NPS looks for insight into medicines use
NPS MedicineWise has launched a four-year program that will collect de-identified patient data from GP clinical software to examine how medicines are prescribed and used.
The program, named MedicineInsight, will eventually involve 500 general practices and will gather data from approximately two million patients.
The program's national manager, Nancy Huang, said it will be the first time researchers will be able to gather clear, real-time and real-world prescribing information from GP clinical software on this scale.
“We will analyse this and obviously validate it, but we'll also provide it back in a meaningful report to GPs so that this can become part of the quality improvement process,” Dr Huang said.
“GPs can start to look at their own prescribing pattern, as well as over time to have comparators with their peers in certain clinical areas, including clinical indicators, adherence guidelines, prescribing patterns and medicines use.
“We would also see the aggregation of all of the practices' data to be valuable for us as a community about how certain new drugs are being taken up or whether we can pick up certain effects of the medications like their benefits or some of the harms.”
HCN has run a prescribing research program called GPRN since 2000, involving a database that collects data from a panel of GPs using Medical Director. However, Dr Huang said MedicineInsight was much larger in scale and would not be restricted to just MD users.
She said the four-year program has been funded by the Commonwealth with a target of 500 practices, including between 2000 and 2500 GPs, and about two to 2.5 million patients. The program is entirely voluntary, and will kick off in a proof-of-concept trial of 10 practices in Victoria before moving to a pilot phase of 120 to 150 practices across four states.
“The aim is to collect data at an individual level so we can build a warehouse to get a longitudinal view,” she said.
NPS will collect the information using data extraction software provided by a technology partner – the name of which Dr Huang cannot yet reveal – which will be installed on practice computers and do the data collection in the background.
“The program will work in the first instance in our proof-of-concept and pilot phases with three of the clinical systems – Best Practice, Medical Director and Zedmed,” she said. “This installed software will work seamlessly in the background, so it doesn't interfere or require the GP or the practice staff to do anything to extract the data.”
All data extracted will be de-identified and encrypted before it leaves the practice, a protocol approved by the RACGP national research and evaluation ethics committee.
“We are using a sophisticated pseudonymisation algorithm which will provide each of the patients with a unique identifier that will be common in our data warehouse,” Dr Huang said.
“The type of data that we are collecting is the clinical information that's in most of the coded fields, not progress notes. It is the information entered in like diagnosis, prescription, referral, pathology results, biomedical measures like blood pressure, smoking status, weight etc.
“The way the software will work is it's not going to be a heavy burden – it is a light process in footprint. At first it will extract the principal data, and we will then periodically time subsequent extracts to just take out what has changed, rather the whole dataset.”
Dr Huang said by extracting data as it is entered in its raw form means the project team will not start with any prior assumptions about what the data means. “But we are using reference datasets so that the different source data will ultimately build into a data model that is common across the different systems, using coding where it is relevant to our reference datasets.”
NPS will develop information packages for general practices and patients about the benefits of the program, but patients are free to opt out if they don't want their data used. Individual GPs' prescribing patterns will not be monitored.
“We have very clear protocols for practice staff and GPs who are taking part in this to both answer the questions but also, within the software that is installed, there is a panel they can click on, so if the patient still feels they want to opt out … [the GP can] press a button and in the next extract they will not include that person's information.”
Dr Huang said while some patients might not want even de-identified data used, focus groups with consumers had shown that they are very comfortable with their de-identified data being used as long as they feel the information is going to help their quality of care.
The MedicineInsight team is looking for volunteer practices to take part in the program. NPS is considering making available a small payment to cover any administrative costs for practices.
Contact details are on the MedicineInsight page of the NPS website.
Posted in Australian eHealth