The effectiveness of inpatient consultation-liaison psychiatry service models: A systematic review of randomized trials, 2021, Toynbee, Sharpe et al

Abstract
Objective
To systematically review randomized trials of the effectiveness of inpatient Consultation-Liaison (C-L) Psychiatry service models in improving patient outcomes, reducing length of hospital stay and decreasing healthcare costs.

Method
We searched databases including Ovid Medline, Ovid Embase, Ovid PsycINFO and EBSCO CINAHL for relevant trials. Two independent reviewers assessed articles and extracted data. The review is registered with PROSPERO, number CRD42019120827.

Results
Eight trials were eligible for inclusion. All had methodological limitations and all were published more than ten years ago. None reported clear evidence that the C-L Psychiatry service model evaluated was more effective than usual medical care alone. All the service models tested focused on providing a consultation for patients identified by screening. Clinical heterogeneity precluded meta-analysis.

Conclusion
Whilst we found no evidence that any of the inpatient C-L Psychiatry service models evaluated is effective, the sparseness of the literature and its methodological limitations preclude strong conclusions. The trials do, however, suggest that purely consultation-based service models may not be effective. A new generation of robust clinical trials of a wider range of C-L Psychiatry service models is now required to inform future service developments.

Paywall, https://www.sciencedirect.com/science/article/abs/pii/S0163834321000530

 

Interesting reasoning. Consultation-liason psychiatry doesn't seem to work. Since we cannot definitely exclude that it might work, we need a new generation of clinical trials testing a wider range of consultation-liason psychiatriy service models.

Normally one would consider the apparent failure of previous attempts a sign that it's not a promising direction.

 

But, that's literally a BPS feature, they deal with biopsychosocial "symptoms" and "causes" whatever the diagnosis, even without a diagnosis. And it rarely stops them from doing meta-analyses anyway, there are several. It's basically the new thing, to compare disparate "interventions" for effectiveness, even including "interventions" that have the exact opposite aims.

I question the seriousness of recognizing that the evidence is weak but that the quality of the evidence is so bad the conclusion cannot be trusted. Which is typical, when BPS research disproves the assumptions they never trust their own results, always assume there must have been a flaw somewhere. Which is an amazing feat of inability to self-reflect and a keen observer would notice that this is the exact opposite of the scientific method.

I don't know on what basis this would warrant trying again, wider or narrower? Especially as the promises were always over cost savings. There has never been another goal in mind and this is a relatively objective one, at least here relying on hospitalization lengths (nevermind the perverse incentives this sets).

But one thing I am puzzled is that Sharpe is currently doing such a trial, at least last we've heard. So what is he doing there reviewing other trials? Especially this late into running an identical one, especially recognizing they are all over a decade old. That can't be right, people who run trials shouldn't be in the business of reviewing trials, there should be some separation.

Also: is there an actual difference between the BPS model and the so-called L-C model? It sounds exactly the same, its thought leaders are literally the same people and the overall quality is also the same. As are the premises and the promises. And everything else, frankly.

It really should be recognized that cost-savings is the only objective here. It states here there is no evidence for it, even though this very model has been used with that intent for decades by now, just with a different name. It was sold and bought on that basis, which past evidence contradicts, although I guess the intent here is to argue that this is what the problem is, that it does work, it's just that past evidence showing otherwise is too poor to be trusted.

This field of study is legitimately regressing with time, it's just getting worse and worse and including evermore politics and controversy. And no one dares question it. How insane is that?

 

There must be a reason for doing this now. Maybe they want the approach discredited so that their CBT looks better or to get more money... whatever. You can be sure patient care is not the priority.

 

Yes I wondered if there was a hint of better-to-do-something-than-nothing "logic" motivating this.

 

Did they use GRADE and still come up with a null result ?