A new report by the Grattan Institute is making headlines with the conclusion that the government could save up to a billion dollars in health care costs with a series of reforms. An overview of Controlling the cost of care: a billion dollar hospital opportunity on the Grattan Institute website suggests that State governments should:
“pay hospitals for treatments on the basis of an average price once all avoidable costs we can measure have been removed”
“make data available to hospitals so they can compare themselves to their peers and see where they can cut costs”, and
“hold hospital boards to account when they fail to control costs”.
The report also points out that it is up to hospital leaders, managers and clinicians to find the most appropriate ways to improve and provides a timeline for implementing change commencing in 2014-15.
One health policy analyst, however, has a different perspective on the evidence.”William Foggin” writes:
In A Study in Scarlet Sherlock Holmes observed that “It is a capital mistake to theorize before you have all the evidence. It biases the judgment”.
Dr Stephen Duckett and his colleagues at the Grattan Institute who worked on the hospital care paper released on 5 March would undoubtedly argue that they have all the evidence to draw the conclusions they have. All over the country hospital managers are lining up to disagree.
But first of all, a big thank you needs to go to Dr Duckett for beginning to ask the right questions about health care – how is it delivered and how efficient are these processes? – rather than simply arguing about who should pay for it.
Of course hospitals can become more efficient by eliminating avoidable costs. The problem is that the National Hospital Costs Data Collection data which underpins Dr Duckett’s work does not constitute all the evidence. While it theoretically includes data on the costs of every patient episode in every hospital, over half the data is cost-modelled – derived from estimates of the level of resources of a particular type that episodes of that sort consume. In other words, the NHCDC uses past estimates of resource use tp estimate current resource use.
A review of the NHCDC conducted by the Independent Hospital Pricing Authority observed that:
“The value of cost modelling was questioned given the widespread use of service weights in the modelling process. Stakeholders raised the following concerns regarding that application of service weights:
- It creates circularity in the costing process
- They lose their currency over time
- It is not relevant to the clinical practices of hospitals if they were developed elsewhere.”
Dr Duckett addresses the adequacy of the data head on, stating that:
“the NHCDC… is the most reliable and comprehensive patient-level data set on public hospital costs in Australia. Most states and the Independent Hospital Pricing Authority use it to set prices. There are concerns about its accuracy, particularly for comparisons between states or costs components (such as medical staff costs or operating room costs). Despite this, we believe the data are reliable enough for our analysis, especially as our recommendations don’t rely on interstate or cost component analysis”
While the report may not make recommendations based on interstate analysis, it presents figures showing state by state differentials. While ill informed commentators are likely to seize on these as support for their current agenda, the extent of these differentials in a relatively homogenous economy highlights the possible unreliability of the underlying data.
Several years ago I saw a comparison of total NHCDC costs by hospital with expenditure by hospital reported to the AIHW in the Public Hospital Establishments Database. With one outstanding exception (a Catholic hospital that reconciled to $2) the results were all over the place, with variations in both directions of up to $100m in some major metroplican hospitals.
Yes, the NHCDC data is the best we have, and the IHPA needs to use it as the basis for the national efficient price. But to use the data in its current state to drive the three year price reduction program recommended by Dr Duckett will simply replicate the trauma that descended on Victorian hospitals in the mid 1990s when casemix was introduced with very little flexibility and reduced budgets.
The report acknowldges that the NHCDC is improving.
“Furthermore, NHCDC data will be even better by the time our proposed changes could be implemented. Hospitals report that since our data were collected there have been improvements in processes, software and resources. More improvements, such as improved timelines, communications, training, and governance, are being planned or considered.”
As these improvements flow through governments can begin to place more reliance on divergence from expected costs as an indicator that there is room for improvement. But even then, improvement will need to involve bottom up process redesign at a hospital level. Funding cuts may focus the minds of those involved, but of themselves will not produce increased efficiency.
• “William Foggin” is the pen-name of a health policy analyst who wishes to remain anonymous.
Thanks William for engaging on this important issue. As you point out, we acknowledged the weaknesses of the NHCDC. I have never been a supporter of ‘cost modelling’, but the main weaknesses are at the ‘cost bucket’ level (e.g. operating theatre costs versus costs for medical staff), not at the aggregate level we use. By its nature, cost modelling reconciles to the general ledger so it is probably safe to assume that a hospital which is expensive on cost modelling data is expensive using any other basis.
We were also conservative in assessing what is ‘avoidable cost’. We looked at the distribution of unexplained cost and accepted that for various reasons some level of unexplained cost is reasonable. Basically, we said that only above-average levels of unexplained costs (compared to the lowest-cost hospital in the state) were avoidable. This ‘buffer’, means that cost data don’t have to be accurate to the last cent. Even with some inaccuracy, they can identify the big differences that need to be addressed.
Finally, I note your last two sentences: “But even then, improvement will need to involve bottom up process redesign at a hospital level. Funding cuts may focus the minds of those involved, but of themselves will not produce increased efficiency.”
I strongly agree with this. My recommendations don’t tell hospitals what to do. Instead, they focus on setting the right incentives and providing the right information.