that being the souls and eternal fates of human beings. People simply could not be trusted with such responsibilities. Only the special priest class could be entrusted with such a role.

The arguments are being put forth by the medical priesthood. Too much as stake, and patients cannot be entrusted with their own lives. However, the genie is out of the bottle. The information is already out there and more will become available to patients every day. Some of it may be unreliable but the same can be said for what is vetted by peer review in our own fields. Furthermore, patients have an inherent advantage over their health care providers. They have much more time to invest in their own care and they have so much more skin in the game.

I saw the Medical Luthers posting their 95 theses on the door today. They are doing it everywhere. There may be health care providers who will excommunicate their Luthers. I am not sure they have much leverage. The relationship is going to be changed forever.

Speaking different languages about health care as a right

Now that the 2012 elections are over, there is a tendency for those whose perspective is more statist to view that they have received a mandate to go forward with their vision of health care reform. From my perspective, the election provides only the starkest of evidence of a closely divided country where a slim majority supported President Obama's vision of state driven health care changes.

Do not get me wrong. Our present system is terribly flawed and the trajectory of spending is unsustainable. There are constituencies who are committed to holding patterns attempting to extract as many rents from the current system as possible, hoping they can reach retirement age before their goose that lays golden eggs goes away. However, among virtually all thoughtful people there is almost complete agreement that what we are doing presently is not viable in the longer term. The issue is not whether change will happen but instead what that change will look like and who or what will drive it.

President Obama and his supporters tend to hold a belief system which places great confidence in the ability of the law to drive change and to structure human systems in such a way to that law is always a positive force. This belief system is predicated on the assumption that what is most important is the intent of the law and that smart people are fully capable of creating rules, often very complex, which can predictably result in improved incentives and improved outcomes. I should also note that professional politicians on both sides of the isle share this set of assumptions, albeit they may have different levels of confidence that the law can be used in different realms to coerce people to behave. Elements of both the left and the right can be all to willing to seize the reigns of a coercive state into order to bully other elements of the populace.

As I see this, the fundamental gulf that exists revolves around what groups view as the priorities of what the governing leadership should be. What are currently viewed as liberals or progressives (and I hate these terms) prioritize moving toward a world which should exists while conservatives (which is what I view myself as being) place priorities on moving toward a similar world but constrained by what can be. As I wrote about in an earlier blog, von Mises noted in his work "Human Action" (https://www.blogger.com/blogger.g?blogID=2308282620289958037#editor/target=post;postID=7269024980841579619), there is really little disagreement on the articulated goals of:

It is obvious that they do not differ from one another with regard to ends but only as to means. They all pretend to aim at the highest material welfare for the majority of citizens.

To me, the most important discussions should focus not on what we aim to achieve but instead on what is actually possible. Going all in aspiring to use the power of the state in order to strive toward a perfect world may sound good, but this experiment has been tried before. The product was not perfection but was instead perfectly awful. What limits our actions in regards to delivery of health care and specifically, what are the limits on state action? Why can't the state use the power of law to simply declare that health care is a right and that ALL those within the borders of the US should have access to health care?

Obviously entire books have been devoted to this and I will not exhaustively delve into all aspects of this question in this essay. The undisputed fact that so much has been and continues to be written belies the complexity involved. However, at the simplest level, you cannot mandate access to what you cannot define, particularly when the definition of what you are referring to is constantly morphing and evolving. Making something a legal right backed by the coercive power of the state requires that you must define the scope of that right. Decentralized private markets and charities are much more nimble in this regard. What one party does at any given time can be distinctly different from other parties involved in similar activities. This is not the case for legal rights. If one cannot define the nature of the right involved, it moves the action, however desirable at a theoretical basis, into the realm of practically impossible. Every decision regarding what is covered can be moved from the personal and medical into the legal realm. If you don't believe this will become unmanageable almost immediately, please refer to the case of Michelle Koselek in Massacheusettes (http://www.cbsnews.com/8301-201_162-57520960/mass-officials-to-fight-murderers-sex-change/).

The issues with defining the scope of any health care right was recently highlighted in the NEJM (http://www.nejm.org/doi/full/10.1056/NEJMp1208386)in a piece by Neuman and Chambers entitled "Medicare's Enduring Struggle to define Reasonable and Necessary Care". In the article the authors note:

Since its inception in 1965, Medicare policy has been guided by legislation mandating that the program not pay for items and services that are not “reasonable and necessary.” Over the years, amid escalating costs and the medical community's embrace of evidence-based medicine, the Centers for Medicare and Medicaid Services (CMS) has struggled to interpret and apply the “reasonable and necessary” criteria. At key junctures, CMS has been thwarted by political pressure or the courts. As Medicare spending takes center stage in the country's budget debates, “reasonable and necessary” warrants a closer look.

Neither necessary of reasonable is definable in any way meaningful way which will create any enduring structure which can address allocation of scarce resources. Neuman and Chambers dance around this issue and conclude by noting:

It may be tempting to believe that the matter will be rendered moot by payment reform and premium-support policies. That is, some may hope that the federal government can simply delegate coverage decisions to other parties, such as accountable care organizations, while forcing patients to consider the value of technologies through increased cost sharing. Such reforms are needed, since they will help move CMS out of the business of micromanaging coverage policy, though the details will be crucial. Offloading financial risk, however, does not absolve Medicare. Although it will shield CMS from certain controversies, questions will persist over how much geographic and socioeconomic variation in technology coverage the country will tolerate in a federal program. Moreover, the steady march of big-ticket, high-profile technology, such as cancer therapies, will demand a single response from Medicare regarding the adequacy and reasonableness of the evidence base.

Thus they admit that the creation of Medicare requires implementation of one size fits all approaches to defining scope. Where will those decisions be made? Whatever the initial intentions of those charged with defining and implementation will be co opted in the realm where all government decisions are made; the political realm. Political allocation of resources yields decisions which generate votes. The political realm works in a winner take all environment where those who can muster tiny majorities can impose their wills on substantial minorities, even if their governance may be unwise.

If the purpose of actions is to strive toward some ideal, whether effective or not, then aspiration alone can be a measure of success. However, I aim for more and am fearful of Utopian aspirations used to justify concentrations of power as a means to attempt what may be theoretically desirable but is almost certainly not attainable in the world of imperfect humans playing politics.

The flaws of encounter based healthcare delivery

I am reading Lawrence and Lincoln Weed's book "Medicine in Denial". I became interested in this book after seeing a video of Dr. Lawrence Weed's Grand Rounds presentation at Emory University which dates back to the early 1970's. I blogged on this presentation earlier this year (http://georgiacontrarian.blogspot.com/2012/07/failure-of-our-intuition.html). Dr. Weed popularized the problem based medical record concept. His views on information collection, analysis, and use were far ahead of his time and concepts that he recognized 50 years ago are only now getting more widespread acceptance.

The concepts are really rather simple. Dr. Weed believes that everything starts with information collection and that to move rapidly toward decision making and action before a broad information collection effort is undertaken leads to disjointed and sub-optimal care. Dr. Weed divides the tasks into choice, collection, and analysis. I similarly divide the tasks into collection, analysis, and decision making and group the choice of which data to collect into the collection step while adding the decision making step which falls as a consequence of the analysis.

Under the current conditions, we attempt to cram everything into a very time constrained office visit. We might do some modest preoperative work ahead of time but for the most part, we experience our daily schedule of patient encounters as a series of agenda-less meetings. We walk in cold and start the process almost entirely from scratch, hurrying to collect relevant data, taking every shortcut imaginable to fast forward to diagnosis and plan implementation. We structure the encounter to limit the amount of any patient push back, with success being measured by how quickly we can get patients to accept our snap judgements and shoot from the hip plans. So much for shared decision making except in your concept of this is that you share your decision and the patient quickly accepts them.

Dr. Weed recognized that many of the functions now undertaken by physicians under hurried circumstances can be delegated to non-physicians under less time constrained conditions aided by software which drives more comprehensive data collection. The net result COULD be that we enter the encounter much better prepared without the need to engage in extensive data collection. With data collection for the most part done and the data presented in formats enhanced to facilitate analysis and decision making, the physician's job would be to facilitate decision making and implementation.

The key to deployment of this model is to put in place the data collection piece. In my own operating unit within a large "integrated" health care organization (and integrated is in quotes for a reason), we have made strides in moving toward collection of structured data in the clinical environments. However, the tools we use are still focused on office based encounters. The electronic tools which we are using which allow for the collection of structured data allow us to do this ONLY when we have the patient in the office. If we have an remote interaction and we place notes in the chart relating to this interaction, it is stored not as structured data but only as text blobs which can be mined for data only with great difficulty.

I suspect that our circumstances are not unique. The electronic medical record a still structured best as a billing justification tool, not a communication and performance improvement tool. It will be a huge waste of money if we spend billions of dollars to deploy a tool which locks us into bad workflow. Key to moving to efficient and less expensive care is offloading tasks such as data collection from expensive providers and moving activities from high cost environments such as hospitals and medical offices to where the patient are. If the only place our electronic tools allow us to collect key information is in high cost environments and that the tasks can only be done by high cost providers, we are sunk.

My blog sabbatical

I have been away from my blog. I can't say for sure why I have not been inspired to write. I suspect it has been a combination of factors: time, the election, Hurricane Sandy, and the natural cycle of writing.

I find politics both consuming and possibly irrelevant. Some much of what candidates say in the lead up to elections is really irrelevant to what they actually try to do or succeed in doing once elected. In many respects political leaders function within political systems very much like our individual conscious selves within our brains. Jonathan Haidt writes (in his book The Happiness Hypothesis) of a rider on a conscious rider on an unconscious elephant.

" I'm holding the reins in my hands, and by pulling one way or the other I can tell the elephant to turn, to stop, or to go. I can direct things, but only when the elephant doesn't have desires of his own. When the elephant really wants to do something, I'm no match for him.
...The controlled system [can be] seen as an advisor. It's a rider placed on the elephant's back to help the elephant make better choices. The rider can see farther into the future, and the rider can learn valuable information by talking to other riders or by reading maps, but the rider cannot order the elephant around against its will...
...The elephant, in contrast, is everything else. The elephant includes gut feelings, visceral reactions, emotions, and intuitions that comprise much of the automatic system. The elephant and the rider each have their own intelligence, and when they work together well they enable the unique brilliance of human beings. But they don't always work together well."

Under ideal circumstances, like the rider a political leader can lead by seeing into the future and help the electorate make better choices. However, it rarely works that way. The electorate generally fails to understand that political tools are very limited in terms of what issues can be effectively addressed. Political entities can pass laws which ostensibly can compel people to behave in certain ways. They also can take money from from one set of people and transfer it to another set of people. Neither of these tools are particularly precise and as political entities become more ambitious, the intended consequences tend to become dwarfed by the unintended consequences of their actions.

Still, people tend to overstate the capabilities of government to solve problems. The

issue is long standing. In the old testament in Samuel, the Israelites demand that they have a King. Basically everyone else had one and it was reasonable to believe that their problems could be more effectively addressed by allocating power to a single wise individual. However, concentrated power generally ends up in the hands of those who desire it most and more often than not, they are motivated primarily NOT by their desire for public service. Furthermore, even if they are fundamentally good, the elephant driver issue will render them less than optimally effective. The elephant will end up going where it wants to go short of actions which incapacitate the elephant. No one is a winner under those circumstances.

I think back upon my undergraduate days when I was friends with a Mr. John ___. He was part of a menagerie of very interesting and entertaining individuals all residing in my dormitory. It was a very different world whose most distinguishing feature was that its inhabitants still retained a very robust sense of humor. Most of us were not living our lives in an attempt to avoid any action which could affect our political careers 20 years in the future. We did things which now in retrospect might have required some degree of explanation on current applications for hospital credentials.

I believe it was during John's junior year that he decided to run for president of the the student body. He was a consummate outsider, not a member of Greek life unless one viewed him as an official member of GDI. We were a cynical bunch, cynical of politics are any level. The Watergate scandal was ongoing and there was skepticism that any of the scoundrels could be trusted. John came up with what I still believe was a brilliant idea. He would run on the apathy ticket and his slogan was to be "Don't vote. Elect John!". It was his plan to claim all the votes of students who failed to vote. His election platform was to dismantle student government and his mandate would come if less than half of the student body voted.

As it turned out, slightly more than half of the student body voted, thus undermining John's claim to higher office. However, the story was not so straight forward. For reasons not entirely clear, some of the students actually voted for John and if one to remove these votes from the total, less than half the student body actually voted. Legitimately John could claim that those who voted for him cost him the election! He was heart broken and I believe that this was his last foray into the quest for higher office.

I think John basically had it right. Be skeptical of the capabilities of what elected offices can accomplish. It is too bad that this message is so hard to translate into enlightened leadership that understands its limits. You can be undermined even by those who are trying to help.