The medical record serves two basic purposes: It provides the treating physician with a reminder as to what information has been gathered, what treatment has been provided and what response the patient has had to the treatment; and it provides that physician with the ability to share such information with other parties as desired by the patient. Because each individual patient has a history of data points within each of these topics that can vary widely, any attempt at categorizing or minimizing, even within a diagnostic framework, results in the loss of information.
I can tell you, for example, that I have 10 patients with major depressive disorder, each treated with sertraline, and each doing reasonably well. That information (plus just a little more) might be sufficient for payment records, quality control, standards of care requirements and liability control, but is quite incomplete when it comes to a true assessment of this patient in the future by either me or by another clinician.
Many electronic medical records focus on the requirements of today's world, such as quality and liability, without focusing on what a medical record is actually supposed to be. They often provide a space for the doctor to freely enter text, but by the time all the check boxes have been checked, menus pulled down and precise selections made for diagnostic entries, there is little time left for the doctor to enter any text at all. In my role as a medical reviewer, where I review medical records for various clients to assist in the determination of diagnosis and disability, I have found EMR-based reports to be, for the most part, worthless. They often have information that is cut-and-pasted rather than written contemporaneously, diagnoses that are based on symptoms which are no longer present, and medication lists that reflect medications actually discontinued long ago according to the pharmacy records. There is no evidence to suggest that error rates will decrease with a transition to EMRs, and evidence exists to suggest that current error types will merely shift to new error types.
In my own field of psychiatry, I've seen that physicians are reluctant to enter the often highly confidential information about their patients that they would have written in their own records without a second thought in the past. The result, as one colleague put it, is that all the electronic charts come in a single flavor: vanilla. And with that, all the useful data with respect to actual medical care are lost.
We all know from our own home computer experience the type of upkeep required: software must be kept up to date, data must be backed up, the entire system replaced every few years when a new operating system requires new hardware. All of this requires attention. In the office environment, that means significant expense where there was once none. Given how many physicians are practicing under considerable debt load from medical school, how reimbursement is failing to keep up with costs, and how liability rates are ever-increasing, the likelihood that physicians even have tens of thousands of dollars to spend on initial and ongoing hardware and software costs with no evidence of improved efficiency or efficacy is nil.
I'm not opposed to EMRs, and indeed spent some of my career working at Apple and various dot-coms, but the time is not yet right to implement the large-scale projects envisioned by hungry businesses eager to enter the medical marketplace.
Stuart Gitlow, M.D.Executive directorAnnenberg Physician Training Program in Addictive DiseaseNew York To submit a letter to YOUR VIEWS, click here. Please include your name, title and hometown.