Is this treatment really necessary? – Daniel Morales crunches the numbers on ‘low-value’ Care
Nearly all medical treatments involve some level of risk; you might be one of the very few patients that has an adverse reaction to a medication, or to the anaesthetic used during surgery.
But many of the treatments that people undertake, are done despite evidence that shows they provide no benefit and may even cause unnecessary harm to the person receiving the treatment.
‘Low-value’ care can include tests that aren’t necessary, surgery that won’t help the patient and medicine that doesn’t deliver good outcomes. Not only is there potential harm to patients; low-value care wastes valuable health care resources and costs money which would better be spent on delivering care that works.
Once classic and well-known example is performing a knee arthroscopy on a patient who suffers from osteoarthritis, says Dr Daniel Morales Silva, a post-doctoral researcher who recently completed an important Digital Health CRC project with partners Lorica Health and Federation University.
Dr Morales says, “Some doctors will still do this routinely for people who suffer from knee pain, it’s quite costly and the evidence now shows that it actually makes things worse for the patient.”
Dr Morales holds two PhDs – the first in pure mathematics, from the University of Ballarat, and the second in Health Informatics which explored healthcare provider performance and risk adjustment, using data from Medibank.
His Digital Health CRC project reviewed various publicly-available risk adjustment models and then identified data science techniques to allow comparisons between clinicians, across healthcare delivery sites and around different conditions, interpreting ways to assess quality, cost and outcomes.
Low-value care complexities
Treatment effectiveness is more complex than simply ‘it works’ or ‘it doesn’t work,’ he adds.
“Risk can be patient-driven – for example, a patient who is over a certain age or who has a complicating chronic health condition such as diabetes or obesity, may face risks of harm if they undergo certain medical treatments,” Dr Morales explains.
The project involved a redevelopment of an existing data dashboard which analysed claims records, bringing more complex and granular capabilities.
Dr Morales says that the project has developed ways to analyse the data so that it can access many more procedures, sorting these by their medicare benefit schedule numbers as well as the diagnosis-related groups. This can identify areas where more ‘low-value care’ procedures are occurring – and can be reported on not just by the type of treatment, but by care location (such as a hospital) and by individual clinicians.
In one example, a data analysis of nearly 10,000 knee arthroscopy procedures – with total benefits paid out of around $98 million – found that nearly 3,000 procedures (or 32 per cent) fit the broad definition of low-value care.
Potential patient harm is the most concerning aspect of low value care – but financial drivers via the private health insurance system may also be a very effective way to prevent unnecessary interventions.
Dr Morales says that an analysis of the 26,000 hysterectomies that were performed in both public and private hospitals over the 2018-19 financial year showed that many of these procedures should not have gone ahead.
“One private health insurer recorded 1,228 hysterectomies during the 2019 calendar year, and at least ten per cent of these would be classed as low-value care,” says Dr Morales. He says the insurer paid over $1.3 million in benefits alone for these low-value care procedures.
“This shows how useful the data can be in reducing significant waste and expense,” he says.
Clinician Reaction
During the project the researchers interviewed clinicians and explained how the app could be used to understand patient outcomes.
“It was interesting to see the different reactions,” Morales mused. “Some clinicians were very concerned about the way the data is aggregated, and by the implications, but others were very receptive and really liked the opportunity to have a very clear way to determine the outcomes of their own practices.”
The most significant barrier to using this kind of data to make clear, evidence-driven decisions about patient choices and outcomes is not a technical one, he says; it’s a cultural shift.
“We already have risk adjustment systems that are available for free use by clinicians, which help predict complications that a patient might face with particular treatments – and they are not universally used or understood,” he says.
“Doctors are very apprehensive about publishing their data and using it to compare different patients, I hear the argument – we are not comparing apples with apples, some patients are more complicated – but the whole point of risk adjustment is to cater for that,” he says.
This project has tied together both the financial and clinical outcomes of low-value care. The outcome will allow administrators to access health comparison data and to group sets of information to demonstrate the value of different procedures to patient and the health organisation.
This data could be the nudge that is needed, to encourage more widespread understanding and adoption of data-driven decision making in our health system.
