Australian patients, clinicians and health policy can benefit from a ‘big data’ approach to measuring patient outcomes following in-patient cardiac care.
Early results from a data linkage analysis of over 100 million healthcare records from more than 1,000 hospitals suggest there is room to improve overall outcomes and consistency between hospitals.
Cardiologist and National Heart Foundation Future Fellow Dr Isuru Ranasinghe told the limbic cardiovascular hospitalisations were common and costly yet the results of care were not fully reported.
“We can tell how many people die in hospital but we’ve never had the ability to broadly evaluate how many people survive down the track. Our project is to expand capacity and to be able to assess outcomes of care not just in hospital but months after people leave hospital.”
In a collaboration with the Data to Decisions CRC, the analysis leverages emerging data linkage capabilities and big data techniques widely used outside the health sector.
The Observing Recurrent Incidence of adverse Outcomes following hospitalisatioNs (ORION) project is just starting to deliver outcomes on heart failure, acute myocardial infarction, bypass surgery and cardiac devices.
Speaking ahead of the inaugural South Australian Cardiovascular Research Showcase, Dr Ranasinghe said the outcomes on heart failure were not as good as might be hoped.
“We’re so good at treating heart attack that we see a large number of people presenting to hospitals with heart failure; a common late consequence. And in those patients we see about 1 in 10 die within 30 days of admission and a further quarter are readmitted to hospitals within 30 days of discharge.”
“So it’s really a revolving door scenario and these people stay in hospital for a lot longer than other patients so it’s a huge cost. It tells you at a national scale how much work we have to do to improve care.”
He said cardiologists would not be surprised as the findings reflected international data and ‘many of us have been wanting this data for a long time’.
The study also highlighted considerable differences in survival rates between hospitals.
“We’ve known for a long time that hospitals are not the same in their capacity, staff, specialist knowledge and systems for quality and improvement. We know there is enormous heterogeneity in how guidelines and pathways are applied in practice.”
He said the findings would be helpful in identifying the hospitals to learn from and which hospitals to target for improvement.
More comprehensive data was useful to clinicians to facilitate those quality improvement efforts and, from a policy perspective, could help determine strategies such as funding and incentives.
Better outcome data was also valuable to patients to ensure fully informed consent.
Dr Ranasinghe said the project was also planning to develop software packages so hospitals could run the code through their own systems to generate information locally and benchmark it against national averages.