A close friend of mine, Tom, has been diagnosed with the rare genetic condition alpha-1 antitrypsin deficiency. Its progress now requires him to seek a liver transplant. Because Tom isn't in critical condition, he is low on the transplant donor waiting list. He can, however, take advantage of the living donor liver transplant technique, which uses a portion of a compatible liver. A number of family members and friends have volunteered to be tested to see if their livers are suitable, so he's hoping that sometime soon he will be able to get a new liver.
As you might expect, before you can get a transplant, you need to pass a battery of tests to see whether you can withstand the surgery. Tom recently updated me on his latest screening trials and tribulations.
Not long ago, Tom was required to undergo a complete stress test, so he trundled off to the local hospital, where a technician performed the test and then proceeded to send the results off to a cardiologist to be assessed. Tom told me that after a couple weeks, his tests came back and showed that he was in good shape for his age. But there was something about them that raised some unspecified concerns. Not sure how much to be worried, Tom scheduled a visit to the cardiologist to see if he could get the issue resolved, since by this time the cardiologist's approval was now the only thing standing in the way of Tom moving on to the donor matching process.
As it turns out, the cardiologist didn't know that Tom was in need of a liver transplant or that he has had four operations in the past six months to stabilize his deteriorating medical condition. Once the cardiologist found this out (and Tom had to tell him because this information wasn't in Tom's records that had been sent to the cardiologist--the cardiologist being only "employed" just to read Tom's stress test), Tom's test results didn't look so bad after all. He related to me that similar episodes of this kind have happened several times as he has gone through the process. The relevant doctors were focused only on their specialties or domains of expertise and didn't have in their possession the full picture of Tom's medical condition.
Tom's experience is what frequently happens in medicine today. Information is captured in independent, scattered silos. There is an incomplete understanding of the problem being solved. Critical context information is missing. Important activities are outsourced. Decisions are delayed. Overhead costs explode due to the rework required to put everything back together again into a comprehensive whole.
The U.S. government's, as well as the health-care industry's, solution to Tom's problem is an initiative to automate health-care record keeping. The idea is to create an electronic health record (EHR) for every person in the United States, which can be sent to any doctor, hospital, or insurer. This electronic record would contain a patient's complete medical history, solving a number of problems. It would help eliminate the source of many medical errors (currently causing up to an estimated 100 000 deaths per year), reduce wasted time and inefficiencies, rid us of redundant or inappropriate medical care, save money (at least US $70 billion to $100 billion per year), and, most important, promise to improve overall patient health care.
While recognizing the benefits of an EHR system to his own situation, Tom is also more than a bit skeptical of all the claims of how straightforward it would be to implement. You see, Tom is one of the most senior computer scientists in the U.S. government. He has worked on large-scale, unprecedented information technology systems for his whole career. Tom knows intimately how hard it is to create, operate, and sustain large-scale IT systems, especially ones that have never before been attempted. The technical, operational and political factors--not to mention the costs and time required--that can doom these projects are routinely underestimated.
For instance, the proposed U.S. EHR system is not really one centralized system but a complex set of distributed systems. Plans are for it to be made up of an unknown but very large number of independently run and managed interconnected national health information networks linking heterogeneous IT systems that reside at the 50 000 or more doctors' offices and hospitals across the nation, each able to communicate seamlessly and securely with one another 24/7.
Interestingly, while technical creation, operation, and management of such a system is a huge undertaking, Tom is more concerned about all the ”little things” that can cause it to place a patient like him at increased risk. For example, the medical information in an EHR system needs to be absolutely trustworthy--complete and correct. What if incomplete or incorrect information is put into the system, which will undoubtedly happen, given that current paper-based medical records have to be converted into electronic medical records? Who is responsible for correcting errors? Who is responsible if incorrect information causes a patient harm?
Tom is also skeptical that having more information necessarily means that better medical decisions are automatically going to get made. When doctors' earnings are related to the number of patients they see, there is a trade-off between better care that electronic records provide and the sheer amount of time required to review a more complete medical record. How often, Tom asks, have you been to a doctor who clearly hasn't read the form you filled out detailing the pain you are currently in, let alone your family history?
Tom doesn't believe that electronic health-care records will keep doctors from ordering their own tests to double-check what other doctors have already tested for either, a supposed benefit of EHRs. In Tom's experience, doctors feel most comfortable with the results their tests provide rather than the results from tests other doctors ordered. Worries about malpractice claims also will likely continue to spawn a number of these duplicative tests. To change this behavior requires two culture changes. The first is in current medical practice, while the second change has to be to society's penchant to pursue malpractice litigation. Either of these culture changes will be extremely hard to accomplish.
There are lots of other issues in creating an EHR system that keep Tom amused, as well. He asks what the probability is that the proposed EHR system of 2006 will look like the EHR system of 2015, the year of the hoped-for operation of a national system. Think back to 1995--how many times has the current underlying EHR technology or thoughts about how to implement EHRs changed during this time? How will future technological change be accommodated? Also, how will "requirements creep" be controlled? Already the idea of what a basic EHR system should look like is morphing rapidly, as everyone from politicians to doctors to patient advocates lobby for their visions of what the system ought to do. Consider how many more "good ideas" will be added to it over the course of the next ten years. To Tom, a national EHR system looks like a very good idea that is likely to die a death of a thousand cuts.
Just before I said goodbye to Tom, he told me his biggest worry is time. The screening process has taken so long that he is worried his disease will trigger some other health problem, which will force him to start the screening process all over again. He mused that the EHR system implementation has the same problem.
[Editor's Note: When this article was about to be published, the author informed us that his friend Tom recently underwent successful liver transplant surgery. He is recuperating and undergoing follow-up therapy.]
About the Author
Robert N. Charette is president of ITABHI Corp., a risk-management consultancy in Spotsylvania, Va. An IEEE member, he is the author of several books on risk management and chair of the ISO/IEEE committee revising the 16085 standard on software and systems engineering risk management.