Ever since short-stay mental health units were introduced in the early 2000s, there has been speculation about whether reduced bed days are in patients’ best interests, or a proxy for insurer cost control. And while both arguments have merit, Mark Sweeney of Avive Health claims the debate is less binary than many assume.
Mark says that while short-stay units can deliver effective psychiatric care, it is likely – and somewhat valid – that insurer incentives prevail in this context. He argues that while other health domains have standardised diagnostics and treatment options, the lines in psychiatry are more blurred, adding an extra layer of risk to insurer funding structures.
“In defence of the insurers, I’m not sure what else they could do here. If you put yourself in their position, they’ve been asked to fund a category of care, mental health, where the product is not standardisable. The providers do not or cannot share comparable outcome data. And that cost curve for them is rising faster than virtually any other health category.
“When you look at what levers they have, they can’t manage the cost of outcomes because they can’t see them. They can’t manage costs through case mix because nobody reliably reports it. They can’t manage costs through clinical pathways because they vary from provider to provider and psychiatrist to psychiatrist. So, they fall back on the only two signals they can reliably see: length of stay and readmission rates.
“So, in my view, there is no doubt that short-stay mental health units are being driven by insurers, but it’s an entirely predictable response to what they’re funding.”
Cross-sectoral comparisons
In contrast, modalities like orthopaedics see most patients presenting in the same way each time. The conditions are definable, the treatment pathways are standardised, the recovery curve is more predictable, and the outcome is more easily measurable.
“When an insurer funds a knee replacement, they know exactly what they’re buying and they can benchmark it,” Mark said. “They’ve got thousands of comparable cases, they can track readmission rates with confidence, and they can easily price an episode because it’s generally standardised.”
Meanwhile, the same diagnostic label in psychiatry can present in varying ways and respond differently to treatments. This is in part due to the heterogeneity of conditions like depression, which one study found to have more than 1000 different symptom combinations, leading some to question the validity of depression as a diagnostic category.
“Somebody with depression could be ready for discharge within seven days, while another could remain treatment resistant for 21 days, before commencing a three-week CMS protocol. A psychiatric patient never presents to us as a seven-day patient or a 21-day patient. We just don’t have that level of predictability,” Mark said.
Psychiatrists often disagree
On top of that, diagnostic discrepancies between psychiatrists pose an extra layer of risk to insurers. On Avive’s electronic medical record, clicking on a diagnostic category shows the DSM-4 diagnosis, along with all subcategories in between – and these can be a source of disagreement for psychiatrists, Mark highlighted.
“We’re relying upon the psychiatrist to give a diagnosis. But even between psychiatrists, there are various opinions about which subset of depression or anxiety a single patient is presenting with.”
What does this mean for short-stay going forward?
Mark claims standardising mental health could lessen financial risk and, in turn, allow improvements in short-stay psychiatric care. While this may be a long way off, he says big data and AI are helping.
“We’re in the process of designing AI to assist psychiatrists with diagnoses so that we can get more predictable responses. But even if we have the same correct diagnosis for two patients, they can still present completely differently. So, it’s certainly not a panacea,” he said.
Data is in its favour, but should be interpreted with caution
Regardless, Mark says the data is in favour of short-stay models, particularly those with planned admission, structured care, clear discharge pathways, and integrated follow-up. Support has also been shown for day programs, crisis stabilisation units, and community-based step-down models.
However, he cautions that supporting research often uses low-acuity patients who are more stable and treatment-responsive and don’t necessarily represent the broader short-stay patient population.
“The research doesn’t invalidate short-stay models, but we need to be careful about its broad conclusions. We can’t yet claim it is better than a long stay model when this selection bias is at play.”
Understanding the comparison with long-stay models is key in the short-stay debate, but Mark warns that the variability in mental health presentations poses a further challenge for its supporting research.
“In orthopaedics, the evidence accumulates cleanly because that intervention is consistent. You can run a randomised trial on knee replacement protocols and the variability between patients is so narrow that you can comfortably predict the outcomes. In mental health, again, that variability between patients from the same diagnostic label is wide.”
Funding reform is needed
Even so, Mark believes short-stay can “absolutely work” in psychiatry, but that it needs to be funded as the entire journey of that patient, not just short episodic care.
He highlights that longer stays can be costly under the current funding mechanism.
“Essentially, we lose money after day 21 if the patients in our hospital, but we also cannot dictate how long a patient should stay. We have some who have been in our hospital for 90 days and clinically, we have to respond to these patients – not the funding.”
Further insight
Sharing more on how short-stay reform might look, Mark will join an expert panel at the upcoming Short Stay Hospitals Forum, hosted by Informa.
This year’s event will be held 3-4 August 2026 at the Crown Promenade Melbourne.
Learn more and register your tickets here.