The measure of success for compliance letter campaigns is often a reduction in prescribing volume, rather than appropriateness of care. That’s not very patient-focussed now, is it?
A new Australian study of “nudges” away from low-value care found that interventions made very little difference to clinician behaviour, sparking questions about what exactly compliance campaigns aim to achieve.
The cluster randomised trial study, published in the Canadian Medical Association Journal last month, looked at clinician and patient behaviour across eight NSW public hospital emergency departments.
Although the study primarily looked at emergency departments, GPs have long been the primary target of so-called nudge letter compliance campaigns, wherein the Department of Health, Disability and Ageing sends out letters to a group of practitioners it has deemed as high-volume prescribers.
There were two nudge-style interventions, both aimed at reducing low-value prescribing and imaging requests for people presenting with low back pain.
One intervention was patient-facing and consisted of an educational poster, and the other was a clinician-facing pop-up alert within the electronic health record which was triggered by imaging requests and opioid administration.
Throughout the intervention period, though, there was actually a slightly higher prevalence of opioid prescribing at discharge and a roughly similar level of unnecessary imaging.
Lead author Gemma Altinger, a health sciences researcher and PhD candidate at the University of Sydney, told The Medical Republic that these results highlighted flaws in the existing evidence base.
“A lot of the existing examples of nudge interventions in healthcare settings, particularly from the US, generally show that these sorts of interventions do have an effect, but I think what that highlighted to us is the difference in the quality and the appropriateness of that evidence,” she said.
“… A lot of the evidence has tended to be before and after studies, not these cluster randomised control trials, which bring a much higher quality of evidence, particularly if you’re thinking about how you would scale this across multiple health services.”
The endpoints for most existing studies, Ms Altinger pointed out, look at the reduction in the total volume of imaging or the total volume of prescriptions.
“Very few of them – none that I’ve seen – are looking at the appropriateness of that test or treatment for the individual patient in front of them,” she said.
“So I think when we apply these more rigorous lenses to [nudge interventions], it’s a different sort of outcome we’re looking at.”
Related
A more specific example of this, she said, was when a previous compliance campaign successfully led to a reduced volume of opioid prescriptions, but a concomitant rise in gabapentinoid prescriptions.
“[Compliance campaigns] might be reducing one thing, but are they increasing another, where in a few years’ time that’s going to be the thing, rather than focusing on what does the patient actually need, how are we improving the quality and the safety of care, and how are we supporting clinicians to be able to make that judgement call in the moment?” Ms Altinger said.
Furthermore, most previous randomised control trials of clinician-directed nudges have also focussed on the implementation of high-value care, rather than the de-implementation of low-value care, the article noted.
Fewer than half of the clinicians on the receiving end of the electronic patient record notification could recall seeing the nudge alert, even though the research team had gone to efforts to reduce alert fatigue by ensuring that the opioid-prescribing alert was muted for repeat requests by the same clinician.
“It was triggered only for patients who presented with the triage of back pain, but even that lacked sensitivity and was limited by the EMR,” Ms Altinger said.
“We’d be able to increase the sensitivity of this by introducing additional variables, like if the patient was over 65 years old, then NSAIDs might be contraindicated and, for them, opioids may be the more appropriate choice.
“Or for someone who had a fracture or had an impact or trauma, for them imaging might be more appropriate.
“Those alerts were still triggering for those patients, which may have reduced the effectiveness of them.”
