Understanding patientexperiences from miningprimary care dataCentre for Health InformaticsFilippo GalganiAdam DunnMargaret WilliamsonMalcolm GilliesGuy Tsafnat
General Practice EMRs• Aim: measure quality of care for a range of conditions in a diversepopulation using GP EMR data.• Dataset: longitudinal data (2.5 million Australian patients) includingprescriptions, diagnoses, pathologies, referrals• Patients’ journey: grouping patients by experience to detect relevantpatterns in data over time..
Big Data Problems• Data collected to keep patient history:– Dealing with missing information– Inconsistency– Combination of short text fields (not coded) and numericalvalues• Doctors’ time constraints make data entry inaccurate• Progress notes not available (privacy issue)• Patients may visit other practices (thus missing information)• Events happen irregularly
Conclusion• Data mining on GP EMRs is challenging due to thenoisy, messy and sparse nature of the data• Analyzing journeys is possible, it required:– Temporal reasoning (infer missing events)– Natural Language Processing (solve textualinconsistencies)– Machine Learning (predict missing information)– Domain knowledge (for modeling)
Acknowledgment• This research was funded by the Australian Department of Healthand Ageing through the NPS MedicineWise as part of theMedicineInsight Program.• I wish to express my gratitude to:Malcolm Gillies and Margaret Williamson from NPSAdam Dunn and Guy Tsafnat from UNSW• Thank you for the attention