Feature

Using data to shape the future of health services


Amanda Lyons


7/09/2018 2:36:10 PM

newsGP spoke with two organisations running primary care data projects that show the potential value of secondary patient data for planning – and predicting – health services at a population level.

General practice data can be extremely useful in predicting future health services and informing current ones.
General practice data can be extremely useful in predicting future health services and informing current ones.

The ongoing debate over My Health Record highlights the sensitive nature of handling information from medical records, particularly in terms of the secondary use of patient data.
 
An example of such secondary use is when data collected for one purpose, ie to register a personal account, is then used again for another purpose, such as targeted advertising. This is how a company like Facebook monetises its social media platform.
 
However, not all secondary use of data is carried out for monetary purposes.
 
newsGP spoke with representatives from two organisations that collect secondary patient data from general practices to help anticipate and even shape future medical services and needs: Associate Professor Chris Pearce, Research Director of Outcome Health; and Gloria Antonio, Executive Manager, Health Insights and Business Delivery for NPS MedicineWise.
 
Their respective systems are known as Population Level Analysis and Reporting Tool (POLAR) Data, which works with Primary Health Networks (PHNs) in Victoria and New South Wales; and MedicineInsight, which collects information from more than 700 practices and 3300 GPs.
 
Protecting patient privacy
POLAR and MedicineInsight operate on a population level and all patient data is de-identified.  However, the way patient privacy is handled is still very important.
 
Associate Professor Pearce believes robust processes are required to ensure the ethical collection and use of general practice data.
 
‘You have to have a very good governance and ethics structure,’ he told newsGP. ‘We have ethics approval from the RACGP and we’ve got a strong governance structure around the use of the data with all the privacy controls.’
 
POLAR exercises strict controls over the de-identification of the data it uses.
 
‘The data is extracted from the practice and de-identified at the practice, and then it goes to the PHNs [Primary Health Networks] and they use it for their population health planning,’ Associate Professor Pearce said.
 
‘The data is identifiable within the practice, so they can do things like generate lists of patients who haven’t had flu shots or whatever they like.
 
‘But, at population level, we’ve got rules. For example, we don’t release data that is in too-small groups, because that becomes identifiable.
 
‘It’s all national-standard stuff, and it’s the sort of information that’s used all around the world and the risks are pretty low. Even if people hacked our database, they wouldn’t get anything they could use.’
 
MedicineInsight also ensures it has careful structures in place in relation to patient data.
 
‘All requests to use data are assessed by an independent data governance committee, comprising GPs, researchers, data security and privacy experts, external academics and legal advisors,’ Ms Antonio told newsGP.
 
‘Patients are not required to give written consent due to the non-identifiable nature of the data collected; however, participating practices inform patients … through poster displays and information leaflets.’
 
MedicineInsight also takes care to conduct its research to a rigorous ethical framework and practise transparency about the final results.
 
‘The ethos of the program is that all access must align with our mission to enable people to make and act on the best decisions about medicines, and be for the public good,’ Ms Antonio said.
 
‘Cleaning’ the data
When GPs enter information into a patient’s medical record, they don’t do so with a view to aggregating it on a larger scale. As a result, information in patient records is not categorised or classified in a way that assists analysis, such as by disease, whether there has been a recent hospital visit, patient age or gender, or so on.
 
This difference in process means primary care data needs to be categorised before it can be analysed, which takes a lot of work.
 
‘We’ve spent a lot of time cleaning and coding the data, so what we use after these programs is much more useful than just raw extracts of the GP’s data,’ Associate Professor Pearce said.
 
‘We go through and code all the diagnoses, the medications, pathology, immunisations, and that’s allowed us to develop chronic disease mappings and clinically useful tools to make analysis a lot easier.
 
‘It is time-consuming, but once you’ve done it, it makes the next steps a lot easier.’
 
Both systems use general practice software to extract the data they need.
 
‘Our MedicineInsight program securely extracts monthly longitudinal, de-identified, whole-of-practice data from the clinical information systems of consenting practices, which NPS MedicineWise analyses and synthesises into tailored practice reports for each participating practice,’ Ms Antonio said.
 
‘Our national network of highly-trained clinical services specialists then provide these individual reports back to practices through in-practice visits.’
 
Getting results
Once the data is collected and analysed, MedicineInsight and Outcome Health use it to help provide insights and guidance for the primary care sector.
 
MedicineInsight provides individualised reports to its participating practices.
 
‘[Our reports] enable GPs to review their own patterns of prescribing and patient care and make quality improvement decisions based on contemporary, real-time data insights and clinical evidence,’ Ms Antonio said.
 
It also works towards a larger goal of supporting quality improvement across general practice as a whole.
 
‘The data [we collect] reflects activities within general practices, including the conditions and risk factors that patients have, the medicines prescribed, vaccines delivered and the results of pathology tests,’ Ms Antonio said.
 
‘This makes it a valuable source of data for research and to inform primary healthcare policy development and health care decision-making at local, regional and national levels.’
 
Meanwhile, Outcome Health has been using POLAR data to help design and create population-level monitoring and prediction tools that can help predict problems before they become major health events.
 
‘People go to GPs, they don’t necessarily go to hospitals. So general practice is the place to pick up a lot of stuff before it becomes a big issue,’ Associate Professor Pearce said.
 
‘We’ve been doing some work on almost real-time monitoring of diseases or adverse events. We have used [the data] to generate a risk prediction tool at the general practice level, predicting the risk of someone going to the emergency department in the next 30 days.
 
‘We’ve done a bit of work with Monash Department of General Practice on emergency department stuff. We’ve also just got the Digital Health Authority test beds.
 
‘There’s quite a bit of activity, and quite a bit in the pipeline, as well.’



general practice data health informatics medical records patient data





Comments



 Security code