- From: Chintan Patel <cop2101@gmail.com>
- Date: Sun, 11 Sep 2005 23:03:23 -0400
- To: "Amit Sheth @ LSDIS" <amit@cs.uga.edu>
- Cc: public-semweb-lifesci@w3.org
Amit, In the recent controversial article in JAMA on Role of Computerized Physician Order Entry Systems in Facilitating Medication Errors (http://jama.ama-assn.org/cgi/content/abstract/293/10/1197) the authors give some interesting numbers on frequency of occurrence of various medication errors *facilitated* due to a CPOE system In context of activities in this group, I believe the above article gives us a stepping stone to show the benefits of SW technologies to solve the current problems. In the article, the authors group errors broadly in to (1) information errors generated by fragmentation of data and failure to integrate the hospital's several computer and information systems (2) human-machine interface flaws reflecting machine rules that do not correspond to work organization or usual behaviors. I believe (1) sounds very much like a typical Semantic Web use-case.The challenges, nevertheless require looking into complex organizational workflow patterns and somehow making the semantics explicit (of not only the data but the processes) to make the integration smoother. There are also some definitive stats on medication errors in the IOM report http://www.nap.edu/books/0309068371/html --chintan On 9/7/05, Amit Sheth @ LSDIS <amit@cs.uga.edu> wrote: > As we transition our first Active Semantic Documents application, > Active Semantic Electronic Medical Record into our partner organization, > a cardiology practice (see > http://www.w3.org/2005/04/swls/ > or http://lsdis.cs.uga.edu/projects/asdoc/ ) > we have an opportunity to measure effectiveness of a > SW technology in a live setting. > In this context, I am looking for norms/studies that could > give us percentage of medical records > that have errors/imperfections of a particular type (eg, drug-drug > interaction, > failure to check on allergy, %age of cases when patient could > have been prescribed same class of drug that is cheaper, etc.). > Thoughts on exactly what to measure (please see the demo so > you could see what could be relevant) would be welcome too. > So far we have planned to log each rule activation/violation (drug > interaction, > drug allergy, %age time physician changed the drug based on > preference/cheaper alternative based on patient's insurance, > diagnosis-treatment mismatch, etc.). > > Thanks in advance, > > Amit Sheth > > -- Chintan O. Patel PhD student Department of Biomedical Informatics Columbia University http://www.dbmi.columbia.edu/~cop7001/
Received on Monday, 12 September 2005 03:35:40 UTC