12 April 2017
Focusing on unwarranted clinical variation to assess performance
Every patient is different and this means that healthcare varies – it must, if it is to respond to patient needs, expectations, social circumstances and capacity to manage their own care . However, there is increasing evidence that healthcare also varies for reasons that are not related to patients’ needs and expectations. While some variation is warranted, it can also be unwarranted.
This sets up something of a paradox care should vary and be person-centred, responsive and tailored yet at the same time, care should be consistent with regards to the application of clinical guidelines and differences across geography or time are often met with concern.
This is why assessment of how the system performs should focus on unwarranted variation – where healthcare does not align with evidence-based care applied in the right way and in the right amount to address patient needs, expectations or preferences.
Over the past five years, BHI has been working to develop measures of unwarranted clinical variation. That work, done in conjunction with other NSW health pillar organisations, has focused on outcomes following hospitalisation for a range of conditions – heart attacks and heart failure, stroke, pneumonia and other lung conditions, hip and knee surgery.
Our approach is to focus on variation across hospitals in mortality and readmissions, once we have taken account of differences in patient characteristics such as age, sex and other illnesses that make it more or less likely for them to die within a month of hospitalisation; or to be readmitted in the month following their discharge. Using this approach helps to pinpoint areas of unwarranted variation – where patient outcomes differ significantly from what we would expect them to be.
Today we are releasing our latest results. This work has been a massive undertaking. Data from more than 210,000 patients have been used to build a detailed picture of variation in outcomes across 79 public hospitals in NSW over the three-year period from mid-2012 to mid-2015. The data is available in a range of formats – in print and online – and has been tailored for different audiences including patients, clinical experts, local providers and managers, regional decision-makers and system leaders. Summary documents provide the findings ‘at a glance’, methodological reports explain how measures are calculated, the main reports show the detailed analyses, individual hospital profiles detail the characteristics of local patient cohorts and outcomes, and a rotating cube presents a wide-angle view of key points.
The work builds on a 2013 BHI report that provided the first set of hospital mortality data to be publicly released in NSW. Following that release, we heard about the efforts of staff in many hospitals to examine and improve the care they give to their patients. This release provides feedback to each hospital on progress, highlights hospitals with better outcome measures and points to areas of care where the variation we currently see in patient outcomes may be unwarranted and requires further investigation.
TagsAccessibility and timeliness 4 Ambulance services 2 Appropriateness of healthcare 1 BHI - general 4 Challenging ideas 1 Data 5 Effectiveness of healthcare 2 Efficiency 1 Elective surgery 2 Emergency department 1 Healthcare services 2 Hospital care 3 Patient experience 3 Safety and risk 1
Kim is the Acting Chief Executive of BHI. She has extensive experience in health services research in Australia and internationally.
Jean-Frederic is the Chief Executive of BHI. He is currently on secondment to the Agency for Clinical Innovation.
Hilary is our Director, Strategic Relations. She has extensive experience of developing policy and strategy around the role of information in improving health and healthcare, mostly in the UK.
Renee Carter is a senior analyst at the Bureau of Health Information. She holds a Masters degree in health and social policy from the London School of Economics and a Doctorate in epidemiology from McGill University.