Background

Challenging behavior is prevalent in people with learning disabilities (Emerson et al. 2001) and has been demonstrated to have a number of negative effects on services, staff and service users (Allen 2009), including abuse (Emerson et al. 1994), placement breakdown (Allen 1989), and high staff turnover (Felce et al. 1993).

Positive Behavior Support (PBS) has been shown to be successful in addressing issues of challenging behavior and of achieving behavior change for individuals (Carr et al. 1999; Lucyshyn et al. 2002; La Vigna and Willis 2012). It came to the fore in the 1980’s & 1990’s and has grown from its origins in Applied Behavior Analysis (ABA), from where it has borrowed a range of assessment and intervention strategies, to incorporate principles of person-centred planning and normalisation (Carr et al. 2002; Dunlap and Carr 2009).

Some controversy remains about the definition of PBS, and whether it is in fact a separate discipline from ABA. For example, Carr et al. (2002) defined it as “an applied science that uses educational and systems change methods to enhance quality of life and minimise problem behavior” (p.4), and Horner et al. (2000) as “an approach that blends values about the rights of people with disabilities with a practical science about how learning and behavior change occur” (p.97). However others have questioned this definition and queried whether PBS is indeed a science, and in particular whether it is separate from ABA. Wacker and Berg (2002) stated, for example, “we are not sure that any definition of science could be used to categorize PBS as an applied science that is distinguishable from the scientific methods from which it is derived” (p. 25). Johnston et al. (2006) concluded in their examination of PBS that “it is difficult to characterize PBS in ways that will not invite disagreement” (p.69).

However, despite these controversies, there are a number of elements of PBS which are largely agreed upon within the PBS literature. In 1990, Horner & colleagues published what is regarded as the first clear definition of PBS—although at this point they called it ‘non-aversive behavioral support’. They emphasized 9 main characteristics: has an emphasis on lifestyle change; based on functional analysis; utilizes multi-element support; focuses on ecological changes; emphasizes antecedent control; teaches new skills; uses contingent and non-contingent reinforcement; minimizes use of aversive approaches; utilizes proactive and reactive approaches.

In 2002, Carr & colleagues added to this definition with their influential article “Positive Behavior Support: Evolution of an Applied Science” by providing additional definitions of the characteristics of PBS. These additional characteristics were defined as: lifespan perspective or a long-term focus; stakeholder participation; systems change; flexibility with respect to scientific practices and willingness to utilize other theoretical perspectives. Following the publication of Carr et al. (2002), the majority of PBS writings reflect these definitions (e.g. Dunlap 2004; Allen et al. 2005; La Vigna et al. 2005; Kincaid et al. 2006).

Despite the apparent effectiveness of PBS in addressing challenging behavior (Carr et al. 1999), its use remains limited in services for people with intellectual or developmental disabilities. In a study of 500 people with learning disabilities living in residential support in the UK, only 15 % were recipients of any kind of behavior programs (Emerson et al. 2000). It is not clear from this study if these programs were PBS-based; however the findings indicated the low use of any kind of behavioral program. Allen et al. (2005) propose a number of possible reasons for the low level of PBS use: the labor-intensive nature of interventions; a bias against ABA-type approaches in community care settings; lack of specialist staff trained in PBS; and no statutory requirement within community care settings for PBS to be provided for individuals with challenging behavior. A further reason for low level of use could be the lack of care staff skills in implementing PBS.

Training of staff teams to use PBS is therefore important for a number of reasons. Firstly, despite the effectiveness of PBS, restrictive and aversive practices continue to be used (Deveau and McGill 2009; Emerson et al. 2000). The risks associated with restrictive physical intervention are clear (e.g. Leadbetter 2002). Studies have also indicated a potential over-dependence on psychotropic medication to manage behavior. For example, in one study service users were three times more likely to receive anti-psychotic medication than behavior support (Emerson et al. 2000), despite a lack of evidence for its effectiveness (Tyrer et al. 2008). In comparison, PBS is non-intrusive and likely to be regarded as least restrictive.

Secondly, creating skilled and confident staff teams who can implement PBS plans with accuracy and consistency may enable individuals with challenging behavior to be supported in ordinary community care settings and therefore minimize the need for expensive, often out-of-area, specialist placements. It is likely that these models of support do not allow for the building of skills and competencies within the individual’s home environment and by the primary caregiver, most likely either a social care provider or family (DOH 1993).

Thirdly, creating skilled local staff teams is likely to be a positive alternative to relying on external specialist additional support teams. These specialist support teams who typically provide assessment and intervention input to individuals within their own environment are one alternative to specialist residential placements. These have been shown to be effective (Hassiotis et al. 2009) and are recommended in the UK government’s good practice guidance ‘Services for People with Learning Disabilities and Challenging Behaviour or Mental Health Needs’ (DOH 1993). However there are also limitations to the resources of specialist teams. For example, McClean et al. (2005) reported that only 48 % of people with severe challenging behavior are on team caseloads. The number of specialist teams required to provide PBS to all the individuals who might benefit from it is not likely to be available (Sprague et al. 1996). Training service providers’ direct care staff in PBS may therefore address issues of coverage.

The objective of this review is to summarize the results from published studies in the last 20 years regarding outcomes of PBS staff training in relation to either children or adults with learning disabilities. Studies which include either outcomes for staff (e.g. changes in knowledge or skills) and/or outcomes for service users (e.g. changes in rate or severity of challenging behavior, impact on quality of life, reductions in use of restraint) will be considered.

Method

Identification of Literature

Searches of Google Scholar and of electronic databases (Web of Science, Pub Med, and PsychINFO) were conducted in October, 2012 to identify relevant studies, published in English from 1990 onwards. In each database searches were made firstly for “positive behav* support”, and then with additional search terms such as “training” and “learning disab* OR intellectual disab*”. Following this, searches were also made under “challenging behav*” with the same additional search terms added in.

In addition to these initial searches, reference lists of studies found by the initial search were then accessed and their relevance assessed. Finally, Web of Science and PsychINFO were used for citation searches in order to identify further studies citing those already identified.

All studies were assessed for relevance by reading the description of the training provided. If the study described the training as PBS then the study was included.

Inclusion Criteria

Studies were included that described outcomes from PBS staff training. If the study identified itself as providing staff training in PBS then this was accepted. That is, the description of the training was not analyzed or assessed as to what exactly it covered; self-identification as being PBS was sufficient. The reason for this inclusion criterion is that, as noted above, there remains some controversy over definitions of PBS, in particular whether it is a separate entity from ABA. Thus, rather than attempt to define PBS & differentiate it from ABA (an exercise beyond the scope of this current study), it was decided to conduct this review based on studies that describe themselves as providing staff training in PBS.

Participants for the studies in this review were individuals with intellectual disabilities and challenging behavior, and/or the staff that provide their support. Studies only relating to outcomes for family carers were therefore not included, although some studies that report outcomes for both staff and family carers were.

Due to the limited number of studies found, criteria for inclusion were kept wide (i.e. any type of format or design of training in PBS; training of any length; any type of research design, including with or without control groups, with or without reliability measures).

No limitations were placed on type of outcome measures, whether based on outcomes for staff or for service users.

Included & Excluded Studies

The criteria stated that the term “Positive Behavior (al) Support” had to be used somewhere within the study to describe the training; this was most commonly (but not always) in the Method section where the training was described. However, Baker (1998), for example, did not mention PBS in his Method section where he included a table showing the content of the training sessions. However, within the introduction to the article, Baker stated that his purpose was to “provide…training…on positive behavioral support strategies.” Therefore Baker was included in the current review.

Some studies referred to PBS as part of the general approach of the agency or unit where the training took place. However, these studies did not meet inclusion criteria, as the described training was not deemed to be PBS-based. For example, Allen et al. (1997) discussed the use of PBS within their unit and referenced the La Vigna et al. (1989). However, the study described behavior management training with a focus on reactive strategies, and therefore it was excluded. Likewise, Allen and Tynan (2000) was excluded for similar reasons. Tierney et al. (2007) referred to the fact that they discussed the importance of using PBS plans within their training session. Yet the title of their training was “Understanding and Responding to Challenging Behavior.” Thus, this study was excluded, as it did not entail training in PBS.

Studies using closely related terms were not included (e.g., Grey et al. 2002; La Vigna et al. 2002). For example, Grey et al. (2002) use the term “multi-element behavior support” to describe their training. Although “multi-element behavior” is sometimes used to describe PBS (for example in McClean et al. 2005 & Dench 2005, the terms are used inter-changeably), PBS was not mentioned specifically within this article. Thus, the study was excluded.

Following the search process outlined above, studies were identified as being part of the review. In order to provide a reliability check, the second author independently carried out a check on the included and excluded studies using the specified criterion. There was 100 % agreement with the inclusion and exclusion decisions of the first author.

Results

Studies meeting the above criteria were summarized into a table with information about author, date of publication, country where research was conducted, participants involved, sample size, design of study, outcome measures, reliability and results. This information is summarized in Table 1, in chronological order starting with the oldest studies first.

Table 1 Details of included studies
Full size table

Although there is extensive research on generic challenging behavior training available, there is a limited literature published regarding outcomes of PBS training. 14 studies were identified by this review including 4 from Ireland, 5 from the US, 3 from the UK, 1 from Canada, and 1 from Australia. Table 2 shows the length, format, and content of the training.

Table 2 Length, format & content of training
Full size table

The studies vary as to whether they focused on staff outcomes (6 studies), service user outcomes (4 studies), or both (4 studies). In order to consider the studies and their outcomes in more detail, the 14 studies have been split into those including results for staff and those including results for service users.

Staff Studies

Participant numbers for the 10 studies that included staff outcomes ranged from 11 to 386 (i.e., 11, 16, 22, 32, 33, 38, 79, 169, 275, 386). Participants included staff from a residential and vocational support unit; staff from community services, including residential, day-support and specialist; professionals with a bachelor’s degree; direct-care staff; students undertaking a university diploma; students undertaking a distance learning course; nurses; behaviour specialists, teachers and unregistered staff from specialist health resources.

Outcome measures used in each study to evaluate different variables are outlined in Table 3.

Table 3 Staff outcome measures
Full size table

Service User Studies

Participant numbers for the eight studies that included service user outcomes, ranged from 5 to 138 (i.e., 5, 25, 32, 35, 37, 60, 61, 138). Outcome measures used in each study to evaluate different variables are outlined in Table 4.

Table 4 Service user outcome measures
Full size table

Discussion

This review presents data from 14 studies of PBS training. There are considerable differences between studies. The length of training is a fundamental difference, with a range from a 1-day training course, to a 2-year University Diploma course, including 57 days training. Comparisons between outcomes from different trainings must therefore be viewed in this light. The studies also vary considerably in their measures. Specifically, some measured staff outcomes only, some measured service user outcomes only, and some measured both. Studies focusing exclusively on staff outcomes included a range of measures. Specifically, some focused on cognitive and emotional responses; others on knowledge acquisition; and some on skills acquisition. Despite these differences, however, there are a number of observations that may be made.

A number of studies only considered changes in staff knowledge, attributions, or emotional responses (Lowe et al. 2007; McGill et al. 2007; Gore and Umizawa 2011). While significant changes in these variables were found following training, changes in staff behavior were not measured. Thus, it is difficult to extrapolate these results to changes in staff practice. It might be logical to assume that changes in staff beliefs, for example, about challenging behavior may lead to changes in support to individuals. However, there is little evidence to support this assumption. Stokes and Baer (1977) referred to this as a ‘train & hope’ model (i.e. delivering training and hoping that staff will implement positive changes back in the service).

Of the 10 studies focusing on staff outcomes, seven reported on staff skills. Of these, four studies reported on skills applied in practice (i.e. where PBS strategies are implemented and results for service users are reported). The remaining three studies (Freeman et al. 2005; Browning-Wright et al. 2007 and Kraemer et al. 2008) focused on related skills (i.e., writing BSPs). The four studies that focused on skills applied in practice focused mainly on person-focused training (i.e. trainees are supported to write and implement an FA and BSP for an individual; Baker 1998; Dench 2005 and Crates and Spicer 2012). In addition, Reid et al. (2003) used on-the-job checks and assessed role-play to teach a range of PBS skills. This opportunity to apply learning in practice appeared to clearly meet the guidance laid out by Dunlap et al. (2000) in their article on PBS training.

It seems likely that skills practiced are more likely to be retained than those that are not. It is likely that research focusing on staff behavior, rather than knowledge, feelings of efficacy, or causal attributions may have more of an impact on actual practice. The literature in relation to Active Support may be helpful (e.g. Jones et al. 1999, 2001) as it stresses on-the-job coaching as a necessary element of effective staff training and has demonstrated outcomes in achieving change of practice in staff’s own workplace.

In terms of length of training, results are mixed. Some short sessions were shown to have a clear impact on staff skills in developing BSP. For example, Browning-Wright et al. (2007) used a 1-day session and reported that post-training staff members were four times more likely to write BSPs rated as good or superior. Generally, however, the studies reviewed here appeared to show that longitudinal training that allowed trainees time and opportunity to practice skills within the workplace, and to complete assignments between formal teaching sessions, was more likely to be effective in achieving real changes in practice and in teaching staff lasting skills (Baker 1998; Freeman et al., 2005; Dench 2005; Crates and Spicer 2012). However, attributing changes following training to the actual training rather than to other organizational changes is difficult. Specifically, changes in management systems and organizational procedures may have impacted results (e.g. Lowe et al. 2007). Controlled comparison groups may be required in future studies seeking to isolate these variables.

Of the 8 studies including measures of outcomes for service users, 4 were based on work done in Ireland by (i.e., McClean et al. 2005; Dench 2005; Grey & McClean 2007; McClean and Grey 2012). All four showed significant reductions in challenging behavior following longitudinal person-focused training (i.e. training that supports staff to design and implement PBS strategies for specific service users (McClean et al. 2005). Positive effects were demonstrated both via real-time behavior recordings and by use of the Checklist of Challenging Behaviour (Harris et al. 1994). In addition, both Baker (1998) and Crates and Spicer (2012) showed considerable reductions in challenging behavior via longitudinal trainings. Baker demonstrated a reduction from baseline for five individuals following a 3-session person-focused longitudinal training course. Although this training was shorter than the four previously discussed studies, each session was a month apart, and follow-up data was taken 2 months after the final training session. Thus, it may be described as longitudinal training also. Crates and Spicer (2012) utilized a training-for-trainers format based on the La Vigna et al. (2005) model; and provided a training model similar to those described by previous studies (e.g., McClean et al. 2005; Dench 2005; Grey et al. 2007; McClean and Grey 2012). Although it is difficult to draw firm conclusions given a limited number of studies, it is possible that when staff skills are improved via longitudinal, person-focused training, positive changes in levels of service user challenging behavior may also occur.

Limited evidence exists with regard to links between PBS training and improvement of support/increased quality of life for service users. Only one study of the 14 reported directly on quality of life outcomes for service users. Dench (2005) attempted to evaluate outcomes for service users following a longitudinal PBS training course. However, despite positive anecdotal evidence of lifestyle changes, these were not able to be to be measured and reported. This may have been due to the Quality of Life Questionnaire (Schalock and Keith 1993) being insufficiently sensitive to measure these changes.

In terms of looking to the future, there is a need to be creative and innovative in how training is delivered, both to maximize resources and to achieve the coverage of training needed. Use of information technology and video training are both measures which have demonstrated some element of success (Macurik et al. 2008; Sailor et al. 2000), and may be useful in this context. Further exploration of their use in relation to PBS training would be appropriate.

In addition to creative use of technology, coverage could also be increased by utilizing a training-for-trainers model (Crates and Spicer 2012). Crates and Spicer demonstrated that ‘second generation’ training could be as effective as that delivered directly (e.g., La Vigna et al. 2005). This is an encouraging model in terms of achieving high quality training with reasonable coverage. Rotholz and Ford (2003) also utilized this approach in that trainees who successfully completed the PBS training for supervisors were eligible to enroll in a 2-day trainers’ course and were assessed in their training skills. (This study is not included in the review as the data were already reported in Reid et al. 2003). These elements of training may be useful as a means of furthering an organization’s ability to provide PBS training to their staff.

There are a number of implications for research and practice in PBS training arising from this review. There is a need for considerably more research on the impact of PBS training for service users in terms of frequency of behavior and also episodic severity (La Vigna and Willis 2005). In particular, more outcome data and research is needed in terms of quality of life outcomes for service users. In particular it would be useful to explore what supports good generalization and maintenance of benefits from PBS training; thus, studies including long-term follow-up would be particularly helpful.

In addition to service user outcomes, research is also needed to examine what makes training effective for staff and service users. This may include considering issues around impact of organizational environments that promote PBS training and practices. These might include the impact of the training level of trainees’ supervisors; the potential impact of good contextual fit (Albin et al. 1996) on implementation of PBS plans; and how organizational systems, such as the Periodic Service Review (La Vigna et al. 1994) may be used to support implementation.

In all of these areas, it would be ideal to see use of more powerful single case or group experimental designs in order to isolate the influences of different variables and allow clearer conclusions. More of an emphasis on the use of established measures for ensuring the reliability of the data presented would increase the confidence we would have in study results.

Conclusions

Research into the impact of staff training in PBS has focused more on outcomes for staff than service users. The eight studies that considered outcomes for service users demonstrated reductions in challenging behavior. However, improved quality of life for those individuals was not systematically linked to the training. It would be useful for future research to focus on evaluating impacts on the lives of service users resulting from staff training in PBS. Although it seems likely that staff training which supports the practice and application of skills within the workplace of candidates attending the training will be more effective in producing positive changes in the lives of service users, research is needed to support this assumption.