Abstract
There have always been a range of treatments for mental illness. Evidence exists of ancient trepanning, and through the ages other techniques have been used such as blood letting, exorcism, confinement, dietary interventions, environmental interventions, talking therapies of various modalities, industrial therapies, insulin comas, and ice baths among many others. In the past the view was held that physicians were people of such sober judgement and fine knowledge that ‘all remedies whatever are at the disposal of practitioners to reject or employ them under the sole guidance of their own judgment’ [1]. But in more recent times, for reasons of efficacy and economics, there has been greater concern with using treatments for which there is the best evidence of a positive outcome.
In order to achieve this it is first necessary to know what works and what is less likely to produce the desired outcome. Despite it now seeming like the term ‘evidence-based medicine’ has been around forever, it was only in 1992 that it was used for the first time [2]. The article in which it was used described a new approach that would be based on clinicians being able to access and appraise the literature in order to determine the treatments with the most evidence and then to use this to guide their practice. As more information became available, the volume of it soon overwhelmed the ability of an individual to synthesize. Shortly afterwards the Cochrane Collaboration was established to systematically gather and publish evidence related to health, and in the mid-1990s clinical practice guidelines began to be developed. Between 2003 and 2005 a range of clinical practice guidelines for Australia and New Zealand were commissioned by the Royal Australian and New Zealand College of Psychiatrists and published in this Journal, for a range of disorders [3–8]. These guidelines described pharmacological, behavioural and psychological interventions for the various disorders that were supported by evidence.
Through the 1990s and into the early years of the current decade there was a great effort to collate data and issue statements about which treatments worked. For example Gaebel et al., in reviewing guidelines for schizophrenia, found 27 sets of guidelines from 23 countries [9]. So after all this effort, are evidence-based treatments based on guidelines being implemented? Despite the massive effort involved in the collection of data and construction of systematic reviews, meta-analyses and practice guidelines over the last 15 years there is evidence that recommendations of guidelines for both psychosocial and pharmacological treatments are not routinely translated into practice [10]. For example, in a national survey in the USA, West et al. found that:
rates of conformance with the [PORT] guideline recommendations were significantly lower for psychosocial recommendations than for psychopharmacologic recommendations: for psychosocial recommendations rates ranged from 0 percent to 43 percent, whereas for psychopharmacologic recommendations rates ranged from 30 to 100 percent.’ [11]
Similarly surveys of people with mental illness show that very few of them are gaining access to evidence-based treatments [12], while other studies have found that the level of prescription practices that adhere to guideline recommendations is low [13], [14].
The real question, perhaps, is no longer ‘What works for whom?’ as Roth and Fonagy asked [15], but rather, why we do not implement what we know works. Surprisingly, there is very little research that considers the degree to which guideline recommendations are implemented [16], [17], or where they are not, what barriers prevent adherence to them.
Evidence-based interventions have the potential to produce better outcomes as well as save money. In order to try and identify areas where there has been a failure to implement evidence-based interventions in an Australian and New Zealand context, and to identify barriers to implementation, the Australasian Society for Psychiatric Research (ASPR) established a Knowledge Dissemination Group (KDG). A key task of this group in its early stage has been to identify barriers to the implementation of evidence-based interventions. In order to do this the KDG has collected examples of failure to implement evidence-based interventions. ASPR members who nominated these examples were asked to seek input from other expert colleagues in order to identify barriers to the implementation of the interventions. This editorial discusses the process that was followed, provides examples of areas where there has been a failure to implement evidence-based interventions, and discusses the barriers that were common to a number of areas. The purpose is twofold. First, to identify and begin a discussion of the barriers that impede the application of evidence, and to initiate a wider conversation about this. Second, to propose some solutions that may be the basis for action to reduce the barriers.
Process
In the middle of 2007 the KDG, through the membership of ASPR, solicited examples of areas where an evidence base existed for effective treatment, but where that treatment was not being routinely implemented. This resulted in five areas being submitted: (i) non-pharmacological interventions for antipsychotic-induced weight gain [18]; (ii) vocational recovery in mental illness [19], [20]; (iii) brief interventions for alcohol use and misuse [21]; (iv) family-based interventions in schizophrenia [22]; and (v) evidence-based treatment of post-traumatic stress disorder [23], [24]. As part of this solicitation process, the researchers who responded were asked to identify barriers that they felt existed to the implementation of these interventions. Two of us (EK, AFJ) then collated these into four thematic groupings: patient barriers; mental health practitioner barriers; other practitioner barriers; and health system barriers. The five respondents then met and further worked over these themes. This resulted in two further barrier areas being identified (Table 1). In total there were 41 identified barriers over six broad themes.
Barriers and consensus ratings for five areas
EBI, evidence-based interventions; PD, professional development; PTSD, post-traumatic stress disorder.
Each of the five respondents then contacted other experts in their field and rated each of the potential barriers that had been nominated by the group, arriving at a consensus (Table 1). Barriers were rated on a 4-point scale: 0, no barrier; 1, mild barrier; 2, moderate barrier; and 3, major barrier.
In order to identify barriers common to the range of areas each item was given an average score, which was the sum of the ratings across the five areas divided by five. Barriers that had an average score ≥2 were considered significant barriers.
Findings
There were 16 barriers that were identified as being significant to the implementation or utilization of evidence-based interventions. Interestingly, the only area in which there were no significant barriers identified was that of knowledge dissemination. This would indicate that clinicians are able to access information about the evidence base.
Lack of knowledge
Lack of knowledge was a significant barrier in three categories, possibly each with a slightly different focus. For patients, lack of knowledge about interventions was identified as a barrier. This is consistent with other discussion of this issue recently [10]. Another barrier in relation to patient knowledge was knowing where to go to get help, or who to get it from. Independent support for this barrier comes from community surveys showing significant gaps between what the public believe about treatments for mental disorders and what the evidence shows is effective [25].
Among both mental health clinicians and other clinicians (e.g. general practitioners), lack of knowledge in the form of skill or training was rated as a significant barrier. When taken in conjunction with the dissemination of knowledge not being a barrier, this may indicate that there is an absence of training when new interventions are developed, leaving people feeling unable to implement them in their practice. Alternatively, it may be that there is training, but clinicians are not accessing it. Possible reasons suggested by our experts are that clinicians feel that guidelines for treatment threaten their clinical autonomy, or they lack the time and resources either to be trained or to adapt their practice to incorporate the new intervention.
Health service barriers
The area in which the most barriers were identified was health service barriers. In this area nine of 11 items were endorsed across the five intervention areas as being significant barriers to implementation of interventions. The only two that were not were poor leadership (and it is good to know that that is not itself a barrier) and generic models of case management. All other items in this category rated significantly.
Lack of access
Given that clinicians lacking the necessary skills and knowledge is a major barrier to implementing interventions and that health services have barriers to providing services, it is not surprising that lack of access to evidence-based interventions was seen as an important barrier for patients.
Societal stigma
While stigma of help-seeking was not rated as a significant barrier, stigma in society as a whole was. Societal stigma may contribute to many of the barriers mentioned here, such as preventing widespread knowledge about what help is available, and where, and reducing pressure on policy makers to fund appropriate services.
Potential solutions
This process has been undertaken in order to identify barriers to the implementation of evidence-based interventions in mental health in Australia. The present results demonstrate that there are two key areas that are barriers to implementation. These are lack of knowledge and skills; and the health system.
Knowledge is possibly the more easily addressed area of these two. The National Institute for Clinical Excellence (NICE) in the UK issued guidelines for the treatment of schizophrenia in 2003 [22]. These are acknowledged as being the best set of schizophrenia guidelines in the world [9] and yet it was found that the rate of implementation of the guidelines was low [16]. The response to this was to design a more structured way of implementing the guidelines with ongoing evaluation to assess how this was progressing. In following this lead, one proposal is for the funding of not just the guidelines, but an implementation process that would include the provision of training in techniques that are supported by the evidence. Of course this evidence would need not only to show symptomatic improvement, but also an economic imperative that would add to the argument for the standard implementation of the technique. Increasingly, it would be advisable for researchers of psychosocial interventions to be as mindful of economics as the developers of pharmaceutical interventions are. Where funders believe money will be saved, it is conceivable that they will be more likely to sponsor training in that technique.
The health system is a much more difficult target to address, and yet it was the area that most rated as a barrier to the implementation of evidence-based interventions. A number of the barriers relate to fractures in the system of care: for example, the difference (and possibly distance) between alcohol and drug services and mental health services, or mental health services and employment services. Another barrier relates to the philosophy of a service taking precedence over the implementation of new treatments. The barrier created by different funding sources (e.g. state vs federal government in Australia) may prevent the development of new interventions in a service funded by one funding source where the intervention is seen as the responsibility of the other funding source. This area by itself would be a fruitful one for discussion about reducing barriers to implementation.
A final barrier identified in the health system area is lack of provision of information to empower patients. A recent article by Rosenman et al. proposed that one of the solutions for implementing evidence-based practice would be a more informed group of patients who would confront clinicians with their knowledge and demand evidence-based interventions [10]. This was in response to their finding in a straw poll of colleagues that, although most were aware of the guidelines, few had read them. A basic problem still remains, which is, from where do the patients get good unbiased knowledge with which to confront their clinicians? As noted in that and other articles, guidelines quickly become dated and it is sometimes hard to know the provenance of information on the internet.
A potential solution to many of these barriers may lie in the mechanisms for determining public funding of interventions. It is interesting to note that all of the examples put forward of failures to implement evidence-based interventions concerned psychosocial rather than pharmaceutical interventions. This may indicate that the process for implementing new pharmaceutical therapies is clearer than is the case for psychological and psychosocial interventions. Australia has a formal mechanism for approving the funding of new pharmaceuticals, the Pharmaceutical Benefits Advisory Committee, which formally examines the costs and benefits of new products before deciding whether their cost should be subsidized by public funds. There is no equivalent mechanism for the review of evidence on psychosocial interventions and for linking the evidence of benefit to the provision of public funding. There may be merit in considering a parallel mechanism for formally considering the evidence on psychosocial interventions and providing Medicare funding only for psychiatric and psychological services that involve evidence-based approaches.