Oh behave …

I recently read an article that speaks about physicians’ behavior and how to change it. The author of the article specifically discusses changing doctors’ behavior in relation to new software and technologies. The conclusion of the article is that you need to clearly demonstrate a benefit of the software to the doctors. If you can show them that it saves time and/or money, or that it streamlines their daily work, then the doctors will be far more accepting. This requires formal training and to some degree, evangelism. In this case, evangelism refers to finding key people who are major supporters of the new system. These people then share their enthusiasm with their colleagues and as such, get them excited about the software’s new features.

I appreciate the context in which this article was written. The author also consulted with a very senior individual who deals with the design, planning and implementation of technology projects. So, the author really made an effort to find valuable information about how to deal with this problem.

My problem with the conclusion of this article is that doctors are notorious for resisting changes in behavior. There are a significant number of medical Journal articles that speak to the importance of washing your hands. As obvious and trivial and expected as this sounds, doctors have a very low compliance rate with washing their hands. When handwashing is enforced, infection rates in various medical departments in hospitals, drop dramatically. In other words, many patients could be spared prolonged admissions due to infections if doctors simply washed their hands before seeing their patients.

I once heard a formal discussion on this topic at a conference on quality of care. The speaker spoke of varied ways to encourage doctors to wash their hands. I was stunned. How can it be that you need to coerce doctors into following one of the most basic practices. But, you cannot deny the research. One way to “encourage” doctors to wash their hands, was to put cameras over the sinks in the various departments. The moment the doctors felt that they were being watched, suddenly the handwashing rate went from a dismal low level to over 85%. If this sounds like children in a kindergarten, what can I say.

In the company I previously worked at [for 21 years], I spent 11 years developing and implementing my own electronic medical record. I am extremely fortunate that I studied computer science before medicine, and I have decades of experience in both programming and clinical medicine. Over the last few years, I made multiple efforts to change certain behaviors of the physicians who worked in the company. I should point out that this was a service that set up and ran community-based emergency rooms. At the time I left, there were 12 such clinics, seeing well over half a million patients a year.

What I discovered was that emails that explained new features and even one to one personal requests, had little effect on physician behavior. At one point, I built a web interface that allowed me to review high-risk cases remotely. Every day, I would review 100 such cases and I would email comments to the treating physicians, in the event that there was room for improvement in the care provided. Over time, the physicians internalized the fact that they were constantly being watched. And it clearly affected the way in which they worked. At times I would even see notes left for me, via the EMR interface, with the doctor’s expectation that I would see the note via my regular review.

In other circumstances, when it was necessary for the doctors to follow a specific guideline and order certain tests, once again, compliance was poor. The way in which this was dealt with was through the software. The software was updated to look for certain patterns in the clinical picture of the patient, and on this basis to pop up a window that demanded an explanation for why a specific test was not ordered. So for a patient with upper abdominal pain and high blood pressure, if the doctor did not order an ECG, the software would require an explanation for the lack of this test. In some cases, with very high risk patients, I would even get an SMS indicating that a high risk patient was being treated, possibly, inappropriately. Over a very short time, the doctors fell in line and started ordering the ECGs in order to avoid the pop-up. I had the same experience with getting the doctors to order pregnancy tests, when indicated according to protocol. At the point that I left the company, there was a whole range of such pop-ups constantly monitoring the recorded data about the patient, and popping up warnings as needed.

There is no question that via the software, I and other senior medical staff effectively infantilized many of the doctors working in the clinics. A significant part of the  EMR became code that monitored physician behavior. The results of this experience were presented multiple times at various conferences in Israel. I personally definitely became an evangelist for developing systems that removed the freedom from doctors to  ignore standard of care practices. There were occasions where doctors became offended by the degree of monitoring. And they openly expressed their dissatisfaction with having the pop-ups constantly forcing them to follow guidelines. Each of these doctors was told in turn that there was no justifiable reason to ignore guidelines, but if there was, they had the opportunity to explain themselves through the software. Over the years,two physicians quit due to this degree of oversight. Personally, I was happy to see them go, as they clearly felt that they were smarter than official guidelines from the major medical associations.

As artificial intelligence becomes smarter [and it definitely is becoming smarter], it will be possible to have more and more decision-making become automated. Even in medicine, especially when there is a detailed guideline for how to manage a specific condition, the computer will almost invariably provide better care than a human, who can forget or may simply refuse to follow the protocol. The truth is, that doctors should not need to deal with cases where the care for the patient follows a research proven guideline. The doctor can definitely be involved when there is a complication, or to explain the whole process to the patient. Ideally, the computer will free the doctor to have more time to develop the doctor-patient relationship, even in an emergency room environment.

This reality frightens a significant number of physicians. They insist that computers will never replace them. At least for the next couple of decades, the intent is not to replace the physician but to assist him or her in their daily work. Anything that can be standardized can be translated into a computerized protocol. Especially with the assistance of alternate medical professionals, like physician assistants and nurse practitioners, smarter computers can guide a great deal of most people’s care. Once again, physicians should not need to spend time with these cases, except to be a source of information and to consider the emotional and psychological state of the patient.

Despite the often times heard statement that the computer made a mistake, in my decade of experience with my own EMR, it was extremely rare that the problem was the software. I remember even one case where data had been deleted and the secretarial staff absolutely insisted that no one was responsible. They insisted that the computer had simply lost the data. Unbeknownst to the secretaries, I had programmed in an internal log of each transaction. And lo and behold, I found that the data had been actively deleted using the access code of one of the senior secretaries. This secretary absolutely denied that she had actively deleted the data. The secretaries also knew that they were not allowed to share their codes. So either she deleted the data or she gave her code to another person who then used it to delete the data.  Not surprisingly, she also insisted that she had not shared her code with anyone.

So here is a case where despite clear recorded evidence of inappropriate behavior, the employee insisted that somehow she was not responsible. I have had similar experiences with the physicians who used my EMR. The simple lesson is that humans are fallible and will even lie to protect themselves. Software that has been tested and used for years, will not fail. While there is always a chance of a bug presenting itself even after a decade of use, computer error is extremely rare. Computers do not forget, and they do not get tired. Maybe, it’s no wonder that  they frighten some physicians.

Now, I will ask a question. How do you imagine medicine being practiced in 50 years, when artificial intelligence and robotics will definitely be advanced enough to replace almost all physicians? Will you personally insist on a human caring for you, when the research evidence is that computers are faster and more accurate than human doctors, in diagnosing and choosing treatment for a patient? I have heard the laughter of many a doctor, when I present this scenario to them. Let them laugh. This will be the new reality sooner rather than later. I personally would rather prepare myself for such a reality [by expanding my skill set and looking for other ways to contribute to better health care] than be left behind. What do you think?