Profiling variation in hospital care quality: Clinical, Statistical and Policy Considerations
The speaker for the SA Branch May meeting was Dr Isuru Ranasinghe, a clinician, researcher, and passionate advocate for improving patient safety and quality, currently working at the University of Adelaide and the Basil Hetzel Institute.
Isuru began his talk highlighting current concerns about safety and quality of hospital-based care in Australia, elaborating on efforts to publicly report variation in quality, and the specific clinical, statistical and policy challenges of profiling institutional variation in quality using national data on complications following implanted cardiac pacemakers and defibrillators as an example.
He then referred to a report by the US Institute of Medicine and British Medical Journal: “To Err is Human” which highlighted a significant number of deaths due to medical error. Isuru also mentioned adverse events attributable to hospital care, resulting in permanent harm or extended hospitalization, as well as a billion dollars added costs to health system.
In Australia, 10.2 million hospitalisations occur annually for treatment of a range of acute and elective conditions across >1000 geographically dispersed public and private hospitals and providing this complex care gets exceedingly costly. Health services and policymakers are increasingly expected to provide high-quality hospital care in the face of rising demand and cost although measuring and profiling the variation in health care quality remains challenging.
Being well acquainted with the US Medicare system of “building quality measures for public reporting”, Isuru incorporated his ideas into the Australian health care system, where he successfully managed to merge national cardiovascular data to measure outcomes of care in a low cost way. He examined the overall complication rates that occurred in hospitals or caused re-admission to hospital to deduct the safety rate and quality of care provided to patients who had a permanent pacemaker implanted. Isuru finished his talk by discussing the methodological consideration of estimating the hospital level and a patient’s level risk standardised complication rates using hierarchical (2-level) logistic regression models. Following the talk the discussion continued over dinner at Lemongrass Thai restaurant.