The power of self correcting systems: why feedback is essential for the creation of enduring entities

I remember when I was a trainee at the NIH seeing two brothers from South America with a rare disorder. They were born with the absence of nerve fibers allowing them to feel pain. They were young adults, having survived chuldhood..sort of. They had no fingers, toes, lips, or nose. They were missing a key system that provided them with critical feedback in real time and as a consequence, their physical structure became seriously degraded.

Pain of some sort is a crucial element of any system or entity which aspires to be enduring. There will always be places where we should not go and activities which we should not undertake.  A world without pain (or at least without sufficient pain) would be a world without necessary feedback. It would be a very dangerous and undesirable place.

We have examples of where particular individuals have ended up operating in niches where they have received insufficient feedback of the negative variety and ultimately ended up in very bad places. A recent article in the NYT described the rise and fall of the career of Dr. Dietrik Stapel ("The Mind of a Con Man"). The article, written by YUDHIJIT BHATTACHARJEE, describes a Dutch social scientist who fabricated data over the span of two decades. As noted in my previous blog, the Roman playwright  Terence noted, “One easily believes what one earnestly hopes for”, and by all accounts, Dr. Stapel's work tended to cater to the belief systems of the social sciences world.

Science is supposed to be self correcting and I guess if one thinks about the time in the geologic sense, Dr. Stapel's career and the errors he has created were discovered fairly quickly. However, my question is how many Stapels are out there fabricating data and remaining undiscovered. Stapel's problem may have been he to was too successful and ambitious.

In medicine and biomedical research, we have had our share of con artists. John Darsee had a meteoric rise as a resident and cardiovascular researcher at Emory University in the late 1970's, landing him a Harvard faculty position. When his work was scrutinized, evidence of falsification of data went back over 15 years and at least at three different institutions. Feedback to Darsee was almost all positive, based upon his remarkable (and unrealistic) productivity, viewed with an uncritical eye. Despite what the phone commercials may say, more is not always better.

One of the problems with feedback is that measurement tools available fall back upon simple and measurable things; almost always money. Take for example two more contemporary individuals; Bernie Madoff and Dr. Midei. In the case of the former, he was clearly a scam artist who thought he could get away with fraud forever. What people focused on were Madoff's rates of return and they assumed he was a genius. As it turned out he was a crook. Madoff's feedback was almost all positive until it was not. One must think he woudl have some sense that he could not keep this up forever. However, I suspect that he did not have a good sense of just how long a human lifetime is and that it is longer than he could keep dancing and deceiving.

In the case of the latter (Dr. Midei) , it is more nuanced but provides perhaps an even better justification for improved feedback loops in medicine. Dr. Midei was a celebrated interventionalist cardiologist until he fell afoul of Medicare. I previously blogged on him (Radio-fence-principle).  In an excellent article in Forbes (Mark Midei Can't Get a Job Taking Blood Pressure At A Walmart - Larry Husten), one gets perhaps a more sympathetic view of Dr. Midei. The article touches upon the peer review process, which was actually set up by Midei at St. Joseph's Hospital. Here was a process by which physicians like Midei could garner feedback.

I asked Midei about stories I had heard from a cardiologist who had attended the conferences who said “he could never see the lesions before a stent went down the coronary artery.” Midei responded that it was “regrettable that that comment wasn’t raised at the conference, that’s what the conference is for. There were many times when controversial issues were raised.”

I asked Midei about the issue in another way, if he understood that it may have been extremely difficult, given Midei’s power and position, for colleagues to criticize his work. ”I get it and I know how you get around it, which is randomized, blinded review.” But, he pointed out, prior to his case this type of review was never performed. “Nobody else did it at the time.” He added: “they all do it now.”

The article touches upon the broader environment which Midei was embedded in. The author goes on to note:

I asked Midei whether, in response to the overwhelming amount of positive feedback he had received over many years, from his colleagues, the hospital, and the medical community, it was possible that he might have developed the feeling that, godlike, he could do no wrong, and that almost anyone could benefit from his procedures. But, he said, “they’re not accusing me of pushing the bounds of intervention, they’re accusing me of exploitation of patients. But that’s not me, that’s just not in my DNA.”

With this answer I’m not sure Midei really responded to my question, but he did answer it somewhat obliquely a moment later, agreeing that the standard of practice had changed over time: “I know what the standard is now, and it’s different from what it was then. Even in 2009 after the appropriateness criteria were published my practice had changed.”
At this point he made an important concession: “so it’s impossible for me to look back at some of these cases and say” that they would all be justified today. “It’s possible that things look different today than they did then.”

But Midei insists on a key point: he is very confident that his standards were “no different” than the standards at St. Joseph’s and other hospitals. 

While the excuse that "everyone else was doing this" should not be an excuse, it does underscore the broader feedback loop environment sending signals to individuals such as Midei. The signals he received were essentially all positive until they were not.  Dan Areily's work underscores a basic human tendency to stretch the truth and we have to understand that there exists a spectrum of individuals. We all have an element of con artist in us. For us to thrive, we need environments which provide us with consistent and timely feedback when we have pushed the envelope. It may seem painful at the time but it does not approach the pain we will experience in its absence.

More on cancer screening futility

For every complex problem there is a simple solution... and it is wrong. - H. L. Mencken

Dr. Bloodgood had a simple idea. Identify cancer in its early stage and cut it out. As it turns out, this appears to work well for certain cancers such as cervical cancer. However, this observation does not appear to be particularly generalizable. It has taken almost a century and many attempts before the reality of the futility has begun to sink in.  In today's New York Times Magazine there is an article published by Peggy Orenstein, a breast cancer survivor who writes of her medical and intellectual journey "Our feel good war on breast cancer".  It is a well written and reasoned article, written by a sophisticated patient from a patient perspective. That journey has led her to the following conclusions:

"It has been four decades since the former first lady Betty Ford went public with her breast-cancer diagnosis, shattering the stigma of the disease. It has been three decades since the founding of Komen. Two decades since the introduction of the pink ribbon. Yet all that well-meaning awareness has ultimately made women less conscious of the facts: obscuring the limits of screening, conflating risk with disease, compromising our decisions about health care, celebrating “cancer survivors” who may have never required treating. And ultimately, it has come at the expense of those whose lives are most at risk."

I am not sure that it will have any immediate impact. These concepts have been around for more than a decade and have failed to gain traction. I am biased in that I think they are correct but perhaps I am wrong. I doubt it. Dr. Bert Kramer wrote in his article in the Annual Review of Medicine entitled: Cancer Screening: The Clash of Science and Intuition^{* (}Annual Review of Medicine - Vol. 60: 125-137 (Volume publication date February 2009) 

As Roman playwright Terence noted, “One easily believes what one earnestly hopes for.” 

Science, medicine, politics, and belief all converge. The screening industry is embedded into the business model for medicine and has strong allies among non-profits. It is not limited to breast cancer and it is a powerful model for public involvement. Our intuition creates the state where the default is to believe that screening should and does work and there a huge financial incentives to keep this model in place. 

Attempts to disseminate information which is contrary to our intuition will be pounced upon by believers as to be motivated by less than virtuous motives, particularly at a time where health care becomes more and more resource constrained.  However, it is even more important that this be sorted out in a time of increasing scarcity. If this a waste of time and resources can be devoted with better return elsewhere, it is stupid to continue to throw good time and money toward activities that provide no real value. I have few doubts that we will continue to throw "public" resources in the form of insurance dollars at this but I can't help but wonder if individuals would spend their own precious personal resources if they were aware of the likely return (or lack thereof) on their investment. 

   

A contrarian view of medication compliance II

I previously blogged about my recent patient encounter which a patient who had been less than entirely compliant with her medication for her chronic and uncomfortable condition. In this blog, I want to expand upon my thoughts regarding patients and medication compliance. In the previous piece, the patients I wrote about had every reason to want to take their treatment. They were uncomfortable and sought relief from their discomfort. They had a current problem which gave them little or no opportunity to forget they had a problem.

However, when we prescribe medications, we often do so to prevent some sort of symptomatic and obvious problem that MIGHT happen at some undefined point in the future. We therefore make recommendations to patients who might feel absolutely fine in the present. They are by most measures and definitions well people. We may be selling them on something they neither need nor want.

 I can't help but think of Clayton Christensen' observation in the "Innovator's prescription" where he studied patrons of a fast food restaurant who purchased milk shakes on their way to work. When the restaurant tried to increase sales, they did so by offering more variety. It did not work. The customers wanted convenience, not variety,  and sales went up when a self serve station was deployed. His take on this scenario was that it is difficult to sell people a product that they simply don't desire.

In that same light, when we treat asymptomatic states and make every effort to get people to undertake activities from which they gain no immediate (and perhaps no long term) benefit(s), are we not trying to sell them a product which, if adequately informed, might be something they do not want? This is particularly an issue if our desire to push therapies goes beyond evidence that these interventions provide actual benefits to those treated.

 Take for example treatment of moderate hypertension. Quality metrics and physician bonuses may be tied to hitting specific milestones of hypertension control in managed populations. Even under the best of circumstances, one needs to understand that there will be hundreds if not thousands of patients treated for every one who might benefit. In fact, recent data would suggest that for modest hypertension, there may be no benefit whatsoever. (http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD006742.pub2/abstract)  Given the vagaries of the data, one would anticipate that patients should be ambivalent about buying a product that likely will provide them with no person benefit. This may only get worse as our strident pleas to address critical issues are followed up my "never minds" which come as a consequence more complete assessments..

The automobile dealer from which I bought my care contacts me on a regular basis trying to get me back in their shop to have regular maintenance done. They are vested in getting me back there and I am sure they can make the case that it is important from a safety perspective to have my brakes and tires checked and to make sure that no critical part will fail at a critical time. Manufacturers have installed various warning lights to create a sense of urgency to prompt us to bring our vehicles in, even if they appear to be functioning fine. These lights send the signal that something bad might happen unless we act. We and most of the public understand they are in business and their actions are driven by their desire to maintain cash flow and a healthy bottom line. Individual vehicle owners may benefit from their efforts to drum up business as well the the only certain benefits is that obtained by the dealership.

There is a point where our actions and recommendations the health care industry as to how patients should behave and adopt our recommendations will be viewed by the general public in the same way as how we view our car dealers calling us and telling we need to have work done. When we make a recommendation for a single patient or groups of patients that a specific treatment is desirable, how likely is it that an individual patient will benefit from the specific intervention? For the most part we don't know and if we do not know, we should be careful about being particularly strident about adoption of our recommendations or critical of those who fail to heed our warnings.

A contrarian view of medication compliance

I saw a patient back yesterday who has suffered from a long standing and uncomfortable but not life threatening condition. It was very gratifying in that she was markedly better. However, the road to arriving at this more desirable place was a rather circuitous one. I had been treating her for more than one year before I became aware that she was less than fully compliant with my recommended treatments.

This was to me a a rather surprising discovery which I was fortunate to have the tools on hand to define. I was able to see when she had filled her prescriptions and to measure drug levels. It was all the more surprising since what I was treating was primarily symptoms driven. I tend to believe that this is a circumstance where compliance tends to be very high. When one is constantly reminded of the reason for treatment, one tends to remember it is important to comply with treatment.

I know this first hand. I am very bad with medications. I was started on a lipid lowering agent a number of years ago and I know that I was very bad about taking it every day. However, more recently I have developed a bit of prostatism and an oral agent to treat  has been instituted. If I don't take this, I get a relatively quick reminder that I failed to be compliant. I think my compliance with all my daily medications has improved.

However, what about my patient who mysteriously failed to take her medication directed toward her very uncomfortable state? Why in the world was she less than faithful about taking her medication? As it turned out, the treatment for her condition was very slow in onset, requiring weeks of consistent compliance to reap the benefits of treatment. While I was confident that treatment would work (based upon my previous experience), she remained skeptical, and the skepticism only deepened after no dramatic effects were seen after a few weeks. She failed to voice that skepticism, at least immediately, and the conversation only got around to this after I become aware of her poor compliance with treatment. She did not feel comfortable with voicing her skepticism but also did not desire to look for help elsewhere.

The lesson I cam away with from this encounter was not that my patient was stupid or unreliable. It was that people are complex. Our brief encounters with them rarely allow us to begin to understand their motivations and fears. The lack of understanding cuts both ways. We are essentially strangers to our patients. They do not know who we are, what motivates us, and why we make the recommendations that we do. There has to be a huge element of trust for them to adopt our recommendations and it begs the question; Why should the individual in front of us at any given moment trust us? What have we done to gain their trust?

Why the academic and research world needs a healthy growing economy;

We have spent a great deal of time recently discussing the impact of budget sequestration. Academic medicine is remarkably dependent upon Federal dollars, whether it be through grants or Federal health care dollars, and the effects of Federal cuts are significant in the short term. However, whatever effects we might need to deal with now are likely trivial compared to the effects we will see if we continue with our present business model.

I previously wrote about the crowding out effect of entitlements spending on biomedical research, the latter being squeezed as part of the shrinking discretionary pool of dollars.(Entitlements and Biomedical Research)
What I wrote back then stills holds now, perhaps even more so. If academic medicine and university based research is going to be heavily dependent upon federal research support, being aware of and supportive of policies which are conducive of a growing economy and tax base is essential for any long term prospects. Presently, that is not the case.

We need only to look across the Atlantic Ocean to see what happens when demographics and uncontrolled entitlements place a squeeze on economies. They have attempted to build states with rich social support packages upon the backbone of shrinking and aging populations and hope to address the growth and revenue shortfalls with higher taxes and creative debt. They have been able to hold on as long as they have since they have relied on the US military shield and they have borrowed recklessly. Western Europe has become militarily irrelevant and it is on the way to becoming economically irrelevant.

Europe continues to trail behind the US in terms of support of research funding.
(EMRC statement on medical research in the EU)

Both in absolute and relative (to GDP) terms, the US spends more on health care (about 6400 € at purchasing power parity exchange rates). Moreover, about 50% of all public US money for research is going to medical research and that translates to their public spending for biomedical research being about 3.5- fold higher per capita ( 143 € for 2009) than is the case in Europe.

And it is only going to get worse as the wealth generating machinery of Europe becomes consumed by its debt and entitlement supporting functions. To have an enduring and robust bio-medical research presence, we need a growing economy and manage entitlement growth in the long term. Sequester is impactful but it is just a small taste of what we face if we do not get the diagnosis right.

It is just too complex to be managed manually

I spoke to my daughter tonight. She is in nursing school and has almost finished up her first clinical rotation on a floor dealing with gunshot wounds and abdominal surgery. What she has been surprised by is how chaotic the environment is. After the rigorous didactic prelude, clinical care does not quite jive with any of the  pre-clinical work. There are too many tasks for any given nurse to handle. There are huge opportunities for error. No one seems to be in charge.

I have made the same observations and during my 30+ years of dwelling in a variety of clinical contexts, I see the chaos only becoming more obvious. Medicine has become much more complex during those 30 years. Our options have expanded and so has our ambitions. With each additional choice the complexity grows not in an additive fashion but in a geometric fashion.

Atul Gawande has written about this, focusing on acute care and hospital based settings. However, this explosion in complexity is every bit as evident in the ambulatory environment. In some sense it may be worse since most of our interventions happen in a much less controlled environment which provides less consistent feedback for patients and providers alike.

I am reminded of a quote from Thomas Sowell in his book "The Quest for Cosmic Justice", where he notes:

Unfortunately, the inherent scarcity of resources means that all of the evils that are preventable seriatum are not preventable simultaneously

Each of us tries to practice in a seriatum world and by the nature of how practice has evolved, we make every attempt to focus on pathology in our specific arena of specialty. It is our way to make our practice world manageable. However it tends to create the illusion that  increasingly complex approaches are desirable since their deployments appear to be possible, at least in the short term and when viewed in isolation. Those deploying them in isolation do not see the web of interconnected and complex problems, but only the simple series of events which they feel obligated to address in order to make a particular intervention. Elements or occurrences which may develop outside of this are viewed as being "not my job".

This leads to even more complexity which drives those practicing into greater focus on even narrowing areas of expertise and ownership of responsibility. It is a cycle which results in greater and greater complexity and less and less integration of care.

I am reminded to a host of other industries where control of complex systems have been moved out of human hands. Airline pilots increasingly are managers of computer systems with fly airplanes. Manufacturing has moved from human controls to computer controls as well as use of robots to replace human labor. In each of these environments it was recognized that the human mind is not the ideal tool to manage huge amounts of information.

In health care, we have not formally come to recognize this but informally those involved have recognized this long ago. Individually, practitioners have realized this and their individual responses to being placed in unmanageable situations has resulted in improvement of their individual circumstances and worsening of the care delivery system in general. Their individual practice lives become more manageable as the general practice environment becomes more complex and chaotic. No villains here; only the wrong incentives.

Playing Where's Waldo with the medical record

There is nothing worse than trying to function and make decisions without access to relevant information. Sometimes that takes the form of a sick patient showing up on your doorstep without the benefit of any reliable medical history. Almost as bad (and perhaps sometimes worse) is someone showing up with the equivalent of War and Peace volume of non-indexed records and an expectation that I review them on the fly and render the definitive opinion within a few short minutes. Did I mention that they had leveraged their life savings to travel from half way across the galaxy to see me only to be scheduled in that generous 15 minute slot?

Yes preparation and information are critical for being in a position to make rational and likely helpful recommendations to patients. Thank God we are on top of this with our electronic medical record with ALL of the relevant information at my beckon call. Not so fast....

Perhaps the only thing worse than no information is a sea of information, and no real way to effectively search and identify what is important and relevant. I was overseeing a resident clinic today and saw a patient with a confusing presentation which perhaps suggested some sort of multi-system disorder which could heart, skin, kidney, or GI tract. We sent searching for relevant studies, first focusing on a cardiac echo. Would I find it under reports? imaging? cardiac? Did he have it here or was it an outside study and could it be one of the sea of scanned documents organized chronologically by when someone got around to scanning it? Is there any sort of reasoning behind how they are labeled or titled? Who am I kidding?

As I am clicking on every imaginable tab and folder, I have a flashback to when my children were younger. We used to read books of the "Where's Waldo" series, particularly at bedtime. They would chose the WW books because they knew they could defer the lights out hour almost indefinitely because it would take forever to find Waldo in the sea of figures he was hiding in.

That is where we are with the EHR. We have created the Where's Waldo equivalent. No real index. No real search function. Just rare morsels of critical information required to make good decisions hidden in a sea of useless distracting medical noise. Where's Waldo?????

Sharing as a form of control

https://www.youtube.com/watch?v=WbuIl6phdco

This where clinical research MUST go.

The problems with not knowing

Ben Goldacre presented a TED talk entitled "What doctors don't know about the drugs they prescribe"
https://www.youtube.com/watch?v=RKmxL8VYy0M. The lesson is relatively simple. Due to basic flaws in the structure of how we deliver care, evaluate therapies, and distribute information, we cannot trust the information we work with in medicine.