FISCHER@RUTGERS.ARPA (03/03/84)
From: Ron <FISCHER@RUTGERS.ARPA>
Apparently because of fierce competition, much current information,
particularly with regard to algorithms, is proprietary. Worst in this
regard (a purely personal opinion) is HP who seems to think nobody but
HP needs to know how they do things and physicians are too dumb to
understand anyway.
...
They offer an advantage to small hospitals by offering verification of
the analysis by a Cardiologist (for an extra fee).
What the latter seems to say is that the responsibility for accepting
the diagnosis is that of the local cardiologist. I cannot see a
responsable doctor examining a few runs of a program's output and
proclaiming it "correct."
A hedge against complaints of computers taking over decision making
processes from human has been that we can look at the algorithms
ourselves or examine the reasons that a system concluded something.
If this information becomes proprietary the government will probably
license software for medical purposes the way the FDA does for new
drugs.
Imagine a testing procedure for medical diagnostic AI programs that is
as expensive and complicated as that for testing new drugs.
(ron)
[Ron makes a good point. As a side issue, though, I would like
to mention that H-P has not been entirely secretive about its
techniques. On March 8, Jim Lindauer of H-P will present a seminar
at Stanford (MJH 352, 2:45PM) on "Uses of Decision Trees in ECG
Analysis". -- KIL]jwb@ecsvax.UUCP (03/09/84)
An extract of a previous submission by me mentioned the overreading of an ECG interpretation by a cardiologist. What I meant (and what was not clear) is that the cardiologist is looking at the raw ECG, not the output of the computer (although a lot of preprocessing is often done which is hidden from the cardiologist--this is another different problem--at least it looks like what you would get from a standard ECG machine). On a related issue, medical decisions regarding the treatment of an individual patient *have* to be made by the local physician treating the patient (at least that is long standing medical practice and opinion). The overreading offered by the remote services is a look at the reconstructed input ECG by a Board Certified Cardiologist and is intended to be analagous to a "consultation" by a more experienced and/or specialized physician. The name of the service is Telemed, not Telenet as I incorrectly typed. Committees of the American Heart Association and the American College of Cardiology are attempting to set standards for computer (and human) interpretation of ECG's. A snag is that different preprocessing of the ECG's by different manufacturers makes it rather uneconomical to acquire a large number os "standard" ECG's in machine readable form. I think the FDA is looking at all this and I think under current law they can step in at their whim. So far they seem to be waiting for the above groups to present standards (since they don't seem to have the resources to even start to develop them within the FDA). Jack Buchanan Cardiology and Biomedical Engineering UNC-Chapel Hill Chapel Hill NC decvax!mcnc!ecsvax!jwb
julian@deepthot.UUCP (Julian Davies) (03/09/84)
As a matter of human engineering, I think "expert" programs for practical use must be prepared to explain the reasoning followed when they present recommendations. Computer people ought to be well aware of the need to provide adequate auditing and verification of program function, even if the naive users don't know this. The last thing we need is 'expert' computers that cannot be questioned. I think Weizenbaum had a valid point when he wrote about programs that noone understood. And I would be unhappy to see further spread of computer systems that the human users cannot feel themselves to be in charge of, especially when the programs are called 'intelligent' and the technology for answering these questions about the reasoning processes is fairly well established. Julian Davies
mcmillan@eosp1.UUCP (Tobias D. Robison) (03/11/84)
The correct way for a cardiologist to "take responsibility" for a computer ECG analysis is to overread EVERY SINGLE ANALYSIS. The cardiologist looks at tracing printouts of the ECG monitored signals, and determines, to a degree of accuracy he finds ethically acceptable, that the computer analysis is reasonable. Please bear in mind that in all cases of pathology where the patient will be treated, the cardiologist will probably go over both the tracings, and a computer matrix of measurements of key parts of the wave forms, with a fine toothed comb. It is usually very easy for a cardiologist to take a quick look at a tracing and determine whether the patient's heart is abnormal. - Toby Robison allegra!eosp1!robison decvax!ittvax!eosp1!robison princeton!eosp1!robison (NOTE! NOT McMillan; Robison.)