Title: Osler Ward and the Duke Cardiology Data bank

Contributor: Frank Starmer

In the late 60s, NIH announced the availability of funds to study myocardial infarction. Andy Wallace, Joe Greenfield and I put together a proposal that would bring the first computer to Duke Hospital. I was the engineer and my job was to develop the software to monitor a variety of signals from patients. In addition, we developed a questionnaire for capturing baseline information. With these tools, we followed the progress of patients on the CCU and the Surgical Recovery Room.

All went well for a couple of years until there was a change in the leadership of the Dept of Surgery. With the change in leadership came a change in the Surgical faculty and suddenly the surgical component of the system was no longer being used. Concerned that something was wrong with the software, I went to the recovery room and watched what was happening. It seemed to me as an engineer, that post op management was pretty straight forward and that my ability to produce beat-to-beat changes in cardiac stroke volume to six decimal precision and continuous ecg analysis was information overkill. I was discouraged by my lack of understanding about what clinical medicine was all about.

Within a few days, Dr. Stead came by our computer area (adjacent to the CCU) and commented, "well Frank, now that you've discovered that I was correct about patient monitoring, what are you going to do with all this expensive computer equipment we bought you". I looked at those steely blue eyes and said that I needed his help. He recommended that I learn some medicine so that I would be better prepared to support doctors in what they did, instead of what they said they did. He arranged for Galen Wagner to teach me Physical Diagnosis. I finished Galen's program and Dr. Stead immediately announced the next step in my education: to round with him on Osler.

Through 3 or 4 years of rounding on Osler with him, I had more than my share of fear and trembling sessions, but I gradually understood better what doctors did and how we might adapt the computer system to support these needs. I saw the doctor as a people enabler, and I saw myself, an engineer, as a doctor enabler.

During our patient monitoring era, I programmed a tool for acquiring data. I had programmed another tool for searching the data and extracting subsets of similar patients. Actually the search program was Dr. Stead's idea - he simply wanted to find all patients that were identical with respect to whatever he wanted to measure. We knew this was a nutty idea, since we had 100s of variables and only 30 patients - but we did it none the less. After a year, we had a couple of hundred patients and suddenly we were able to find subsets of patients containing more than one person. Dr. Stead's idea was that if we could identify similar patients, monitor their long term outcome, then we could make better decisions - evidenced-based medicine in 1969.

One day, we were rounding on Osler and gathered around the bed of a young lady with syphilis. The intern made a complete presentation to which the JAR and SAR had no additional comments. Dr. Stead looked around the room and asked: "what is this lady's problem"? The intern repeated his diagnosis and around the room he went, never acknowledging whether the answer was correct or not. Finally he got to me and asked, "Frank, what is this lady's problem". I had watched her during all the presentations and clearly saw that this was one very angry young lady. Obviously, the diagnosis of syphilis was not the answer he was looking for so I said - "I think she is mad at all of you - perhaps because of her problem". He replied "go to the head of the class". Leaving the room, we walked out into the hall and he said, "Frank, you've learned all the medicine I can teach you, now go back to the computer and make the Data Bank work".

Of course, I could not make the Data Bank work, but by building tools that enabled Galen, Bob Rosati, Vic Behar, Fred McNeer, Rob Califf and others to chase their curiosities, the Data Bank did succeed. The rest is now well known history.

Title: Avoiding Academic Failure with an "I"

From time to time, Gene and I struggle with the question of how you test students (junior learners) - and this inevitably leads to the question of whether he ever failed someone or not. His answer, which he has repeated a number of times, is quite simple.

"I never failed anyone. I realized that there will always be slow learners and fast learners and it is unreasonable to think that a single academic session can accomodate both extremes. Consequently, if I determined that a student was a slow learner, then I gave him/her the grade of "I" (incomplete) with the understanding that the student would continue their learning and eventually replace the "I" with a passing mark".

This seems like a perfectly reasonable approach assuming that there is no financial penalty for slow learners. Our experience has been that they may be slow in one area, but quite fast in others. Are medical schools still flexible enough to accomodate slow learners?