Philadelphia Reflections

Electronic Medical Record

TRS-80 Model I

When the first home computers became available in 1980, I bought a TRS-80 from Radio Shack and tried to automate the paperwork in my office. It was fun, and the idea was exciting, but after a few years I gave it up. I was willing to accept a partial electronic record, which was limited to data input that was already automatic, in order to get something useful, without making extra work for myself. Over thirty years later, an embroidered version of it was working satisfactorily in my office, but it was far from complete, and I had not been able to persuade a single colleague of mine to use it for more than a trial. To finish the story, when I retired I was advised I had to retain my records for seven years. During those seven years, I had exactly two requests for any of the thousands of records.

The reader is entitled to suppose my particular electronic record was a poor attempt, but now that it's a requirement, I hear nothing but the same old complaints which made me conclude the electronic record was not ready for prime time, as they say. First of all, it takes too much of the doctor's time for data entry. In effect, the bulk of the record was typed in by me. An ophthalmologist friend, concerned with a much more limited data set, would use the computer (he had a MacIntosh) to jot down notes, and then turn it over to his secretary to expand into something readable and useful. He had three secretaries, by the way, so he was the fourth secretary in the system. Some other ophthalmologists expressed interest in his approach, but so far as I know, no one else adopted it. More recently, my physician daughter used the Kaiser-Permanente system which is considerably more sophisticated, and found the data entry imposed on her was burdensome. In fact, she devised a sort of measure of the extra work. She was staying at the office an extra two hours a day, falling further and further behind. So, she requested a 10% reduction in her patient load (and salary), but still fell behind. So, she dropped her workload to 60%, and found she then enjoyed the experience, although not the 40% reduction in income. From this I derive the estimate that the electronic medical record reduces physician efficiency by 40%. She says most of her colleagues just fake it to maintain their income. It's hard to translate a 40% impairment of physician efficiency into a reduction in quality, but one can make one's surmises.

I'm pretty proud that my daughter was unwilling to compromise her principles for money, but that really isn't the end of it. Both my system and the one she uses would record the patient's laboratory work instantly in our offices, which was very gratifying, except all of the eleven hospitals where I was on the staff refused or were unable to send the hospital laboratory work to my office for the inpatients. That was summarized on the typed discharge summary they mailed me, often several weeks after discharge, and added to the office record, that was also true of x-ray, pathology and electrocardiograph reports, so the office record was fairly complete, but hardly automated. But even if automated in some way, it was pretty unreadable. Quite a few records ran to a couple hundred pages, which is not something a doctor is likely to read before each visit. It provided me with a useful record, but I certainly was not going to spend the time to organize it. What the computer needed to to was to organize a system of periodic summarization, but to be useful it had to condense the summary to 10% at most of the raw data. When some company produces a program that will take the input load off the doctor's back, and also automatically produce concise and accurate summaries, well, let me know. My present position is that EMR is not ready for everyday use, it's going to cost an enormous bundle of money to force that conclusion to be changed, although here and there, for some specialties, it produces a few benefits. Meanwhile, let me tell you a story about how EMR remains a tantalizing dream.

We have half a dozen hospitals which deliver babies in central Philadelphia, all within twenty blocks of each other. We also have an ambulance service which operates on the strict rule of taking every patient to the nearest hospital. Pregnant ladies still go shopping or to the theater when they are at term; it's sort of a form of bravado. And so, it is inevitable that quite a few of them get caught in labor, call an ambulance which obediently takes them to the nearest hospital. But their doctor is waiting at another hospital, and has their records, which the receiving hospital does not. I can tell you with the best will in the world, it takes at least an hour, often more, for a hospital to get a call, find the record, and fax a hundred pages to the hospital that needs it. By the time all this confusion gets sorted out, the baby has been born without any use of records. No doubt all the confusion could be straightened out by an efficiency expert, either by periodic fire drills at the hospitals, or raising Cain with the ambulance people, depending on who is most likely to prove helpful. Meanwhile, we dream of an electronic record to solve a problem that is pretty low on the priority list of all concerned. It's a good idea; it's just not ready.

It's worth a little time to ask why the hospital computer departments have proven to be so resistant to merging with physician office systems. The answer offered by some observers is that hospital IT departments report to the accounting department, who don't want to negotiate with a new set of bosses. There's something to that, or rather there was something to that, thirty years ago. Furthermore, the Nursing Profession has long maintained a resistance to being left alone with sick patients, taking orders from doctors who are miles away in their offices. Therefore, this attitude surfaces in an adamant resistance to verbal orders, demanding an order be signed by a doctor on the spot. No doubt all these anecdotes have variable application in different institutions, but they give a hint that resistance to the electronic record is not entirely based on technology, but it is no less real.

Originally published: Wednesday, February 05, 2014; most-recently modified: Thursday, May 16, 2019