Monday, November 30, 2015

Information Systems, Good Medicine and the Search for Black Swans

Many years ago, when I was in medical school, learning the art of the physical exam, I was taught that the exam should always be done in the same order each time, starting with the skin, moving to the head and then working your way down, leaving the rectal exam and the exam of the genitalia for last (more for the comfort and the dignity of the patient).  It was drilled into me that it must always be done from the right side of the bed or exam table.  That consistency and discipline in performing the physical exam the same way each time was considered a critical element of good medicine. 
While the idea of a physical exam in general may be seen quaint and outdated in our era of telemedicine, scans, apps that can monitor bodily functions and other imaging technology, the discipline of the physical exam and even the emotional impact of that exam is hardly outdated.  Abraham Verghese, the esteemed author, physician and educator has written and spoken about the importance of the physical exam in both a ted med talk and in the medical literature and his words are worth reviewing.  Verghese emphasizes the ritual part of the physical exam, and the importance it holds in bonding the physician and the patient.  He writes,

“…….the ritual of the bedside examination involves two people meeting in a special place (the hospital or clinic) wearing ritualized garments (patient gowns and white coats for the doctors) and with ritualized instruments, and most importantly, the patient undresses and allows the doctor to touch them.  Disrobing and touching in any other context would be assault, but not part of this ritual which dates back to antiquity.”
He goes on to say,

“We propose that if the ritual is short changed, if it is done in a cursory fashion, if it is not done with skill and consideration, if its sacredness seems to be violated, then the transformation (which in this case is the formation of the doctor-patient bond, the beginning of a therapeutic partnership and the healing process) does not take place.”

His words reflect the way a good physical exam helps form the basis for the doctor-patient relationship.  However that is not what I will choose to dwell on in this post.  Rather I want to discuss the ritual itself.  The very rote and disciplined way the examination is done each and every time a physician, especially an internist or family physician sees a patient. 

Back in medical school, I was taught to perform the exam very precisely each time for a simple reason.  I was told that if I did the exam the same way each time, it would make it more likely for me to notice anomalies.  I would notice something that was not like the thousands of other exams I had done.  In many ways, medicine is extremely repetitive.  Because of this repetitive nature of practice, it is very easy for physicians (and nurse practitioners) to go on auto pilot – to become reflexive in their actions instead of remaining cognitive with every patient and through every exam.  At the same time, high quality medicine is all about finding the needle in the haystack: noticing the unusual exam finding that might suggest an illness that could turn into tragedy.  In order to find that “black swan” (defined as an unpredictable or unforeseen event or diagnosis, typically with extreme consequences) one must sift through thousands of exams that are very usual and very common.  One must feel a lot of normal livers and listen to many normal hearts to find the one that is indicative of a disease.  It is so easy not to notice!  It is so easy to take shortcuts and not go through the discipline of the exam. 
The result can be a missed diagnosis and ironically a more expensive evaluation, as unfocused diagnostic evaluation often called shotgun medicine, requires numerous tests which are unnecessary if instead the exam is thorough.  An example is the 50 year old patient who comes in with chest pain.  On the basis of that history and a lax approach to the physical exam a cardiac evaluation would be done which could include stress tests, and even cardiac catheterization.   If one however does a physical exam and notes that the pain is in one dermatome (an area of the skin supplied by nerves from one spinal root) and that small vesicles can be seen, the diagnosis of shingles is made and no further evaluation is needed. 

I sat once on an airplane next to a psychologist.  He told me that he had been an airline pilot and had developed a psychological test to see if prospective pilots would be good in the cockpit. He left the pilot’s seat to develop this test as he felt his ability to improve airline safety would be greater in promoting widespread use of the test than in his flying planes.  The key to being a good pilot was to be highly intelligent but not too intellectual.  You had to do the same things every time you took off and landed.  On each take-off, you had to make a decision as to whether to abort the take off and it had to be a conscious decision.  You could not allow your mind to wander (in his definition an intellectual is one whose mind does wander) or have the take-off become so reflexive that you don’t notice small differences from the normal take-off.  It seemed to me that the similarities to medicine, especially primary care medicine, were apparent and so it also seemed to Dr. Atul Gawande.
Atul Gawande noticed the use of checklists in the cockpit of an airplane and it led to his writing “The Checklist Manifesto” which advocated for checklists in medicine, similar to the checklists in aeronautics.  In a real sense, Dr. Gawande was only harkening back to the medical school truism of finding ways to do repetitive tasks, such as the physical exam, in such a way that that steps in the processes were not missed and  abnormalities were more easily noticed.  Checklists are a way to highlight deviations from the norm and to ensure that we do everything in medicine in a disciplined way.

However, as we started to computerize our routines, and started to monitor them using modern information technology including EHRs and other electronic tools, we may have lost our way a bit.  The goal of routine leading to more easily noticing deviations from the norm took second place to pure efficiency and blind adherence to protocol.  The idea of the best doctors being those who were complete even compulsive and who found the black swans was replaced by the idea that the best doctor was the one who could prove adherence by having all the computer fields filled and prove efficiency by doing so most quickly. 

Computers allow us to copy and paste and pre-populate fields for the sake of efficiency.  If we do the same exam each time and nine times out of ten the exam is normal, doesn’t it make sense to just prepopulate the normal exam and doctors can then change the results as needed?  However having the areas prepopulated does collide with one of the secrets of medicine, namely that doctors are people and they work towards incentives and towards the path of least resistance as other people do.  When you have a computer record field that is prepopulated, and you don’t think something will be abnormal, and you have a waiting room full of patients, you tend, as a person, to just accept the prepopulated answer and skip the actual exam component.  By doing so, your organization (and more and more physicians are employees of larger organizations) will receive higher reimbursement for a more intensive exam because it is documented that more was done on the exam.  The physician will get a higher report card grade, leading to a better bonus because they did more complete exams or so the computer says. 

By the same logic, in the hospital setting, in which multiple physicians and nurses all having different roles and responsibilities may see a single patient, the ability to copy and paste someone else’s exam since it is likely to be the same as yours, is not seen as dishonesty or poor medicine, only as a way to be more efficient and more productive.    The fact that it removes an internal quality check and that each health professional may notice something missed by another is not factored into the development of the system.  If someone is sick enough to require hospital level care it only makes sense to have multiple checks in place to guard against mistakes that could cost someone their life. 

As an example: for one patient, the head, eyes, ears, nose and throat exam (HEENT) was presumably done because it was pre-populated.   In reality the physician skipped actually preforming that part of the exam.   The patient being seen had known Irritable Bowel Syndrome and came in with constipation and abdominal pain that was not very different from when seen during the previous visit.  The rationale was that the HEENT exam was bound to the unchanged from the previous visit.  But this time could have been different and the good doctor, would have either changed the pre-populated fields to reflect the head was not examined, jeopardizing his or her own job evaluation by the health care organization, or actually done the potentially unnecessary part of the exam, lengthening the appointment in such a way as to either not have time to focus on the problem at hand or run late in the clinic and hurt his or her service evaluation, also harming the job evaluation. 

This is not a problem with the information technology.  These tools can help us achieve and monitor the exam routines I learned more than 40 years ago.   What is missing is the proper design of these systems to support those noble goals of finding the black swans and focusing further testing and therapeutics, rather than the simpler goals of efficiency and adherence to a norm.  Missing from our currently designed systems are the aspects of human engineering that understand how patients, physicians, nurses and therapists really work and think.  We cannot afford to ignore the need for constant diligence to find something unexpected – something unusual that will make the health professional stop and think and perhaps save someone’s life.  It is not an impossible challenge.  It can be done and must be done for the promise of information technology in medicine to truly be met.  

Sunday, November 22, 2015

Technology, the Triple Aim, and the Three Laws of Robotics

The triple aim of health care (improve the patient experience, improve population health and lower per capita cost) remains the driving force in health policy since Donald Berwick, Thomas Nolan and John Whittington first described it in Health Affairs in 2008.  Information technology is now being used in ways unimagined in the past to drive the triple aim, often with less success than had been anticipated.  The challenge has been and continues to be, using information technology effectively while maintaining the aspects of medical care that require the human touch.  The real challenge is how to harness the power of the computer in what must remain a caring pursuit. 

It is clear that our quest to incorporate computers into medical care and perhaps even direct medical care, risks losing the essence of medicine – the humanism and caring tradition that should be paramount.  I have written before in an earlier post of my own experience in the hospital and the fact that I felt ignored as the staff worked diligently to address the need to answer the computer’s demands (My Recent Hospital Stay and the Care of the Computer).   Yet I know that our ability to properly use information technologies will help improve medical care.  The real question is, and will be, will the health care system effectively use technology to improve care and foster the humanism inherent in care or will the technology itself define a new system which is driven only off the zeal for efficiency and the best science of disease, but which leaves the hands-on humanistic and spiritual tradition of medicine in the dustbin of history. 

When I was in college, in the late 1960s, early 1970s, I had a friend who was getting a degree in the early field of computer engineering and programming.  He was a strange sort of guy who used to rail about the evils of computers.  He would talk about how they were actually imbued with malign intentions and evil spirits.  When asked why then he wanted to go into computer science, he would answer that someone had to control them in order to defeat the evil that was inherent in them.  In today’s world, and especially in today’s world of computers in medicine, I wonder if he wasn’t on to something. 

Certainly in popular culture the ideas of computers having an evil dimension and even dominating humans is examined both by those at the cutting edge of science and by those in the arts.  Physicist Steven Hawking has been one of the more vocal scientists who have warned of the risk that computers, via the use of artificial intelligence, could, “spell the end of the human race.”  Bill Gates and Elon Musk have also voiced their concerns in calling for more research on the potential for computers to learn to “think” for themselves and evolve themselves and where this could potentially lead including the possibility of information technology controlling human action.  As recently as June of this year, Steve Wozniak, co-founder of Apple, created media buzz by declaring that in the future, humans will be the pets of computers.   Science fiction has explored this concept for many years, with Isaac Asimov inventing the “Three Laws of Robotics” which aimed to protect mankind from the control of machines.  He then proceeded to build stories to show the inadequacy of the three laws in protecting humans and humanity.    As written by Asimov the three laws are:

“A robot may not injure a human being, or, through inaction, allow a human being to come to harm.  A robot must obey orders given it by human beings except where such orders would conflict with the First Law.  A robot must protect its own existence as long as such protection does not conflict with the First or Second Law. 

The laws all stress the need to help men and women however Asimov’s stories show the reality of unintended consequences when good intentions can cause bad even evil results.  The reality is that when human arrogance believes that control can be total over the world around us (including the technology we create) the results often can prove us wrong in painful ways.  In medical care, this may prove to be especially dangerous as the results are highly personal and could be life-threatening.  Medicine is filled with low probability, high consequence events and information technology is invariably designed for populations, rather than the black swans – the unpredictable, rare but high impact event that can radically affect a person’s life.  Our systems approaches to date have also not taken into account the values, beliefs and social structures that we all live under. 

While the three laws of robotics do not directly address the same issues as the triple aim, the idea that certain hard wired goals, or laws can address all eventualities, and all permutations, is similar.  Both the triple aim and the three laws of robotics are inherently good: however any laws developed by man that are ultimately taken as holy writ and hard wired into computer systems, can be interpreted in such a way as to create pain for individuals.  In medicine, one of the dangers of technology being programmed to address certain components of the triple aim is that while service and quality are implied in the first aim, it is not stated clearly who defines either service or quality.  More and more we have studies that show that the “system” definitions as determined by those who design and run the health care organizations are different than physician and nurse definitions which are different than patient definitions.  A certain decision algorithm, being driven by information technology, may not support the goals of population health and lower per capita cost but, due to the unusual nature of the disease and of the patient’s psychosocial situation, may help that person in need. 

David Shaywitz, one of the best thinkers in health care and a blogger for Forbes, recently wrote a short blog entitled, “First, We Devalued Doctors; Now, Technology Struggles to Replace Them” in which he describes  the challenge of trying to have technology drive personalized medicine which depends so much on knowing the psychosocial dynamics of each person being treated.  He writes, “I realized there was something that seemed a little sad about the idea of developing extensive market analytics and fancy digital engagement tools to simulate what the best doctors have done for years – deeply know their patients and suggest treatments informed by this understanding.”

I agree.  It is sad that society may be abdicating the sacred trust of knowing the person to a computer rather than to a caring professional.  But it is not too late to change the new paradigm being written.  We can effectively find a way to control the computer and use the capabilities inherent in that technology to augment the humanism of the professional helping those who are in need.  We can prevent a purely technological approach to the triple aim going the way of the three laws of robotics in literature and being the fodder for tragic stories of individual pain.  It will take new information technology and new approaches that are carefully designed to foster humanism, as defined by the patient and the family.  Caring can be improved if we learn how to use information technology in a way that supports and helps our professionals focus more on understanding each patient as a unique individual and not just a set of pathologies.    My strange friend from forty years ago had confidence that he could fight the inherent evil of computers and I too am confident that we take harness the power of technology to improve the professionals’ ability to care for patients.

Monday, November 9, 2015

Apps, Attention, and Obesity

At one point in my career, since I have a strong background in nutrition (I am a Fellow of the American College of Nutrition and was board certified in clinical nutrition) I thought I would make my fortune by writing the ultimate diet book.  It was really very simple.  On each page I would print the following message:

Of course that was many years ago and now the current version of my book is achieved by creating the ultimate weight loss app.  The many apps that have appeared are as equally effective as my book concept would have been.  Or so it seems in a very nice study this month in the journal Obesity.  The Duke team, one of the best in the country in treating obesity, investigated whether apps on a cell phone helped young obese adults lose weight.  The answer was no.  In their words, “this behavioral intervention did not lead to significant weight loss relative to control at any trial time point.” 

In some ways, this is not surprising.  Shlomo Benartzi, author of the book, “The Smarter Screen” and a professor and co-chair of the Behavioral Decision-Making Group at UCLA’s Anderson School of Management, makes the point that computer screens, or cell phone screens are “changing the way we think making us more impulsive and reactive.”  For behavior change to occur especially when it comes to overeating, becoming more impulsive is not likely to drive a successful strategy.  He notes, that “people think faster on screens, and this can lead us to become more reliant on our instinctive responses and initial impressions, even when they are misleading and incorrect.”  So while the app may tell you that you are eating too much, it is unlikely to drive behavior to make you eat less impulsively. 

I can have multiple apps on my cellphone, to manage my diet, my activity, my sleep, my heart rate, my medications, and my medical records.  Many of these apps are based on the theory that people just need more information about what they are doing and how they are doing it in order to change.  The most elegant apps (technologically) also send messages to your doctor or to a nurse if you or your numbers are out of line.  However the question is whether all these fragmented sources of personal information actually help us to modify our behaviors. 

This focus on making it simple to keep track and give us feedback on our own activities appears on face value to have some benefit.  We can potentially be able to keep records and “share” the information with the doctor or a nurse who can then intervene.  However it also can create information overload for both the person who owns the device and the health professional expected to intervene on the basis of the information.  Benartzi points out the experience from the VA Health System which has wonderful alarms within its medical records which doctors routinely ignore because there are too many alarms and too little time and attention.  It turns out that too much information – too many alarms – leads to a “scarcity of attention” – a term used by Benartzi – and that scarcity is amplified in our screen based world.

Benartzi compares our new use of apps and computer screens to Adam Smith’s treatise “The Wealth of Nations” written in the early stages of the Industrial Revolution.   He writes, “While Smith associated scarcity with a lack of material resources – during the Industrial Revolution, people needed more coal and wood and land – the most important scarcities of the information age are psychological, and caused by our new abundance of information. “ 

When I was a health consultant, advising the Fortune 100 companies on the health programs they put in place for their employees, I would often point out that they were spending huge sums of money on multiple health programs which ranged from weight loss, to stress management to disease management to pharmacy management, each with their own separate information and interventions for their employees and the employee’s families.  I would then tell them that the only thing that they were actually achieving was confusing the hell out of the same people they were trying to influence and whose behavior they were trying to change.   While at that time I admittedly did not know all the decision science behind my consulting point, it turns out that what was obvious to me has now been proved experimentally in numerous studies. 

In the world of apps, it is hard to have a full picture of the person trying to make the behavior change.  Life circumstances may make good diet and exercise virtually impossible due to competing priorities, financial and time constraints, and emotional factors.  It is virtually impossible for an app alone to address these points that are unique to each individual and that are often constantly changing. 

In the article on cell phone weight loss apps, the Duke team makes the point that behavioral change principles are often absent in the app based commercial obesity products.  Current obesity treatment guidelines recognize the need to evaluate optimal frequency and duration of contact on an individual basis which is hard to design into a pure app solution.  They conclude that you need a combined approach that may require “the scalability of mobile technology, the social support of personal coaching, adaptive intervention design and more personally tailored approaches.”

Apps and technology do have a role.  However their role is in making already effective programs and approaches scalable, easy and consistent.  They are facilitators of solutions rather than independent solutions.  Part of our challenge, is how to use these powerful tools, in conjunction with people and other support and operational processes to truly help people to lose weight, control their blood pressure, take their medications and maybe even just enjoy their lives more. 

The apps alone, in providing information, will be hard pressed to address the psychological and life factors that require the personally tailored approaches the Duke authors in the article speak about.  In many ways, our challenge for the foreseeable future is how to harness the power of technology in the search for the more individualized approaches that address people’s health, their life situations, their psyche, and even their spirit in the quest to improve their lives.