Service evaluation of a Human Givens Therapy service for veterans

Abstract

Background

PTSD Resolution is a charitable provider of mental health support and treatment via Human Givens Therapy to members of the Armed Forces Community. This study utilized data from their client database to evaluate outcomes of their services.

Aims

To determine whether treatment by the service resulted in improvements in mental health; to the extent possible, compare these outcomes with those of NHS IAPT (National Health Service – Improving Access to Psychological Therapies) services; and to evaluate how other factors, notable reported stress levels and prior treatment, impacted outcomes.

Methods

Administratively collected data from the service provider collected between 2014 and 2016 were utilized. Clients still undergoing services were excluded. The CORE-10 (10-item Clinical Outcomes in Routine Evaluation) measure was used as the outcome measure. Demographic factors examined included age, sex, employment, accommodation, relationship status, distress at the time of presentation and prior treatment history. The effect of the factors was analysed using χ² test as well as linear regression and Poisson regression analyses according to the nature of the variable.

Results

PTSD Resolution clients appeared to show a similar degree of improvement as IAPT patients. Job-seekers and those who are living alone benefited less from the treatment. Despite high levels of distress and prior treatment among this client group, these factors did not seem to affect treatment outcomes.

Conclusions

The services of this provider appear to be an acceptable alternative for IAPT treatment, to the degree such a comparison can be made given differences in measures and client groups.

Introduction

Numerous charitable providers in the UK aim to provide mental health support and treatment to veterans, outside the provisions of the NHS. Such providers provide alternative treatment services and/or gateways to services that may provide faster access to care or simply be preferable or more convenient for veterans.

PTSD Resolution is one such charity offering free a talking therapy intervention for members of the Armed Forces Community (including reservists and dependents) using a Human Givens Therapy (HGT) approach [1]. HGT incorporates concepts from several therapeutic paradigms, including cognitive and behavioural therapies, as well as addressing the unmet emotional needs of clients. As a relatively new therapeutic approach, it has yet to establish a strong evidence base. PTSD Resolution delivers its services through a network of over 200 HGT practitioners; veterans who access their services can do so without the need for a referral. PTSD Resolution aims to provide up to six one-to-one, 1-h HGT sessions focused on treating mental health problems associated with military trauma including but not limited to post-traumatic stress disorder (PTSD). Clients may have more than six sessions if they require.

This study aimed to examine the outcomes of PTSD Resolution clients to determine (i) whether their symptoms of mental illness improved during the treatment; (ii) to the degree to which, where comparisons were possible, treatment outcomes and attrition compared with those of NHS IAPT (National Health Service – Improving Access to Psychological Therapies) services and (iii) how other client attributes, particularly levels of distress at presentation and prior treatment experiences, affected treatment outcomes.

Methods

This service evaluation examined data from clients referred to PTSD Resolution therapists between April 2014 and February 2016. Anonymized data that had been routinely collected by PTSD Resolution were provided to King’s College London for service evaluation. For the purpose of the analysis, we have included only closed cases (ended therapy or dropped out).

Clients of PTSD Resolution are asked to complete a brief mental health measure when attending sessions: the CORE-10 (10-item Clinical Outcomes in Routine Evaluation) [2]. The CORE-10 measures outcomes and progress of clients’ therapy and is a 10-item short measure derived from the 34-item CORE – Outcome Measure (CORE-OM) [3]. It has been used for routine assessment of therapies in secondary mental healthcare to evaluate mental health. Studies have shown that CORE-OM is highly correlated with the Clinical Interview Schedule – Revised (r = 0.77) [4], and it shows good test–retest reliability and validity against numerous other instruments [5]. CORE-10 shows similar agreement with other measures as the CORE-OM [2]. CORE-10 yields sensitivity and specificity values of 0.92 and 0.72, respectively, for depression at a cut-off score of 13 and a clinical cut-off score for general psychological distress of 11 with a reliable change [6]; this study used a cut-off of 11 to determine caseness, and a change in six points on the score as representing ‘reliable change’, as recommended for this measure [2].

Clients were asked to provide a range of standard socio-demographic questions: sex, age, current employment status, accommodation, and their relationship status and dependents. Additionally, they were asked how distressing their problems were for them at initial presentation, on a scale of 0–10 of increasing distress; if they had received prior treatment, and to rate such if they had received it on a scale of 0–10; and current use of pharmaceuticals.

Analyses were performed using the statistical package Stata version 11.2 [7]. Basic comparisons of proportions between categories of demographic variable were performed using χ² tests. Some analyses utilize Poisson, rather than linear, regression, as the distribution of outcome values does not approximate a normal distribution but rather a Poisson distribution (i.e. having the highest value at zero, with decreasing frequency of each subsequent outcome, such that the mean outcome approximates the variance).

Results

Table 1 shows the demographic distribution of 504 PTSD Resolution clients who had ended their treatment (whether by completing treatment or dropout).

Eleven per cent of the PTSD Resolution sample were female, which is broadly in line with the characteristics of the Armed Forces: 10% of current UK Regulars are female [8], and 11% of the veteran population were female in 2014 [9]. For comparison, 8% of serving personnel were female in 2001 (around the average time of leaving for this sample) [10]. The mean age was 42 years, and around two-thirds were taking some form of medication. Forty-five per cent were employed, which is somewhat lower than 54% employment rate in the IAPT sample, and substantially lower than employment levels in the veteran population in general (75%) [11]. Over half were living with their partner, but a substantial proportion was living alone (31%). A minority (37%) were married; this is substantially lower than the veteran population in general (64%) [9].

Similar to IAPT services, not all who enter the programme receive therapy. Within the PTSD Resolution sample, 85% attended more than one session (including those sessions listed as assessment, therapy or completion, as any of these may contain at least some treatment, but not screening or follow-up sessions). By comparison, 64% of IAPT users do not return for further treatment after the initial contact [12]. However, these rates are not directly comparable due to differences in types and definitions of contacts. Those in employment were most likely to receive therapy (χ² = 21.1, P < 0.01); this was the only recorded demographic variable which has a significant effect on whether or not individuals receive at least some therapy (i.e. at least one session which might include treatment, as indicated previously; henceforth, these will be referred to as ‘therapy sessions’).

Sixty-five per cent of PTSD Resolution clients had a planned ending; no demographic factor had a significant effect on achieving a planned ending to therapy. Of those with a planned ending (i.e. treatment was completed as determined by the provider), 97% (n = 315) attended at least one therapy session, while of those with an unplanned ending (i.e. attrition), 62% (n = 110) attended at least one therapy session. Those in the 45–54 years age group attended significantly more sessions (i.e. slightly less than one more session than the baseline 35–44 years age group), but there were no other significant effects among other age groups. Job-seekers attended fewer sessions; this may be explained by the previous finding that job-seekers were less likely to complete treatment, as those with unplanned endings had significantly fewer treatment sessions. Those on courses of medicine had more sessions of therapy. Possible explanations for this observation could be that such individuals are accustomed to making more contact with medical services, or that they had greater needs hence requiring more therapy sessions or that willingness to take medication is indicative of motivation to improve.

Forty-two per cent of those defined as cases according to their first CORE-10 measure were no longer a case by their last recorded measure; this is comparable to recovery rates of 37–45% of those using IAPT services [12,13]. Of those who had a planned ending, this recovery rate rose to 52%, which is again similar to IAPT users (59%) [12]. We found that 65% showed a reliable change in CORE-10 score (i.e. a decrease of 6 or more), suggestive of reliable improvement. This rose to 67% for those who were cases at first measurement, comparable with the average of 64% in IAPT [13]. We also found that 42% of cases at baseline were both no longer cases and showed reliable improvement by last measure; this compares with an average of 40% of IAPT users [13].

Those with unplanned endings and those who were unemployed but not seeking work had worse treatment outcomes (Table 2). Despite receiving more therapy sessions, long-term sick and pharmaceutical users did not demonstrate a significant difference in score change.

Eighty-two per cent reported having received prior treatment from some other provider. Among those who reported prior treatment, there was no difference between those rating prior treatment as not at all helpful with the remainder in the proportions receiving therapy from PTSD Resolution (χ² = 0.0021, P = 0.963), having a planned ending (χ² = 1.33, P = 0.249), or in number of therapy sessions attended (regression coefficient = −0.21, 95% confidence interval [CI] −0.93 to 0.52, P = 0.571).

Among those who received at least one session of treatment with PTSD Resolution, there was no significant change in the CORE-10 score between those with poor prior experiences and the remainder (regression coefficient = 1.76, 95% CI −1.43 to 4.95, P = 0.276).

Reports of distress are generally high; 40% selected the highest score and 91% selected a score from 7 to 10 (Figure 1).

The association between distress and the covariates was analysed using Poisson regression (see Supplementary Appendix 1, available as Supplementary data at Occupational Medicine Online), by reversing the scores for distress (i.e. converting score 10–0, and vice versa); thus, an increase in score represents a reduction in distress.

Females reported being more distressed by their symptoms, as were older clients, the long-term sick, those living alone and those using pharmaceuticals. Notably though distress did not predict having an unplanned ending.

Those who were less distressed had lower initial CORE-10 scores (coefficient = −0.92, 95% CI −1.49 to 0.35, P = 0.002) (Table 3), but the relationship with attendance at therapy sessions and distress was borderline (coefficient = −0.19, 95% CI −0.39 to 0.01, P = 0.065), and score change to last recorded measure was not significantly different between distress scores (coefficient = −0.21, 95% CI −0.96 to 0.54, P = 0.580).

Discussion

Where such comparisons were possible, we found that PTSD Resolution clients had similar rates of recovery and reliable improvement of symptoms (42%) as IAPT patients (40%) [13]. These data are similar to those found in a prior pilot study using clients of the same practitioner network as in this study (55%) [14] and its 5-year evaluation (51%) [15].

PTSD Resolution clients were broadly representative of veterans in terms of sex and ethnic origin allowing for the time elapsed since leaving service. A little under half (45%) of clients were in employment; this figure is much lower than UK veterans in general, and somewhat lower than the IAPT population (54% in the 1-year IAPT report) [12]. Furthermore, nearly a quarter (24%) were unemployed due to long-term sickness and additionally, nearly two-thirds of the client group were on some form of medication. A large proportion were single (39%), and a much smaller proportion were married (37%); once again this figure is substantially lower than the general veteran population (where 64% are married).

Clients most frequently attended six sessions (22% of all clients attended exactly six), in line with PTSD Resolution’s target. Even among those with unplanned endings, a substantial proportion (42%) attended more than one session (and 5% attended six or more). By comparison, 36% of clients of IAPT services did not return for further care after the first session [12]. IAPT services observed 60–66% treatment completion [16]; we found that 65% of PTSD Resolution clients had a planned ending. This is consistent with the findings of a previous pilot study of this therapist network, which had 65% planned endings [14], and a little below the findings of the 5-year evaluation [15], which had 73% reaching a planned ending (but 21% of that sample did not have ending type recorded).

While clients who had unplanned endings were less likely to experience improvement, there were no clear risk factors predicting unplanned endings. We did find that clients who were seeking employment attended fewer sessions. These findings could indicate a shortage of resources for such clients, such as limited finances or time available to access care, or that as job-seekers, they were more focused on seeking employment than attending treatment sessions. Whatever the reason, job-seekers did not receive as much benefit from therapy as other employment groups.

We also found that people who were living alone were less likely to benefit from therapy. This may be because social support provided by, for example, spouses and partners is associated with better prognosis in treatment, possibly because close social contacts encourage engagement in therapy. This finding is very much in keeping with previous research showing that good social networks are protective of mental health in veterans [17].

We identified that more than half of the sample had received some form of prior treatment; most of those who did rated their previous treatment experiences as very poor. We did not find, however, that receipt of previous treatment (regardless of how such treatment was rated) influenced the effect of treatment with PTSD Resolution.

At the time of initial presentation to PTSD Resolution, most clients reported high levels of distress regarding their symptoms; this was particularly true for females, the long-term sick, those living alone, and those on medication. While distress was correlated with higher CORE-10 scores at initial presentation, it did not otherwise affect treatment progression (i.e. use of services and outcomes).

Clients on medication showed less overall benefit from treatment, despite higher attendance and CORE-10 scores at presentation, as did the long-term sick. This may be an unavoidable consequence of the severity of their symptoms, or an indication that some other treatment pathway might be more appropriate for those suffering a high or sustained symptom load.

This study is an evaluation of existing services rather than a controlled trial, and hence no comparisons with a control group (whether waiting list or alternative treatment regime) are available. Long-term follow-up is also unavailable, so any remission or reliable change identified cannot be assumed to persist. The CORE measure used is not one which is commonly used in studies of PTSD or other mental health difficulties, making it difficult to perform comparisons with other literature. Thus, comparisons with IAPT data (including those involving recovery and reliable improvement) cannot be made directly. We were unable to determine how long clients had experienced their symptoms and could not thus ascertain how much improvement in symptoms might have been related to expected resolution of non-pathological distress as a result of singular stressors (e.g. relationship breakdown).

The main result of this study suggests that members of the Armed Forces Community who attend a PTSD Resolution therapist appear to experience similar rates of recovery and reliable improvement as IAPT users. However, it appears that those who present with more severe difficulties benefit less from the services of PTSD Resolution and may be better served through referral for a more comprehensive assessment of their needs by NHS veteran’s services or charitable providers such as Combat Stress which employ consultant psychiatrists and provide care for those with more complex needs. This study is a service evaluation rather than a trial and direct comparisons with IAPT data were not possible. As such, while these results should be seen as encouraging, further research is needed to better understand how effective PTSD Resolution therapists are compared to more mainstream mental healthcare providers.

Competing interests

N.G. is the RC Psychiatrists Lead for Military and Veterans’ Health and a trustee for two military charities but not the charity from which the data came from for this paper. H.B. declares no conflicts of interest. This study was funded by a grant from PTSD Resolution. The funder had no input on the study methodology or the contents of this paper.

References