Introduction
Emotionally unstable personality disorder (EUPD) is an enduring psychiatric disorder characterised by emotional dysregulation, cognitive distortions, impulsivity and relationship instability. Dialectical behaviour therapy (DBT) equips people with EUPD with skills in managing intense emotions and changing maladaptive coping behaviours.
An earlier service evaluation by Fox et al. (Reference Fox, Krawczyk, Staniford and Dickens2014) explored the effectiveness of a comprehensive DBT programme for 18 women with EUPD, admitted to a specialist low-secure unit. Findings highlighted significant improvements in patient- and clinician-rated outcomes pertaining to symptomology, social functioning, clinical problems and risk behaviours, primarily within the first 6 months of treatment. Nevertheless, the outcome measures had not been selected for the DBT service specifically and thus may not have been important recovery indicators for those accessing the programme.
Since then, the service has transitioned towards a value-based healthcare model. The importance of systematic patient-centred outcome measurement as an indicator of quality of care is fundamental to a true value-based healthcare system (Porter and Lee, Reference Porter and Lee2013). Thus, as part of this evolution, the outcomes utilised by the DBT IPU were selected through a process of patient consultation. As such, the current study was conducted as a follow-up to the previous evaluation of the DBT service, exploring changes in outcomes which those accessing the service consider to be valuable indicators of recovery, rather than outcomes selected by the service. Additionally, whereas the initial evaluation explored outcomes within a singular specialist rehabilitation ward, the current study assessed outcomes across the DBT integrated practice unit (IPU), a multi-unit secure service providing a progressive care pathway across low secure and specialist rehabilitation wards.
Method
Design
A retrospective within-groups design was employed to evaluate treatment outcomes in women admitted to the DBT IPU between July 2017 and December 2019, over a 12-month admission period. Demographic, clinical and risk data were extracted from electronic records.
Participants and setting
Overall, 101 women were admitted to the DBT IPU. The service, which consisted of one low secure and two specialist rehabilitation units, provides in-patient treatment for women with a diagnosis of EUPD and additional complex mental health needs through a comprehensive DBT programme, as previously described by Fox et al. (Reference Fox, Krawczyk, Staniford and Dickens2014). Of these women, 41 had a complete dataset for at least one outcome and thus were included in the study.
Measures
Scores at admission, 6 months post-admission and 12 months post-admission were extracted for the ReQoL20 (Keetharuth et al., Reference Keetharuth, Brazier, Connell, Carlton, Taylor Buck, Ricketts and Barkham2017), COREOM (Evans et al., Reference Evans, Mellor-Clark, Margison, Barkham, Audin, Connell and McGrath2000), and HoNOSSecure (Sugarman and Walker, Reference Sugarman and Walker2007) measures. An alternative HoNOS scoring structure that has been shown to be a better fit for in-patient populations was utilised in the current study (see Maddison et al., Reference Maddison, Marlee, Webb, Berry and Whitelock2016). Security scales were also grouped based upon the two-factor structure outlined by Tiffin et al., Reference Tiffin, Daniel and Carey2011): those assessing ‘risk’ and ‘need’. The frequency and severity of risk behaviours (self-harm, violence against persons and objects) and the frequency of restrictive practices (restraints, seclusions) was routinely recorded on electronic records. Aggregated aggression scores (AAS) were also calculated to account for type and severity of risk behaviour (see Alderman et al., Reference Alderman, Knight, Stewart and Gayton2011).
Data analysis
Where the assumption of normality was met, repeated measures ANOVAs and pairwise post hoc comparisons were conducted to test for changes in outcomes between time points (baseline vs 6 months; 6 months vs 12 months; baseline vs 12 months). Where the assumption of normality was not met, the Friedman’s test and Wilcoxon signed ranks test were instead conducted. The Bonferroni correction was applied to adjust for multiple comparisons (p<.016). As data were not available on each outcome for the whole sample, the number of patients included in analyses varied between each outcome.
Results
Sample characteristics
Most participants were admitted to a specialist rehabilitation unit (n = 31, 75.6%), with the remaining participants residing in a low secure ward (n = 10, 24.4%). Participants were most commonly detained under Section 3 of the Mental Health Act (n = 36, 87.8%), with the remaining participants detained under forensic sections (n = 5, 12.2%). All participants had a primary (n = 40, 97.6%) or secondary (n = 1, 2.4%) diagnosis of EUPD. Age ranged from 18 to 51 years (M = 27.56 years, SD = 9.25). Of those who had been discharged (n = 18), length of admission ranged from 12 to 28 months, with a mean admission length equivalent to 18 months (M = 551.36 days, SD = 148.15).
Main effects
ReQoL-20 (n = 22)
Mental health scores improved across admission (F 2,42 = 5.42, p<.01, η2 p = .21). Scores on the physical health item did not significantly change (F 2,42 = 1.53, p = .23].
CORE-OM (n = 16)
There were significant reductions (improvements) across admission on the CORE-OM for global distress (F 2,28 = 13.28, p<.001, η2 p = .47), wellbeing (F 2,30 = 6.64, p<.01, η2 p = .30), problems (F 2,30 = 10.36, p<.001, η2 p = .41), functioning (F 2,30 = 11.67, p<.001, η2 p = .44), and risk (F 2,30 = 10.39, p<.001, η2 p = .41) scores.
HoNOS-Secure (n = 36)
There were significant reductions (improvements) across admission on the HoNOS-Secure for severe disturbance [χ2(2) = 19.75, p<.001], emotional wellbeing [χ2(2) = 8.96, p<.05], socioeconomic status [χ2(2) = 7.72, p<.05], risk [χ2(2) = 17.40, p<.001] and need [χ2(2) = 12.44, p = .01] scores. Personal wellbeing scores did not significantly change during admission [χ2(2) = 2.36, p = .31].
Risk behaviours and restrictive practices (n = 41)
There were significant reductions (improvements) across admission in incidents of self-harm [χ2(2) = 17.12, p<.001], and the use of restraints [χ2(2) = 13.65, p<.01] and seclusions [χ2(2) = 10.38, p<.01]. The mean AAS, weighted by type and severity of risk behaviour, also significantly changed over admission [χ2(2) = 14.97, p<.01]. The reduction in incidents of violence approached significance [χ2(2) = 5.90, p = .052].
Pairwise comparisons
Table 1 illustrates the results of the pairwise comparisons, conducted to explore changes between time points. Generally, the most significant changes occurred across the 12-month admission period, comparing between scores at baseline and 12 months post-admission.
Table 1. Change in outcomes during 12-month admission to the DBT IPU
T0, baseline assessment; T1, 6-month post-admission assessment; T2, 12-month post-admission assessment.
a median and range values are reported for skewed variables.
* statistically significant result after applying the Bonferroni correction.
The exceptions to this were incidents of violence, and use of seclusions, which significantly reduced in the first 6 months only, and in severe disturbance (HoNOS-Secure) scores, which reduced marginally more in the first 6 months, compared with changes over the 12-month admission period. No significant changes in outcomes were observed in the latter 6 months of admission. Whilst a main effect was previously identified for socioeconomic scores on the HoNOS-Secure, post hoc time point comparisons were not significant on this subscale.
Discussion
The current study aimed to evaluate changes in outcomes for women with EUPD admitted to a specialist DBT service, building upon an earlier evaluation (Fox et al., Reference Fox, Krawczyk, Staniford and Dickens2014), conducted prior to the adoption of a value-based model of care. The findings demonstrate improvements in quality of life (ReQOL), psychological functioning (CORE-OM), health and social functioning, and security risks and needs (HoNOS-Secure), as well as reductions in self-harm and restrictive practices. Whilst reductions in incidents of violence bordered significance, the frequency of these incidents at baseline was already low. Accounting for the type and severity of incidents, analysis showed a significant reduction in AAS scores over admission. Generally, the most significant improvements occurred over the 12-month admission period, suggesting that a shorter 6-month admission duration is not substantial in eliciting significant improvements.
No significant differences were found on measures of physical health (ReQoL), socioeconomic status or personal wellbeing (HoNOS-Secure). Nevertheless, there are a number of important caveats to consider. Firstly, physical health was measured on a singular self-report item only. Additionally, the non-significant change in socioeconomic status, a subscale consisting of the items ‘problems with living conditions’ and ‘problems with occupation and activities’, is likely to be a reflection of the setting itself; patients were detained to a secure in-patient service and thus were out of employment with restricted opportunities for activities. Thirdly, average scores on the personal wellbeing HoNOS-Secure subscale were somewhat low at each time point. Additionally, given that patient-reported ratings of ‘wellbeing’ on the CORE-OM did significantly change over admission, this may reflect a discrepancy in perceptions between clinicians and patients.
The changes reported here are largely consistent with those reported previously by Fox et al. (Reference Fox, Krawczyk, Staniford and Dickens2014). There are, however, some discrepancies in the point of admission for which the greatest improvements in outcomes occurred. Whereas Fox et al. (Reference Fox, Krawczyk, Staniford and Dickens2014) reported the most significant improvements in outcomes between baseline and 6 months post-admission, there was a tendency for the greatest changes to occur between baseline and 12 months post-admission. This may therefore suggest that improvements in patient-valued outcomes take longer to manifest. The exception to this trend was in the frequency of seclusions and violence, where significant reductions were found between baseline and 6 months only.
Limitations
The study reports on outcomes across 12-month admission to the DBT IPU. However, of those who had been discharged, the average length of stay exceeded one year. Further investigation into outcomes over a longer admission period is therefore warranted. Furthermore, data were not available for the whole sample on all outcome measures, and thus the findings reported are reflective of separate subsamples. Finally, the IPU delivers holistic care through a multidisciplinary team to address all areas of need, and thus it cannot be concluded that improvements were a direct result of the DBT programme itself.
Supplementary material
To view supplementary material for this article, please visit: https://doi.org/10.1017/S1352465822000467.
Data availability statement
The current study reports on confidential patient data and thus is unavailable to access.
Acknowledgements
None.
Author contributions
Elanor Webb: Conceptualization (supporting), Data curation (lead), Formal analysis (equal), Methodology (lead), Project administration (lead), Writing – original draft (lead), Writing – review & editing (lead); Alessandra Girardi: Conceptualization (lead), Data curation (supporting), Formal analysis (equal), Methodology (supporting), Project administration (supporting), Supervision (supporting), Writing – original draft (equal), Writing – review & editing (equal); Emily Fox: Conceptualization (equal), Methodology (supporting), Project administration (supporting), Writing – original draft (equal), Writing – review & editing (equal); Paul Wallang: Writing – original draft (supporting), Writing – review & editing (equal).
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conflicts of interest
The authors declare none.
Ethical standards
All authors have abided by the Ethical Principles of Psychologists and Code of Conduct as set out by the BABCP and BPS. Informed patient consent was not sought as the current study was an internal service evaluation, utilising anonymous, routinely collected data held by the service. The current study was an evaluation of the DBT service, based on routinely collected clinical assessments, and thus formal ethical approval was not required. However, approval was sought from the organisation’s internal Clinical Audit and Assurance committee and the services’ clinical director.
Comments
No Comments have been published for this article.