Information Technology Reference
In-Depth Information
processing may increase the recognition rate of barcodes. As robust barcode
recognition algorithms are researched (Gallo and Manduchi 2009 ; Wachenfeld
et al. 2008 ), it is expected that more efficient BCMA may be developed.
6.4.2 Expand the Use of RFID in Healthcare
Investigation of the effects of electromagnetic waves from RFID devices in
healthcare settings has not been adequately performed (Akiyama et al. 2006 ).
There have been studies of the effect of radio waves produced from RFID devices
on medical equipment in laboratory settings (Ministry of Internal Affairs and
Communications 2006 ), but little research has been conducted in real medical
facilities.
There are huge amounts of transaction data that contain patient status, medica-
tion history, and the list of administered drugs with a lot-number. A distributed
processing technology to handle such data efficiently is required to realize realtime
CDSS (Clinical Decision Support System). To effectively work with the life cycle
management of ethical drugs' supply, safekeeping, mixture, administration and
disposal, the standardization of a drug master code and a medical information
model should be developed. Solving these issues will accelerate the expansion of
RFID in healthcare.
6.4.3 Data Mining for Medical Safety
Akiyama et al. developed POAS (Point of Act System), a system that collects the
transaction logs of medication administered to patients. By conducting data mining
on tens of millions of transaction data, they found that warnings about an injection
occurred most frequently when the nurses' shift change occurred (Akiyama et al.
2008b ). Data mining based on the real dynamics of medical treatment is impossible
with conventional ordering systems and medical electronic records. By using
mobile terminals, the practitioner, patient, medications, injections, place, time,
and routes are all recorded, allowing investigation into which part of the workflow
medical accidents most frequently occur. Analyzing the time lag and the route
between ordering and administering medication may optimize stock and material
distribution.
This kind of data mining has already been performed in other fields, but data
mining in healthcare is still in the early phases of development. As I have described
previously, the standardization of the medical information model is still developing,
and we have to combine complicated, structured data to build data mining targets
because the treatment process includes many kinds of administering medication,
and these administration transactions are stored separately.
Search WWH ::




Custom Search