Unlocking ‘Big Data’ medical information

Unlocking ‘Big Data’ medical information

New Zealand researchers have developed a new way of extracting and classifying important information from the ‘Big Data’ stored in hundreds of thousands of electronic health records.

Medical notes in general practice are written directly into electronic records. Much of the information in those records is currently not accessible for an in-depth understanding of why people visit their GP, why differences exist between different population groups’ visits, and how those visits might change over time.

One of the researchers, University of Otago, Wellington Professor Tony Dowell, says that a partnership between clinicians and IT specialists in local primary health organisations has enabled them to unlock the stories within the medical records from primary care consultations, and estimate more accurately the level of childhood illness seen in general practice.

“We have developed and tested a computer algorithm that can extract the ‘natural language’ and free text written by doctors at each consultation and code it into internationally recognised illness classifications,” Professor Dowell says.

The research team tested the method on 5000 medical records by comparing the computer algorithm to the diagnosis of expert clinicians reviewing the notes. Having achieved good results from this process the next step is to explore the stories of three quarters of a million records and estimate the prevalence and burden of childhood respiratory illness in primary care.

All of this information is obtained and used without any identifiable individual patient data being made available to the research team.

Details of the research about the computer algorithm have been recently published in the international journal BMJ Open.

“We have previously had good information about what illnesses are seen in hospital, and what workload that creates for health services. Now we will be able to get an idea of what happens in general practice and primary care.

“This is a big step forward in supporting more effective health planning at every level, for the general practice, local health authorities and at a national level,” Professor Dowell says.

ENDS

© Scoop Media