home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!dtix!darwin.sura.net!wupost!m.cs.uiuc.edu!vela!pd-dom!schuette
- From: schuette@wl.com (Wade Schuette)
- Newsgroups: comp.ai
- Subject: Re: Legality of Expert Systems
- Message-ID: <1992Sep16.000415.14452@wl.com>
- Date: 16 Sep 92 00:04:15 GMT
- References: <1992Sep9.225854.1619@wl.com> <1992Sep11.163619.12477@msuinfo.cl.msu.edu> <ALMOND.92Sep14120951@bass.statsci.com>
- Organization: Warner Lambert / Parke-Davis
- Lines: 50
-
- In article <ALMOND.92Sep14120951@bass.statsci.com> almond@statsci.com (Russell G. Almond) writes:
- > [schuette wrote...]
- >| >A related question is, at what point will a Physician or Hospital be
- >| >successfully sued for FAILING to use an expert system which CLEARLY would
- >| >have PREVENTED the problem which followed.
- >
- >One of the things I would definitely like to see, is an expert system
- >(especially a probabilistic one) used as a defense in malpractice
- >suits. For example, we did not ___ [perform such a test, perform such
- >a procedure] because given the information we had, the probability
- >that such a test would be needed was low [ below ___]. I think this
- >might produce cost effective, as oppossed to cover you neither
- >regions, perscription and testing decisions.
- >
-
- Russell, don't get me wrong -- I'm also serious about doing something
- about malpractice, and have been following this area since 1976 when
- I started working in medical computing, while married to a Health Care
- and regulation specialist attorney, yet. The American Medical Informatics
- Assn. has a lot of work on this.
-
- The probability approach has some problems. For one thing, the numbers
- are notoriously difficult to obtain reliably. For another, they confuse
- jurors and are easy to cast doubt upon, as are the operations which
- combine fuzzy facts.
-
- I was looking more for a system that would get medical personnel up to the
- level of airline pilots, ie, make it EASY to have an automated and dynamic
- appropriate checklist. Of more concern to the patient than suing the Doctor
- is avoiding the problem in the first place. For this, there are often
- suitable checklists: Have you checked for contraindications? Did you
- remember to whatever the whatsis. Whoa, this person has x-tosis, which my
- memory (and online PDR) say contraindicates use of the drug you just
- prescribed-- do you want to see the reference? Etc.
-
- The "expert" need is to keep the list short and relevant and subservient
- enough that physicians will actually USE it. The cleanest area of
- malpractice, in some ways because it IS so obvious, is the list of things
- that everyone agrees should be done 100% of the time in situation X, but
- the physician forgot, or thought the nurse did it, etc. It's a major
- challenge to get THAT under control.
-
- Wade
-
-
- --
- ====================================================================
- R. Wade Schuette | All opinions expressed are personal.| B> |
- Ann Arbor, MI USA | | :) |
- schuette@wl.com | *** "Synergize Scientists!" *** | B> |
-