STATEWIDE LONGITUDINAL ANALYSIS TO PREDICT FUTURE DIALYSIS DEMAND IN WA: PROOF OF CONCEPT STUDY

H MOODY1, H KULKARNI1, B MAI3, T EITHELHUBER2, M BURLEY1

1Health Networks Branch, Clinical Excellence Div – DoH, EAST PERTH, AUSTRALIA, 2Data Engineering, Data and Information Systems – DoH, EAST PERTH, AUSTRALIA, 3Strategy, Policy and Planning Div- DoH, EAST PERTH, Australia

Aim: To predict the future for dialysis demands in the state, predict the geographical areas of highest needs, and to inform palliative care services within WA who are not suitable for dialysis
Background: Traditionally, demand for future dialysis services is based on historical data within state. Significant changes to incidence of dialysis in Australia in 2008 prompted the Renal Health Network (RHN) to explore new methods to predict future dialysis needs. We collected retrospective eGFRs of less than 60 from all but one laboratory in WA from 2010 to 2017.
Method: All eGFR’s of <60 from Pathology laboratories in WA were transferred to the Data Linkage Branch in the Health department of WA. De-identified data that included Age, gender, post code, eGFR, laboratory & date of service were provided for analysis. WA Mortality data and ANZDATA was linked to provide censoring for mortality or initiation of dialysis respectively.
Retrospective data from 2010 to 2017 was analysed. A linear regression model was used to predict the year at which an individual would achieve an eGFR of 7ml/min, the average eGFR at which Australians commence dialysis.
Results: 2,985,552 records in 256,142 persons (Mean 11.7) data points were analysed.
Actual incidence of dialysis, Predicted Incidence and Error (%) in 2013 (352, 276, -22%), 2014 (381, 328, -14%), 2015 (445, 374, -16%), 2016 (426, 435, 2%) and 2017 (504, 516, 2%) respectively.
Conclusion: Our model predicts the demand for dialysis. Accuracy and ability to predict the annual demand improves with longitudinal cumulative data. Established model will be used for future prospective analysis and will explored to predict clinical outcomes.


Biography:
Dr Harry Moody is CoLead of the Renal Health Network, along with Dr Hemant Kulkarni; with keen interests in providing patient centred affordable care closer to the home. Harry is an astute renal physician with his interests in clinical medicine, research and epidemiology.

Categories