Evaluating effectiveness and cost-effectiveness of a group psychological intervention using cognitive behavioural strategies for women with common mental disorders in conflict-affected rural Pakistan: study protocol for a randomised controlled trial

The objectives of this cluster randomised controlled trial (cRCT) are to evaluate the effectiveness and cost-effectiveness of the locally adapted PM+ Group intervention for women in Swat, Pakistan as compared to enhanced usual care (EUC; defined below). The primary objective is to test the effectiveness of PM+ Group in reducing symptoms of psychological distress (anxiety and depression) in women in Swat. The secondary objectives are to assess improvements in levels of functioning, major depression, symptoms of post-traumatic stress disorder (PTSD), perceived social support, and economic outcomes. To complement this quantitative data, a qualitative component will explore intervention acceptability amongst stakeholders including participants, families, and PM+ Group facilitators.

The primary hypothesis is that women receiving PM+ Group will achieve lower psychological distress (depression and anxiety) scores in comparison to the EUC control group at 20 weeks. Secondary hypotheses are that:

This is a two-arm, single-blind, cluster randomised controlled superiority trial conducted in a community-based setting with women in rural Pakistan. A cRCT was favoured to reduce the likelihood of contamination between the two groups. The Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) figure 2013 is provided in Fig. 1, and the SPIRIT Checklist accompanies this paper as Additional file 1.

This community-based study is being conducted in Odigram and Ghalegay Union Councils in Swat District, Pakistan. A Union Council is the smallest administrative unit in Pakistan, each with a population of approximately 25,000. Each Union Council has a Basic Health Unit (BHU) – a primary health care facility staffed by a primary care physician, a Lady Health Visitor, a vaccinator, and 15–20 community-based LHWs. LHWs are female community health workers trained to provide maternal and child health care. They receive no training in the assessment and management of common mental health problems and typically have 10 years of education. Employed by the Government, each LHW is responsible for a catchment area of approximately 1000 people or 150 homes, visiting five to seven homes daily, covering approximately 85% of the population [24, 25]. Odigram and Ghalegay Union Councils have 23 and 22 LHWs respectively. These LHW catchment areas constitute our cluster sampling frame from which 34 LHW catchment areas across the two Union Councils are involved in this study. The LHW role in supporting this study is explained below.

PM+ is a brief WHO psychological intervention based on evidence of established behavioural techniques for addressing symptoms of common mental health problems in low- and middle-income countries. The manualised intervention involves empirically supported strategies of problem-solving, behavioural activation, accessing social support, and stress management training [23, 26]. Prior to implementation, a period of formative work was undertaken to contextually adapt the intervention and training materials for delivery to women in Swat, Pakistan, and a pilot trial conducted [27].

PM+ Group is an adaptation of the individual PM+ intervention. It was developed through 6 months of formative work, including review by 28 international experts. The group intervention (manual available upon request) involves five weekly sessions each lasting for 3 h, inclusive of breaks. The first session opens with psychoeducation, including information on common reactions to adversity, the rationale for PM+, goal setting, and brief motivational interviewing. Sessions 1 to 4 each introduce a strategy: (1) Managing Stress, (2) Managing Problems, (3) Get Going, Keep Doing (i.e., behavioural activation), and (4) Strengthening Social Support that are applied by participants to problems that they are facing. Each strategy is reviewed in subsequent sessions, with application of strategies between sessions encouraged to enhance learning through repetition. The final session involves a revision of learning, education on preventing relapse, and ends with a culturally appropriate closing ceremony. To enhance accessibility for illiterate individuals the programme is structured around locally relevant and appropriate pictorial materials and adopts a narrative format to support engagement and individual disclosure of personal difficulties which can be more difficult in a group format. Specifically, a case example of a woman experiencing common practical and emotional problems is shared each week, with participants following her progress through PM+ Group.

LHWs provide logistical support to PM+ Group facilitators, including the provision of a room in her house in which to conduct sessions, and convening sessions with participants from her catchment area. LHWs do not deliver the intervention, this is done by female facilitators. The PM+ Group female facilitators are recruited from Swat and have a minimum of 16 years of education and no previous experience in providing mental health care. Facilitators received 7 days of intervention training, delivered by the master trainer (KSD) and supported by three in-country supervisors (PA, HN, AM). Prior to attending training in PM+, facilitators underwent a half-day orientation to Psychological First Aid (Urdu version) [28] to sensitise them on how to appropriately respond to current crises a woman may be experiencing. Intervention training included education on adversity and its impact upon mental health, basic counselling skills, delivering PM+, skills in group facilitation, and facilitator self-care. An additional half-day security training for working in unstable settings was also provided. To build confidence and competence following training, all facilitators delivered one practice group each at an accelerated rate (five sessions in 2 weeks), with participants living outside the trial area, under intensive supervision. After conclusion of practice groups all facilitators underwent competency assessments prior to delivery of PM+ Group to cRCT participants. In case of insufficient competency additional targeted training is provided and competency assessments readministered. A co-facilitator who occupies a support role (e.g., providing child care and supporting LHWs with logistical arrangements) supports facilitators in each group. Co-facilitators participated in a half-day orientation to their role delivered by the facilitators.

Consistent with an apprenticeship model [16], protocol adherence is ensured through weekly supervision of the facilitators provided by three trained PM+ Group supervisors based in Islamabad; and fortnightly supervision of the supervisors by the master trainer. Involving all facilitators in a group, supervision lasts 2 to 3 h, conducted via Skype. Supervision entails reviewing the progress of groups including case management of participants and additional refresher training on intervention components and group facilitation skills through role play. The supervisors are in turn provided with supervision by the master trainer, conducted fortnightly via Skype and also lasting 2 to 3 h. Intervention fidelity is monitored through independent observations of 15% of randomly selected sessions of each facilitator against tailored checklists. These fidelity checks are conducted by persons trained in PM+ Group but not acting as study facilitators or supervisors.

Control and intervention arm participants will continue to receive routine LHW visits on an individual basis. The care that they receive is considered to be EUC as usual care for common mental disorders in primary care in Pakistan typically involves no or placebo care leaving the majority of cases of distress undetected and unsupported. EUC will comprise:

Both arms have unrestricted access to EUC through their routine health providers. The intensity of services received in both arms will be measured at 20 weeks as part of the economic evaluation (see below).

Similar community-based intervention studies using change in symptom-based questionnaires, like the Hospital Anxiety and Depression Scale (HADS) [35‐37], have used effect sizes of at least 0.4 when testing treatment as usual groups with limited or no active therapeutic elements, as we propose. Assuming an effect size of 0.4 for the primary endpoint (HADS total score at 20 weeks), with 90% power and 5% significance, an intracluster correlation coefficient of 0.05, and a two-sided hypothesis test with 34 health worker catchment area clusters randomised at a 1:1 allocation ratio, and accounting for 20% attrition, we will need 612 participants (i.e. 306 in each arm). This means that we will need on average 18 participants from each of the 34 health worker catchment areas. A pilot study conducted a nearby Union Council found that these recruitment numbers were feasible and achievable [27], and previous research in the same setting identified a high prevalence of mental health problems in women [38].

Thirty-four LHW clusters will be randomised to the intervention and control arms on a 1:1 allocation ratio using a permuted-block randomisation method (see Fig. 2). A randomisation list will be generated by an independent statistician using SAS PROC PLAN. Allocation of clusters will be carried out by an independent person based at the Human Development Research Foundation, Islamabad. This is achieved by the research team sharing a list of LHWs each with a corresponding number. Allocation status will be communicated to the research team coordinator who will inform respective LHWs after consent for trial participation is obtained from participants by researchers.

If a catchment area is allocated to the intervention arm, the LHW contacts all participants within her cluster to schedule an initial meeting to introduce the participant to the intervention facilitator. This is a necessary engagement step in this setting where service uptake is dependent upon a trusted relationship with the service provider. If allocated to the control arm the LHW continues routine visits to women in her catchment area.

Due to the nature of the intervention it is not possible to mask participants, LHWs, and intervention facilitators and co-facilitators to treatment allocation. All researchers conducting quantitative outcome assessments will be masked, while the qualitative research team will be unmasked. Researchers conducting assessments are residents of the local Swat District and remain distinct to the intervention facilitation and co-facilitator teams with separate offices, logistical arrangements, and administrative management. They are trained and supervised by the site principal investigator (PI; NK).

Prior to each assessment point participants are instructed by LHWs not to disclose to researchers their allocation status. The researcher completes a form to guess which arm of the trial the participant belongs to both preceding and following the conduct of each assessment. The order of assessments ensures that the primary outcome measure (HADS) is administered first to minimise the risk of bias should masking be compromised. If unmasking does occur this will be documented, the assessment halted, and a new researcher assigned to complete assessments with that participant. The CSRI is administered as the final instrument at the 20-week assessment to minimise the impact upon assessments should responses lead to unmasking.

The trial statistician (DW) is blinded to the treatment code when developing the statistical analysis plan and writing the statistical programmes, which will be validated and completed using dummy randomisation codes. The actual allocation will only be provided after locking of the database.

Quantitative data will be completed on paper assessment booklets with assigned participant codes, stored at the research field office at the end of each day. Daily checking of data will be performed by the research coordinator, with queries identified and resolved promptly. Double data entry will be conducted by an assigned data entry team at the Human Development Research Foundation (HDRF) in Islamabad, with discrepancies resolved by a third data entry person. Once in an electronic file, all data will be password-protected, with data managers controlling access to the passwords and the database backed up daily.

Qualitative data will be stored in paper format in locked filing cabinets in the research field office at the end of each day. None of the qualitative data will contain identifiers (name, age, category of respondent, etc.) that may compromise participant anonymity.

All other process data (i.e. supervision forms) will be stored in locked filing cabinets in the intervention field office. Intervention team members have been trained in the importance of de-identifying all notes relating to participant progress through the intervention to ensure confidentiality.

Data will be analysed using SAS 9.3 and SPSS Version 21. Findings will be reported according to the Consolidated Standards of Reporting Trials (CONSORT) guidelines for cRCTs [56]. Primary analyses will be based on the intention-to-treat population and secondary analyses will be based on the per-protocol population. A linear mixed model will be employed for the primary endpoint analysis. The mixed model will have treatment, visit, interaction between treatment and visit as fixed effects, baseline measurement of HADS as covariate, and cluster and subject as random effects. The mean difference between two treatment arms at each visit, together with its 95% confidence interval (CI), will be derived from the mixed model. Covariate-adjusted mixed model of primary endpoint will also be performed by adding prespecified covariates at baseline into the above model. Missing data will be treated as missing at random in the mixed model analysis and no imputation of primary endpoint will be made. To assess the sensitivity of the result to this assumption, the last observation carried forward strategy will be used to compute missing primary endpoints. Subgroup analysis will be performed on the prespecified covariates.

Continuous secondary outcomes will be analysed in a similar way as the primary endpoint analysis. For the analysis of binary secondary outcomes, a generalised mixed model will be employed with treatment, visit, interaction between treatment and visit as fixed effects, baseline measurement as covariate, and cluster and subject as random effect. The odds ratio between two treatment arms at each visit together with its 95% CI will be derived from the generalised mixed model.

All analyses will be described in detail in the finalised and signed statistical analysis plan before unmasking the study.

The trial protocol has received ethical approval from the Institute of Psychiatry, Benazir Bhutto Hospital, Rawalpindi, Pakistan and the World Health Organisation Ethical Review Board. The study has also received support from the Swat District Health Department and People’s Primary Healthcare, and from the Psychiatry Department at Saidu Teaching Hospital, Swat.

Trial recruitment commenced December 2015 and, at the time of manuscript submission, was ongoing. Results of this study are expected in 2017.