Te Whatu Ora Health New Zealand has released a proposal which says AI techniques may help improve healthcare speed and accuracy.
"Clinical coding plays a vital role in collating health data in a meaningful and efficient way", the proposal says.
"However, the clinical coding process faces challenges which leads to delivery latency and inaccuracy.
"AI-assisted coding will speed up our reporting to assist with production planning and policy analysis allowing the organisation to respond more quickly and efficiently to changes."
Clinical coding was undertaken in hospitals to support future planning of services, and to support research and policy development in the health system.
Associate Prof Grant Dick, of the University of Otago School of Computing, said this sort of AI-driven automation was already being integrated into everyday productivity software.
"I don’t see much issue with this proposal. Perhaps, if there was inadequate human oversight of the system, it may lead to erroneous coding of information that is used for reporting and planning.
"However, humans would also inevitably make mistakes in this process, so that’s probably a very low risk."
At present, clinical notes from all inpatient and same day patients discharged from New Zealand public hospitals were clinically coded and recorded in the hospital’s patient management system.
Clinically coded summaries of these discharges were forwarded to national collections where the information was loaded and stored in the national collection national minimum data set for easy retrieval and analysis.
"These codes have been identified and then entered manually by individual teams of clinical coders located around the country.
"There is no common work list which means all their teams work independently.
"Most hospitals code from paper or scanned records as well as some digital systems", Assoc Prof Dick said.
Application closed last Friday. Shortlisted applicants would be informed by January 26, with a view to the contract starting on July 1.
There should be no additional threat to patient safety or privacy, he said.
"This appears to be about bringing AI into the administration of the hospital’s patient management system for more effective and efficient use of available resources.
"It could lead to more consistent reporting of health data, which in turn may lead to better decision making around planning and resource allocation."
