The survivability of dialectical behaviour therapy programmes: a mixed methods analysis of barriers and facilitators to implementation within UK healthcare settings

Dialectical Behaviour Therapy (DBT) [1] is a comprehensive cognitive-behavioural treatment originally developed for adult women who meet criteria for Borderline Personality Disorder (BPD), particularly those who engage in suicidal or non-suicidal self-injury. Traditionally, this client group  has been perceived as “treatment resistant” and considered unsuitable candidates for psychotherapeutic intervention [2]. DBT was the first psychological therapy to challenge the culture of therapeutic rejection for individuals with BPD and has become one of the best evidenced treatments for this client group.

Numerous DBT efficacy trials [3‐11] have demonstrated reductions in suicide attempts, intentional self-injury, anger, depression, hopelessness, and improvements in global functioning [12]. Recent meta-analyses have found moderate to large effect sizes indicating a beneficial effect of DBT when compared to treatment as usual on outcomes such as anger, parasuicidality, and mental health [13, 14]. Furthermore, several randomised controlled trials (RCTs) have examined the application of DBT with other client groups such as older adults with major depressive disorder, eating disorders, and forensic populations [15‐19]. Thus, the data on DBT clearly indicate its efficacy for the treatment of BPD and holds promise for a host of other disorders.

In 2009, DBT was included in the National Institute of Health and Care Excellence (NICE) guidelines as a recommended treatment for females with a diagnosis of BPD and a history of repetitive self-harm [20]. Since then, a number of healthcare providers within the United Kingdom (UK) have included the provision of DBT as a quality improvement indicator in an effort to meet national targets in health outcomes for individuals with serious mental illness [21]. Preliminary efficiency research also suggests that DBT has the potential to be a cost-effective treatment for individuals presenting with parasuicidal behaviour [22, 23]. Indeed, it appears that the potential benefits DBT has to offer is gaining traction within routine healthcare settings.

Notwithstanding NICE recommendations, demonstrable treatment efficacy, and potential cost efficiencies, concerns have been raised about the sustainability of DBT programmes within the UK National Health Service (NHS) [24]. Diffusion of Innovations Theory [25] suggests that innovations must be widely adopted in order to self-sustain. Widespread adoption of a new practice depends initially on innovators and early adopters and how quickly the subsequent late majority can be persuaded to shift. Furthermore, it is proposed that ideas not sustained by early adopters are unlikely to spread elsewhere [26]. Thus, effective implementation is relevant not only to long-term sustainability but also subsequent spread of an innovation.

Other factors that can impact sustainability are those directly related to the innovation itself, such as the ease with which it can be implemented and how well treatment effects observed in efficacy trials will generalise to routine healthcare settings. The DBT model entails a comprehensive programme that structures the treatment environment across different modalities to enhance client’s capabilities (skills training groups), improve their motivation (individual therapy), aid generalisation of new skills (telephone skills coaching), and supervise DBT therapists (a consultation team model) [27]. All of the treatment modalities are informed by a coherent theoretical model with associated therapeutic strategies based on cognitive behavioural principles and mindfulness [1, 28]. The programme is delivered by a team of mental health professionals all trained within the DBT model and the rationale for doing so is to alleviate the stress and anxiety of working with a high risk client group in which change is often slow [27]. Nevertheless, the requirement of a specialist trained team usually involves significant reorganisation of existing services and an ongoing commitment to delivering an intensive specialist intervention. This is likely to have an impact on how well DBT is implemented or, indeed, whether it is even considered viable for adoption within a service.

Deciding to implement a new practice is not a discrete event but a set of interactive dynamic processes. The difficulties of translating evidence-based research into real-world settings is widely acknowledged [29], which has led to a growing body of literature examining the various factors involved in the implementation and sustainability of evidence-based practices (EBPs) [30‐32]. Historically, more attention has been paid to the efficacy of interventions. Whilst such information might help a consumer or agency to select a particular type of intervention, evidence of efficacy alone does not lead to more successful implementation [29], in the same way that simply training practitioners in a new approach does not sufficiently ensure behaviour change [33]. Thus, transfer of innovation needs to be considered within organisational and wider system contexts to ensure that desired change is disseminated, implemented and sustained [34]. However, because organisational restructuring requires changes in service provider behaviour and transformation of systems, translating an EBP into routine practice remains an unquestionably complex and often daunting task.

A number of conceptual frameworks have been developed to aid the process of implementation [29, 31, 35‐37]. Whilst these frameworks differ somewhat on areas of emphasis and terminology, influences on implementation generally relate to the context (outer and inner), the innovation itself (fit, training, efficacy), implementation processes (planning, selection, evaluation), individual characteristics (motivation, skill), and sustainability factors (fidelity monitoring, penetration, outcomes etc.). These components are considered to be interrelated and a change in one may result in change to others. Therefore, due to the dynamic nature of healthcare systems and their external contexts, a given programme or practice may require more or less of each component at any one time in order to be successfully implemented. This represents a challenge for the implementation and sustainability of innovations, as the relative contribution of each component to overall outcome can change, resulting in the need for ongoing monitoring of processes. Such tasks can be greatly supported by the application of a guiding theoretical framework. Only recently have distinct models for sustainability of evidence-based programmes been produced [38, 39], however, most of the elements of these models (Inner and Outer Contextual Factors, Characteristics of the Interventionists and of the Intervention) are incorporated already in conceptual frameworks of implementation [32, 36].

Considering the above, implementing a comprehensive DBT programme in routine healthcare settings is unlikely to be a straightforward endeavour. Preliminary research into the sustainability of UK DBT programmes that underwent an intensive training programme between 1995 and 2007 confirmed that some teams had difficulty surviving [27]. Highest failure rates were found shortly after training ended (i.e. the second year of the programme) and again in the fifth year. Participants identified a number of challenges associated with implementing DBT in their service, which were generally characterised by an absence of organisational support. Conversely, for teams that had implemented successfully and managed to sustain, the presence of organisational support was identified as a facilitating factor.

In an effort to increase organisational support and promote effective implementation strategies, British Isles DBT (biDBT) have begun to offer an alternative training modality. Typically, training involves teams of practitioners participating in two five-day DBT intensive training events that are delivered off-site, which is known as the ‘open-enrolment route’. Each training event is separated by 8 months during which teams commence the process of setting up and starting a DBT programme. With the new mode, the content and structure of the training is the same; however organisations wishing to deliver DBT programmes are encouraged to host intensive training on-site. This requires a greater financial investment and consideration of how to adapt staff roles in order to successfully deliver treatment, with the idea that greater organisational investment will have a positive influence on the implementation process. This change in training delivery warrants further investigation to examine whether it improves implementation of programmes.

The aims of the present study are threefold: 1) to investigate whether early and late adopters of DBT have differential sustainability, 2) to investigate whether change in training method delivery impacts the sustainability of DBT programmes, and 3) to examine factors that act as a barrier or facilitator to implementation by using a theoretical implementation framework to guide assessment.

Consistent with earlier data [27], the highest failure rate for DBT programmes was observed in the second year post-training. Despite this early high failure rate the survivability of DBT programmes compares well with other evidence-based programmes reported in the literature. Cooper and colleagues [44] reported that 69% of delinquency and violence prevention programmes in a state-wide implementation sustained at two to four years post-initial seed-funding. Whilst the National Implementing Evidence Based Practices project reported that 80% of sites sustained at two years post implementation [45] and 47% at six years, although, in the six year data, sustainability rates varied between the five interventions studied from 25 to 69%. DBT compares favourably with these figures with survivability rates of 88% at two years and 69% at eight years.

Differences in the survival rates between the early and late implementers is not particularly surprising, although the different rates of censored data between the cohorts means that the result should be interpreted with caution. Several factors might account for this difference. Early adopters are known to be psychologically different from their peers and often in influential positions [46]. Whilst they may have adopted DBT early they may also have been keen to move on to the next innovation. Secondly, DBTs place as an evidence-based intervention within the UK became more solid with the publication of the NICE guidance in 2009 [20]. The advocacy for DBT within the guideline may have provided an ‘outer context’ support to teams training post-2007, just as publication of the guidance also boosted training in DBT [47].

Traditionally, the translation from science to practice has been a passive process that has usually only involved diffusion and dissemination of EBP information, with the hope that this is sufficient to change practitioner behaviour. There is a current shift towards a more active approach whereby outside experts work alongside organisations to help achieve implementation success and assure benefits to consumers [48]. Results from the present study found that on-site training did not increase the probability of survival. Survival curve comparison of training delivery methods indicated programmes trained off-site had a significantly higher probability of surviving. This is a surprising finding, given that on-site training was designed to increase organisational investment in DBT implementation. However, this finding must be interpreted with caution, as the amount of censored data between the comparison groups was found to be significantly different, limiting conclusions that can be drawn about differences between groups. Notwithstanding this caveat, a possible explanation for the differences may be that those attending off-site training have engaged in a substantial amount of pre-planning and assessment of organisational readiness, and in efforts to obtain management buy-in, have identified an explicit need for implementing DBT in their service setting. In doing so, they are possibly more likely to have actively considered how an implementation plan may be executed. Addressing organizational funding and resources and aligning the innovation with organizational goals are factors known to be associated with sustainability [39, 43, 45, 49]. Teams attending off-site training have typically had to actively pursue funding and gain agreements from their organisation to attend. This may indicate that individuals in teams pursuing this route may possess particular leadership skills that may also relate to sustainability [49, 50]. Attending off-site training provides greater opportunities to network with other teams, allowing for the sharing of experiences and ideas, which prove beneficial to implementation and sustainability. During the second week of training teams present their initial implementation efforts for consultation and feedback from trainers and fellow trainees. In off-site trainings trainees are exposed to a wider range of systems and witness trainers applying the components of the model to these different systems. This more expansive experience may increase knowledge of the core components and principles of DBT. Cooper and colleagues [44] similarly highlighted that greater knowledge of a progammes’ logic model increased the likelihood of sustainability.

Practitioner turnover and financing were the most commonly identified barriers to implementing DBT programmes. This is consistent with findings from other studies [43, 45, 50]. Indeed, these constructs may interact, as difficulty financing new team members was one of the main problems identified when practitioner turnover was high. Financing initial training was identified as a key barrier for some programmes. Although, a few overcame this difficulty by securing initial funding from external sources and then using evaluation and outcome data to secure ongoing funding from their organisations. Other programmes identified difficulties with ongoing financing, whether it was for training new team members, booster training, or accessing expert supervision or consultation. Whilst securing financing is a common theme both in this study and in others [43, 45, 50] consideration is rarely given to the costs of de-implementation and, in the case of DBT, failing to provide an intervention that may produce cost-savings [22]. Developing models that highlight the costs of failing to sustain may prove useful to influence leaders both in the inner and outer context or organisations to continue to support an evidence-based intervention.

A number of facilitators to implementation were identified. Most notably, all constructs within the Individual Characteristics domain were strongly endorsed as aiding the implementation process. A number of respondents reported highly motivated or skilled practitioners, effective leadership of the DBT team, or the presence of a DBT champion as key to overcoming barriers encountered to implementation and sustainability of programmes. This finding highlights how a strength in one or more areas can compensate for weaknesses in others [29]. Nevertheless, overreliance on an individual(s) to ensure effective implementation and sustainability leaves a programme particularly vulnerable to practitioner or leadership turnover. Organisations are dynamic and so the relative contribution of implementation constructs can inevitably wax and wane. This poses a difficulty for organisations because changes in one construct requires adjustments in others. Thus, successfully managing such changes will require effective monitoring and feedback systems to keep a programme on track [48], as well as ongoing availability of resources to do so.

Characteristics of the intervention, a feature in many implementation and sustainability models [31, 36, 38, 39] in particular the quality of the evidence base for DBT, was strongly endorsed as aiding the implementation process. Whilst efficacy data alone maybe insufficient for changing practice, findings from this study indicated that for some programmes research data played a crucial role in securing management commitment to delivering DBT. The quality of the evidence base may be of particular relevance during pre-planning and preparation stages, allowing for organisations to weigh up the suitability of DBT for their service and make as assessment of perceived benefits and ‘fit’ with the context [38]. For populations where the evidence base for DBT is not as extensive or robust, the lack of efficacy data may present a barrier to implementation. In this instance, the opportunity to trial a DBT programme and collect effectiveness data may prove beneficial.

Over half of survey respondents indicated that their programme engaged in practices which are considered pertinent to sustainability, with the exception of receiving supervision from a DBT expert. This is an encouraging finding and suggests that teams are aware of the need for continuous monitoring and collection of outcome data as an aid to sustainability [43]. Given that the highest failure rates for programmes are found within the active implementation stage (i.e. first two years), programmes should also consider identifying and monitoring implementation outcomes, distinct from service and treatment outcomes. Evaluation of implementation outcomes will provide an indicator of implementation success and yield an index of implementation processes. Also, because treatment effectiveness requires successful implementation, monitoring implementation outcomes is a necessary intermediate step to obtaining desired clinical and service outcomes [51].

There are a number of limitations to the study. The first being the small number of survey respondents from inactive teams, which prevented comparative analyses, and limits the conclusions that can be drawn from the findings. Second, the method of data collection prevented exploration of research participants’ interpretation of questions or the opportunity to clarify responses. Although a summary question was included at the end of each survey domain, not all respondents chose to elaborate their responses, limiting the amount of qualitative data collected. Lastly, the retrospective accounts from individual team leaders/members must be interpreted with caution due to problems inherent with self-report, such as post-hoc rationalisation. Future research should endeavour to recruit multiple respondents from programmes to reduce the likelihood of methodological bias, as well as recruit greater numbers of inactive teams to ensure a representative sample of respondents.

Despite these limitations, the present study possessed a number of strengths. There are few studies in the literature studying sustainability beyond the early stages of implementation (post-two years) and none, to our knowledge, that allow the comparison of two different types of training delivery that may have implications for sustainability. In addition, the use of a concurrent mixed-methods approach allowed quantitative findings to be complimented with qualitative information providing greater insight into the complexities of implementation and sustainability processes. The existing implementation literature utilizes a wide range of definitions and terminologies rendering extrapolation of constructs across settings difficult. By using the CFIR as a scoping tool to guide assessment of the barriers and facilitators to DBT, a number of constructs salient to implementing DBT in routine healthcare settings were identified, allowing for refinement of more relevant assessment tools for future research.

Successful implementation and sustainability of healthcare innovations into routine settings poses a challenge; DBT is no exception. However, since the onset of biDBT intensive training in 1995, the sustainability of DBT programmes has remained stable and similar to the rates of other innovations, and higher than others. Given the ever-changing landscape and finite resources of healthcare systems, this is an encouraging finding. Nevertheless, a number of programmes struggle to effectively implement and sustain DBT within their organisation. The particular adaptation to the location of training trialed here did not improve the probability of programme survival. Further augmenting on-site training with additional interventions for both inner and outer-context leadership [49, 50] could potentially improve the outcome of such training . A number of factors hindering or facilitating implementation of DBT were reported. Whilst these factors can vary between and within organisations, comparison with previous research suggests that the main barriers or aids to implementation have remained fairly consistent. Future research should include evaluation of predictive models that allow for testing the relative contribution of each implementation component, in order to identify what works in which contexts.