While Barack Obama’s health care proposals have been the best of all the major candidates this year, they are not without fault. I’ve been critical before of estimates of cost savings which will pay for expanding health care. FactCheck.org presents data suggesting that Obama is overly optimistic about savings from electronic health records. While Obama’s campaign might be sincere in their predictions, after having seen many of the problems with electronic records I fear they underestimate the difficulties in initiating this as well as over-estimating the benefits.
As the RAND study quoted in the post states, it could take until 2019 for 90% of doctors and hospital to have electronic records. Even this estimate might be overly optimistic considering the weakness in the electronic record systems now available. Even if there were tremendous incentives offered which got every doctor to go electronic despite the current ridiculously high costs there is no guarantee we would save much money.
One problem is in communication. Currently different doctors, hospitals, and labs might have computerized systems but it is hit or miss as to whether they can communicate with each other effectively. Factcheck cites a Congressional Budget Office report which states there are forty different vendors and their products cannot communicate with each other. Plus, all forty vendors over-charge for their products and each seems to have serious flaws.
Sometimes the volume of material turns out to be counterproductive. Often in computerized medical records tons of material looks exactly the same, regardless of significance. It is often difficult to find the important information among all the material which is accumulated over the years. In contrast, an old fashioned paper chart with the important reports at the front, often with pertinent sections underlined, often provides the important information more easily.
The systems available today have numerous weaknesses. For example, today I had a diabetic patient come in the hospital who was also in the hospital last month. Neither I or the nurses on the floor were able to get the hospital’s electronic medical record system to reveal the patient’s insulin dose from the last hospitalization. In the old days we could have had medical records send up the written chart and then we’d find the information very quickly. I had a similar situation with another patient a couple of weeks ago in which information from a previous hospitalization could not be easily obtained. In both cases I wound up turning to my charts at the office. Even there we have a dual system–both a computerized system which is used for sending prescriptions electronically to pharmacies and a paper system. For a number of reasons, my paper system is far more reliable in telling me exactly what medications and doses were used. Perhaps that is why I generally have such difficulty evaluating the medications when I receive copies of electronic records on new patients.
I’m not even convinced that using a computerized system to send prescriptions to pharmacies either reduces errors or lowers the risk of medication reactions even though both are cited as advantages. If I’m preparing a prescription there is always the chance for human error but most likely the prescription will say what I intend. I’ve been using a simple system with a word processor and macros to quickly generate printed prescriptions for years, already eliminating the old problems with handwriting.
With staff going through the charts to enter the old data into the system, as well as responding to electronic requests for refills, there has been tremendous room for error. I’ve caught a number of errors from information entered incorrectly, and I fear that when sending prescriptions is as easy as pushing the send button there is a much higher chance that errors will go unnoticed. The software does supposedly screen for medication reactions, but it is far too flawed to be of value. Interactions between medications are commonly considered in writing prescriptions, often with such interactions actually being desirable, and when every conceivable interaction generates a warning message it becomes likely all will be ignored.
Perhaps over time the state of computerized medical records will improve, but we are not likely to see them either implemented or bring about real cost savings in the near future.