This is an open, pragmatic, two-parallel-group, individually randomized controlled trial. Women seeking help from specialist DVA agencies are randomized at a 1:1 ratio to usual DVA agency support or usual DVA agency support plus additional psychological support from a specially trained advocate or support worker. Two hundred fifty women are being recruited into the study at two DVA agencies, both in the United Kingdom. Participants are enrolled in the study for 1 yr, during which time they will all have access to usual support from the DVA agency. The primary outcome is assessed at 1 yr from randomization.
Participants are eligible for inclusion if they are experiencing domestic violence or abuse which has led them to seek support from one of the recruiting sites and are age 16 years or older. Women are excluded if they (1) have a psychotic illness, (2) have a severe drug or alcohol problem, (3) are unable to read English or (4) are currently attending counseling, CBT or other psychological treatments in either primary care or specialist psychiatric services.
Recruitment is taking place at two sites: Bristol Next Link (http://www.nextlinkhousing.co.uk/) and Cardiff Women’s Aid (http://www.cardiffwomensaid.org.uk/). All women who present to the DVA agencies are assessed for eligibility and invited to consider participation as part of the service’s routine intake procedures. Women who are potentially interested in participating are then referred to a female researcher who meets with them face-to-face at a safe and convenient location.
All women referred to the study are provided with a written patient information sheet and have the opportunity to discuss the study with the researcher. Informed consent is obtained from each participant in writing. After the woman consents to participate in the study, the researcher works with her to develop an individualized schedule which includes safety information and the details of friends, family or associates (locators) who will be able to help the researcher to contact the woman if contact is lost. The participant is then asked to complete a baseline questionnaire. On completion, the researcher randomizes the participant by using a remote, independent, automated telephone randomization service provided by the Bristol Randomized Trials Collaboration (http://www.bristol.ac.uk/social-community-medicine/centres/brtc/). Randomization is stratified by whether women receive support from the safe house team or from the resettlement team at Next Link and the refuge team or tenant support team at Cardiff Women’s Aid. The randomization program applies stratification with random blocking. Within each stratum, participants are assigned to the treatment group (T) or control group (C) in randomly chosen blocks of two, four or six. Within each block, the order of assignment is also random, so that in a group of four, for example, the permutation TCTC would be equally as likely as TTCC. For each completed block, the allocation ratio is 1:1, so that the number of Ts and Cs will be equal. Allocation is concealed from the researcher until the moment of randomization.
Over the course of the study, women are requested to complete three further questionnaires, at 4, 8 and 12 mo after randomization. Questionnaires are either hand-delivered or sent to the women, depending on the stated preference. Women receive a £10 shopping token after completing the baseline, 4- and 8-month questionnaires and £20 for completion of the 12-months questionnaire. To retain women in the study, researchers make midpoint contacts with women between follow-up points. When researchers are unable to contact a woman on four consecutive occasions, they begin to work through the locator contacts given by the women at the point of recruitment into the study. When a woman cannot be located, researchers continue to use the woman’s last known details. Women are prompted to return questionnaires at 2, 4 and 6 wk after the questionnaires are sent or delivered. Failure to return a questionnaire by week 7 prompts the researcher to begin working through the women’s locator contacts. Researchers continue to prompt women (or their locators) for the return of questionnaire until week 12 after posting. The researcher will call the woman by telephone and, if no answer is received, leave a voice message (unless it has been explicitly stated that it is not safe to do so). After two failed telephone contact attempts, the researcher will send a text message to the woman’s last known mobile telephone number. If there is still no response, the researcher sends a handwritten card to the woman’s last known address. Failure to return a questionnaire by week 12 (after it was sent or delivered) results in the questionnaire’s being noted as missing. Researchers will continue to issue the next questionnaire in the series even if the previous questionnaire was not returned.
Women randomized to receive psychological support in addition to usual DVA agency support will be assigned a specialist psychological advocate (SPA) who has gone through a 25-day manualized training program from a clinical psychologist with expertise in DVA (RAD). The training addresses the psychological impacts of DVA on women and develops therapeutic skills specifically tailored for this client group. SPAs are trained to work with women with common presenting problems within a single-session model  using a session structure (finding a focus for work, obtaining a commitment to work, working on a specific topic, evaluation of progress and homework) based on the work of Daldrup and colleagues . Topics covered include posttraumatic stress, depression, anxiety, low self-esteem, unresolved anger and managing loss. SPAs are also provided with handouts and self-help resources that can be used with their clients. Supervision of the SPAs is provided by the clinical psychologist, who listens to a sample of recorded SPA sessions and provides feedback through regular telephone or email contact and periodic face-to-face meetings.
Once assigned a SPA, the participant will attend eight SPA sessions, meeting either weekly or fortnightly during the course of 2 to 3 months, with a further two booster sessions, 1 and then 3 months later. During SPA sessions, the advocates will provide time-limited interventions using a variety of primarily CBT psychological techniques, focusing within any one session on a specific presenting problem, such as hyperarousal, sleeping difficulties or parenting problems. The advocate will aim to empower the client to apply therapeutic strategies (such as relaxation, challenging thoughts or goal-setting) to promote recovery from each problem. Each series of eight sessions varies according to the client’s presenting problems, but typically includes discussions about the impact of domestic abuse on mental health and common reactions, typically including posttraumatic stress and low self-esteem. The client selects a specific topic for each session, and the SPA will use written resources to guide identification of signs (symptoms), the association with experiences of abuse and strategies to overcome them within that session, and through homework tasks to be completed between sessions. Each of the eight sessions can have different content for any therapeutic dyad, although SPAs will focus on building the alliance and safety in the first session and address endings and closure in the eighth session. The follow-up sessions 1 and 3 months after the eighth session are focused on consolidating gains.
In addition, the woman will receive the usual support or advocacy provided by the DVA agencies. This support can involve assistance with a variety of health and social issues, including housing problems, budgeting and debt, and legal proceedings. There is evidence that advocacy on its own can reduce physical abuse and possibly improve quality-of-life outcomes for women who have experienced DVA. Importantly, usual advocacy also provides a space for women to share their experiences and talk about their feelings in a safe and nonjudgmental environment.
Control participants will have access to the usual DVA agency support and advocacy as described above. They will not receive SPA sessions, and their support worker advocate will not have received specialist training in psychological methods. The length of time a woman can be engaged with a DVA agency varies, depending on their needs and the service’s policies. Support can range from a one-off session with onward referral to regular meetings over a period of months.
We are not using an attention control group, as this trial is designed (and funded) explicitly as a pragmatic (effectiveness), not explanatory (efficacy), trial, building on previous trial evidence for the effectiveness of psychological and advocacy interventions. Therefore, it is legitimate to have a treatment-as-usual rather than an attention control group. We will collect data on contact time between advocates and participants in both arms and will use that in our interpretation of the findings.
This study has two primary outcomes, both collected at 1 yr after randomization. The first is the Clinical Outcomes in Routine Evaluation–Outcome Measure (CORE-OM) score . CORE-OM is a standardized, validated global measure of psychological distress based on a self-completed 34-item questionnaire, with established sensitivity to change, good test-retest reliability and UK normative data. It was designed to assess efficacy and effectiveness across multiple disciplines offering psychological therapies . The second is the Patient Health Questionnaire (PHQ-9) score , which is a widely used, self-completed measure of depression  with extensive validation in diverse populations with sensitivity to change, leading to its adoption across a wide range of trials internationally.
Some further validated mental health measures are being used: the Generalized Anxiety Disorder 7 scale (GAD-7) , the Posttraumatic Stress Disorder CheckList developed by Weathers et al. , the Composite Abuse Scale to measure type and severity of abuse , the Short Form-12 Health Survey to measure quality of life  and the EuroQol EQ-5D to measure health state utility . All secondary outcomes will be measured at 4, 8 and 12 mo by questionnaire. Data derived from the PHQ-9 and CORE-OM questionnaires at 4 and 8 months will also be treated as secondary outcomes and will be used for imputation in the absence of 12-month outcome.
For the economic evaluation, data will be collected on the women’s use of health, social and criminal justice services. We will also collect data on expenses incurred by the women for attending appointments, such as travel and child-care costs. Resource use by children of participants will also be measured, such as appointments with youth offender teams and school attendance, as many of the effects of DVA on children and young people manifest as psychological, social, educational and/or behavioral difficulties [33–35]. These data will be collected by questionnaire at the same time points as the above outcome measures.
Other variables collected in the participant questionnaires are age, number of children at home, ethnicity, income, occupation and relationship to perpetrator. The following process data are recorded by the domestic violence advocates working with the participant: number of DVA agency sessions the participant attends; the main focus of each session; and advocates’ contact time with participants, number of women entering the DVA service who do not meet the eligibility criteria or who are not interested in participating. We will also collect data on adverse events self-reported by the participant (verbally and/or by questionnaire) or elicited by the DVA advocate following contact with her.
Using pilot data, the sample size is based on being able to detect a difference of 0.5 in the CORE-OM score, a “reliable change index” . This will be the difference in psychological distress between intervention and control groups as measured by the CORE-OM questionnaire 12 months after each patient is randomly assigned. A sample size of 200 participants in total gives a power of 96% to detect a difference of 0.5 on the CORE-OM, corresponding to an effect size of 0.5 and 81% power to detect an effect size of 0.4. Assuming an attrition of 20%, we will need to recruit 250 women.
An effect size of 0.4 to 0.5 is consistent with effect sizes detected in studies of psychological interventions using the CORE-OM as an outcome measure  and is similar to that of CBT interventions on the PHQ-9 and other depression outcome measures . An effect size in the 0.4 to 0.5 range is also comparable to the outcome of trials of psychological interventions for women who have experienced domestic violence (see Appendix 7.5 in our systematic review of interventions ). For example, Kubany and colleagues’ two trials of CBT for women diagnosed with PTSD following intimate partner violence reported effect sizes of 0.9  and 0.33 , respectively, for PTSD and 0.4 and 0.25, respectively, for depression outcomes. Labrador et al.  reported effect sizes of 1.23 and 1.77 for PTSD and depression outcomes, respectively, and Reed and Enright  reported effect sizes of 2.33 and 1.55 for PTSD and depression, respectively.
Nested qualitative study
A sample of approximately 30 women across both trial groups from both recruiting sites will be invited to take part in a series of three interviews spaced across their year of participation. The aim of these interviews will be to explore women’s assessment of their mental well-being and quality of life in their own words, as well as the overall acceptability and perceived effectiveness of the services they received. Topic guides include questions that explore which of the therapeutic strategies (such as exercises to resolve anger, challenging negative thoughts or goal setting) were most useful to the women and why. Interviews are face-to-face and will last up to one hour. With participants’ consent, interviews will be recorded and transcribed verbatim. Purposive sampling based on data collected in the baseline questionnaires will ensure a diverse group of informants in terms of age, currency of abuse, and mental health status.
Trial outcomes will be analyzed and reported in accordance with Consolidated Standards of Reporting Trials (CONSORT) guidelines. Baseline demographic data such as gender, age, duration of abuse and relationship to perpetrator will be summarized and descriptive summary statistics will be provided to check for imbalances between the two trial groups. For variables with continuous measures, we will report the mean and standard deviation, and for categorical data, we will report frequencies and percentages.
Because the allocation will be stratified by site (Bristol vs Cardiff) and type of setting where the treatment is administered (safe house vs within community), the comparative analyses will adjust for these two factors. These will involve normal and logistic regression analyses for continuous and binary outcomes, respectively, and also, where appropriate, will adjust for the baseline scores of the outcome variables. The primary comparative analysis will compare continuous CORE-OM scores between the two groups as randomized using an intention-to-treat regression model. Secondary analysis will adjust for any factors exhibiting substantial baseline imbalance; further secondary analyses will include prespecified subgroup analyses according to the two stratification variables and age, which will be investigated by introducing the relevant interaction with treatment group into the regression model . It is possible that the treatment effect will differ across both sites and settings because of demographic differences between Cardiff and Bristol and differences in community and safe house or refuge settings. It is also likely that treatment impact will differ between age groups, partly in connection with baseline exposure to violence, which tends to be higher in the younger strata of the population [1, 42]. The subgroup analysis by age will use the median as a threshold to separate younger from older women in the sample as well as use age as a continuous variable in two separate specifications. Because rapport with the therapist (in this case, the specialist psychological advocate) is a key component of therapy effectiveness [43–45], we will also account for potential differences in impact at the therapist level in generalized mixed models. To assess the stability of any treatment effect, we will fit a mixed model for CORE-OM or PHQ-9 scores at 4, 8 and 12 months, adjusted for baseline CORE-OM or PHQ-9, using number of treatment sessions as a fixed categorical effect.
The impact of, and reasons for, missing data will be investigated in sensitivity analyses based on multiple imputation by chained equation models (mice) [46–49]. The imputation model will contain all the variables predictive of missingness and of the outcome, as well as those in the hypothesized model . The number of imputations will be determined on the basis of the percentage of data that will be missing, as well as on the precision of the estimate . A random coefficient mixed effects model will also be considered in comparing groups in terms of CORE-OM and PHQ-9 scores. A mixed (fixed and random) effects model will allow us to account for the distribution of errors, compared with fixed effects models  and to deal with missing data were these not to be missing at random .
We will calculate complier-average causal effect (CACE) using instrumental variables techniques to establish a stronger causal link between assignment to treatment and impact in the absence of full adherence to the treatment [46, 52, 53]. In particular, it is possible that different individuals will complete treatment at a different pace than that envisaged by the trial team, and these models will take this heterogeneity into account. To calculate CACE estimates, adherence will be represented primarily by the number of treatment sessions attended. Where a binary version is required, we will in the first instance consider completion of at least four sessions at 12 months from randomization as reflecting adequate adherence.
Although there is literature on the economic impact of DVA, to date there has been little economic analysis of interventions aimed at improving outcomes for domestic violence survivors . The aim of the economic evaluation is to compare the extra cost of providing the PATH intervention with the extra benefits gained to allow judgment about its cost-effectiveness. This process involves detailed individual-level costing of caring for women who are victims of DVA. The cost of caring for women offered the PATH intervention (that is, sessions with the SPAs) will be compared with the cost of domestic violence advocacy alone. The analysis will consider costs associated with the delivery of care from the perspectives of the NHS and personal social services (PSSs), the public sector and society in general. Resource use will be measured from the point of randomization until final follow-up 12 mo later. Data will be collected by means of a questionnaire at 4, 8 and 12 mo, and the analysis will include all resources used by the women unless clearly unrelated to domestic violence, including relevant resources used by or on behalf of children younger than age 17 years in their care.
We will report the direct costs of the intervention. Because the training curriculum was developed prior to the intervention [20, 21], start-up costs pertain only to training of the SPAs. Delivering the intervention requires SPA time, clinical supervision and management by the host domestic violence agency. We will quantify time with the use of timesheets and the original budgets. This time will be valued at advocates’ and trainers’ salaries and management costs paid to the host agencies.
Use of NHS and PSS resources will be captured through the questionnaire. These will include primary and community health services, prescribed medication and use of hospital services. Unit costs for the valuation of these resources will be derived using the method described by Curtis , Department of Health reference costs and the British National Formulary. Community and social care services (such as family therapists and social workers) provide day-to-day support to DVA victims and their children to help them cope with the trauma and aftermath of domestic abuse. We will measure residence in refuges or safe houses by participants during the 1-yr follow-up period, estimating costs from information given by participating agencies and other refuge providers. Use of legal services as a consequence of DVA will be measured using our questionnaire, and we will derive relevant unit costs from current aggregate estimates . Finally, in line with the societal perspective in this analysis, we will also quantify personal and social costs caused by the exposure to DVA. These will include personal outlay for travel, child care and change in housing circumstances, as well as the use of voluntary sector services and time off from work.
We will present results using a cost consequences framework comparing cost from each perspective with primary and secondary outcomes . We will use the EQ-5D data from the study questionnaires to construct quality-adjusted life years (QALYs) and present an incremental cost-effectiveness ratio. This will indicate the cost per QALY gained by adding a psychological intervention to advocacy for women experiencing DVA. Costs and outcomes will not be discounted, as the study will be limited to a period of 12 mo. The effect of uncertainty in unit cost estimates or assumptions about resource use will be addressed in sensitivity analyses. Uncertainty in the cost-effectiveness/utility ratios resulting from participant variation in resource use and effectiveness will be captured by estimating confidence intervals around the net benefit statistic and estimating cost-effectiveness acceptability curves.
The aims of the qualitative analysis will be twofold: first to understand the meaning and perceived effectiveness of the intervention in the context of the women’s lives, and second to deepen our understanding of the process of identity change for women as they move through some or all of the stages from victim to survivor to thriver . The first analytic strand will provide important insights into contextual and motivational factors for the women in relation to the intervention. This will enable hypotheses to be tested about factors likely to act as barriers or facilitators to the effectiveness of the intervention, for example, practical arrangements for SPA sessions, such as the availability of a safe location or child care or finding suitable times to attend sessions, as well as factors relating to women’s expectations or prior experiences in help-seeking and their relationships with provider organizations and individual SPA workers. Comparisons will be made with the help-seeking experiences of women in the control arm of the study. The second analytic strand will be conducted using the perspective of interpretive phenomenological analysis which explores the lived experience of participants . This approach allows for a deeper exploration and interpretation of women’s experiences. Insights gained will enable the research team to critique and add to the theoretical development in the field of DVA, most of which has been generated in the United States [60–63].