evidence based health care

When and how should new therapies become routine clinical practice?


Following on from my last post about when to adopt new therapies – a wonderful colleague of mine (who shall remain nameless, but You Know I Know Who You Are) sent me a copy of this paper from a physiotherapy journal. Bo and Herbert argue that the current way that new therapies become integrated into our daily clinical work is ‘far from optimal because innovative therapies still become accepted practice on the basis of laboratory research alone.’ I agree. Worse still, old therapies that have little evidence to support them continue to be used – even in the face of clinical studies demonstrating that they have no greater effect than placebo.

Bo and Herbert suggest there are several ways that new therapies are adopted within physiotherapy practice. I suggest that there is little difference between the situation in physiotherapy and other health professions!

Clinical experience – this is the traditional way practice in health care evolved. Experienced therapists hand down ‘what works’ on the basis of their observations within their own practice. Sometimes this is more formalised within workshops or conferences, and sometimes case reports are published. The basis for adoption is mainly on the reputation (or charisma) of the founder of the method. This situation continues today within most health professions, but fails to account for biases that are present in clinical practice. Things like patient expectation, natural course of the disorder, placebo, reduction of distress and failing to control for other external sources of influence mean the ‘expert’ can be mistaken. There are some ways to ameliorate these biases such as using good outcome measures and making time for long-term follow ups, but clinical experience alone is insufficient to make good generalisations about ‘what works’.

Theories based on basic science or ‘preclinical’ research (laboratory findings)
In this paper, Bo and Herbert suggest that as physiotherapists began to develop an academic arm, researchers based their experiments on their knowledge of the basic sciences underlying the profession. Often the laboratory-based experiments lead almost immediately to alteration of clinical practice – alternatively, laboratory findings were used to justify existing practice if the results obtained were consistent.

The problem is that while laboratory findings can inspire further study, perhaps leading to the testing of new hypotheses, findings don’t directly translate to what occurs in a real clinical patient. Patients don’t conform to the very tight parameters required for experiments. There are multiple variables involved in their clinical presentation that may be quite important but too complex to be incorporated into basic science experiments. Bo and Herbert suggest that laboratory studies measure impairment-level outcomes, while treatments need to measure disease-specific or disability/quality of life changes. Clinical practice is very different from a laboratory!

The ‘gold standard’ for identifying whether an intervention provides an effect in real people is the randomised controlled trial. When they are well-designed, they can be used to clearly demonstrate that treatment X and only treatment X influences the outcome. I won’t detail why the RCT is such an important method today, only to say that it remains the best way to ensure extraneous variables are controlled for – but they’re expensive, time-consuming and difficult to conduct in a clinical setting. It’s not surprising that I don’t think I’ve ever read about RCT’s for raised toilet seats!

Bo and Herbert go on to describe a six stage protocol to be used when a new therapy is being considered.
Stage 1 – clinical observation, laboratory studies – development phase
A clinician or researcher observes something interesting. Preliminary studies demonstrate that this ‘interesting finding’ is able to be consistently obtained. Some hypotheses are developed and tested in a controlled environment.
Stage 2 – clinical exploration – development phase
The hypotheses are explored in a clinical setting, maybe a prototypical treatment is carried out amongst volunteers, trial and error (or ongoing hypothesis testing based on a theory or model) shapes the treatment. This is a strongly exploratory phase – and quite exciting!
Stage 3 – pilot studies – development phase
Once one or two specific interventions become confirmed, pilot studies in a controlled but clinical environment are conducted. These might be small-scale group studies, case series studies, or small RCT studies. Further refinement of the practicalities of this treatment is made.
Stage 4 – randomised clinical trials – testing phase
If pilot studies are ‘promising’, RCT’s are carried out. A single swallow does not a summer make, however, so one study is rarely sufficient data on which to base wholesale adoption of a new treatment. Replications with different settings, different clinicians, different patients need to be made.
Stage 5 – refinement – refinement and dissemination phase
This stage may involve ‘head to head’ comparison of the new approach with other, more established treatments. Larger RCT’s are needed to identify how subgroups respond to the intervention.
Stage 6 – active dissemination
Now the word can be spread using guidelines, teaching curricula, continuing education – other professionals will learn about it, patients can be informed of the option and even the general population can be advised.

Now I have absolutely no argument with this staged approach to developing therapy. I will, however, suggest that there are very few physiotherapy, occupational therapy, or even psychology interventions that have reached Stage 5 before the treatment is already adopted and described as ‘evidence-based’!

Lack of research expertise notwithstanding, obtaining funding for these studies is difficult. The hoops that need to be gone through, at least in New Zealand, to prepare a research proposal for even a Stage 1 or 2 study makes it challenging for therapists in full-time clinical work to contemplate conducting even basic observational studies. There is an inherent lack of interest from managers of health services to allow clinicians to spend time on non-treatment-related activities, it simply doesn’t pay. Allied health clinicians rarely have the expertise to carry out methodologically strong studies without requiring support from within the profession, within the clinical adminstration and management, and without support from an academic institution.

I’ll jump off that soapbox quickly now!

Where does that leave clinicians when thinking about adopting a new therapy? More tomorrow on what to say to patients, but in the meantime I want to leave you with this thought: until an RCT is able to demonstrate that treatment X has an effect attributable only to itself, and is applicable to the kind of patients within the kind of setting the therapist is working in, all treatments are really ‘experimental’.

BO, K., & HERBERT, R. (2009). When and how should new therapies become routine clinical practice? Physiotherapy, 95 (1), 51-57 DOI: 10.1016/j.physio.2008.12.001

When should we adopt a new therapy?

When should we adopt a new therapy? This is a vexed question for me. New therapies should be more effective, less time-consuming, have fewer negative effects or take less effort than old – or at least offer something positive – before they’re adopted.

I’ve been critical of the way new therapies have come and then gone in pain management to the point where I’m just a tiny bit hesitant to try a new approach until… well that’s the point of today’s post!

There are a few factors I find myself using to review a novel therapy.

The first is to establish whether it really is a new therapy – perhaps what is being called ‘new’ is merely rebranding. Often I’ll go to Cochrane Reviews or similar to see whether the treatment has been reviewed before under another name. The ‘gold standard’ for evidence-based practice is that there have been a good number of randomised controlled trials, preferably double-blinded, that have replicated strong effects in different populations from different settings. This level of evidence is pretty difficult to find in psychological therapies, and I’ll usually read a Cochrane Review only to find the authors asking for ‘more research’! Despite this, it’s surprising how often I’ll find that a ‘new’ therapy has been included in the past, and there are some useful comments within a Cochrane review.

The next thing I’ll look at is whether the new therapy is simply an old therapy being transferred from one disorder to another – a good example of this is the implementation of graded exposure for pain-related anxiety and avoidance. Graded exposure has been widely used in anxiety disorders. With the adoption of a similar theoretical model of avoidance in chronic pain, transferring the same approach from one disorder to another is relatively straightforward, with provisos. It can provide a way to test hypotheses about the mechanisms operating in a person’s presentation – but I’ll usually want to see some randomised controlled trials, or at the very least some clinical trials of the approach before I want to use it on a regular basis.  I don’t really want to experiment on my patients.

If it’s a really novel approach – for example, motor imagery for complex regional pain syndrome – I want to read some of the original research and review the proposed mechanisms the new approach is thought to be using. I’m wary of adopting a new therapy when there is only one researcher publishing findings, or when the clinical group it has been used with is different from the people I see. Anecdotes, or case studies, are the weakest level of evidence on which to base treatment decisions.

Some resources
If you’re new to thinking about evidence based health care, this website by University of Minnesota offers a two-part tutorial. Bandolier website offers a huge range of topics on evidence based health care, and the most recent update includes information on neuropathic pain and also fibromyalgia.

Cochrane Collaboration has the reputation as being one of the most critical review processes, although individual reviews vary in quality – overall, however, their reviews are well-regarded and worth a look first.

To simplify your search, this website SUMSearch, is provided by UT Health Science Center, San Antonio, and trawls through a number of search engines to find evidence based information on health care.

For some great links to internet databases, go to (the) health informaticist – and read their regular blog posts too for some thought-provoking and often chuckle-making commentary on things informational on the net.

**Quick rant**

What really bugs me is when there is a good level of evidence to suggest that a specific intervention has no effect or has a limited effect, therapists continue to recommend it!  A case in point is the use of splints for rheumatoid arthritis.  Another is back braces.  And another is carrying out ‘safe manual handling’ training to prevent back pain.

A therapist once said to me ‘if we stopped using all the things that don’t have evidence, we wouldn’t have anything we could do’ – erm…that’s the point I’m trying to make!  Perhaps not intervening at all is a more helpful option than continuing to do something that has no effect.

**End of rant**

More on science and therapy this week – if you’re not keen on science, be afraid, be very afraid!

On evidence and practise

An opinion piece to restart my blogging after my lovely holiday…

I’ve been reading ABC Therapeutics blog where Chris Alterio writes in response to a long comment by Michele Karnes suggesting that occupational therapists (and by inference all health care providers) ‘should be made aware of treatments that are offered to clients/patients, whether it is traditional or non-traditional, a long existing treatment or new one. This enables our OT profession and professionals to better educate the people they treat and interact with.’

I don’t have any particular concerns about this part of Michele’s comment – but I do have a problem with this part ‘while Evidence Based Practice is on all of our minds, and ultimately the best to utilize with our patients, if we only used treatments for all of these years we would have missed out on the many treatments that OT’s have historically (and still) use.’ (my emphasis)

It raises some concerning things for me – and while I don’t have answers for all of my concerns, I hope to stimulate some debate at least.

Chris writes in his blog ‘Just because people seek out alternative energy healing interventions doesn’t mean that it constitutes appropriate or ethical practice. In an article published in the Journal of the American Medical Association on this topic an author writes: “Given the extensive use of CAM services and the relative paucity of data concerning safety, patients may be putting themselves at risk by their use of these treatments. Only fully competent and licensed practitioners can help patients avoid such inappropriate use... Physicians can also ensure that patients do not abandon effective care and alert them to signs of possible fraud or danger.“‘

I’d add that licensing in itself does not inevitably lead to patients being helped to avoid inappropriate treatments. I also add this:

I think it also takes a critical and educated mind, a systematic approach to reviewing evidence, and considerable determination not to be swayed by forceful opinion. (more…)