Interpersonal Counselling Versus Perinatal-speci c Cognitive Behavioural Therapy for Women with Depression during Pregnancy Offered in Routine Psychological Treatment Services: A Phase II Trial

Jonathan Evans (  j.evans@bristol.ac.uk ) University of Bristol Jenny Ingram University of Bristol Roslyn Law Anna Freud Centre Hazel Taylor University Hospitals Bristol NHS Foundation Trust Debbie Johnson University of Bristol Joel Glynn University of Bristol Becky Hopley University of Exeter David Kessler University of Bristol Jeff Round University of Bristol Jenny Ford University Hospitals Bristol NHS Foundation Trust Iryna Culpin University of Bristol Heather O'Mahen University of Exeter

during pregnancy and concern amongst clinicians about prescribing them, meaning that psychological interventions are particularly important at this time (Sockol et al 2011).
The current treatment recommendation for mild to moderate depression during pregnancy, is supported self-help approach using the principles of cognitive behavioural therapy (CBT) (NICE 2014). However, the provision and take up of CBT treatment for antenatal depression is low. Women who are pregnant receive lower rates of mental health interventions than those outside the perinatal period (30% vs 50%), and most of these treatments involve medication (McManus et al 2014). This is concerning given the worries pregnant women have about medication. Furthermore, there are important limitations of CBT for pregnancy as a model for less experienced therapists because: 1) without adaptation and additional training, CBT has few explicit strategies to manage many of the problems that are common for women with antenatal depression, including role transitions and problems in communication. This is critical as con ict in relationships and poor social support are the strongest risk factors for antenatal depression (Lancaster 2012). 2) low intensity CBT does not manage problems with grief (i.e., miscarriage, still-birth, termination, loss of parents at time of birth of child) that contributes to depressive symptoms. 3) CBT is not designed to involve the partner.
In contrast, Interpersonal Counselling (IPC) a brief treatment derived from Interpersonal Therapy (IPT) is based on relational theory. It acknowledges that although depression is multicausal, a key precipitating factor occurs when problems in interpersonal relationships trigger symptoms of depression, such as low mood or sleeplessness, and these symptoms further compromise these relationships. It uniquely focuses on approaches that help manage changes in role and losses (e.g., miscarriage, still-birth, termination, previous loss of would be grandparents) and the impact of these on relationships and mood. By directly approaching these issues, IPC addresses what service users report are signi cant worries and concerns for them during pregnancy. Furthermore, the brief nature of IPC makes it appropriate for mild to moderate forms of depression, and can be delivered by those without specialist mental health training, thus making acceptable and appropriate psychological interventions for pregnant women more widely available.
IPC has been clearly developed and manualised as an intervention allowing assessment of delity to the model. Although there have been very few studies of the effectiveness of IPC and none in pregnancy, one study of depression in primary care in Italy found it to be more effective than antidepressants (selective serotonin reuptake inhibitors) particularly for those with less severe depression (Menchetti et al 2013). A small feasibility study of IPC for antenatal depression in the US amongst low income mothers indicated high satisfaction with IPC and some improvement in mood (Lenze et al 2017).
There are therefore good reasons to hypothesise that this form of therapy would be more acceptable during the antenatal period and particularly effective in treating depression at this time. We aimed to test the hypothesis that a large scale RCT of IPC in routine NHS services is feasible. Intervention IPC involved up to six 30-45 minute sessions, with the option of inviting a partner or signi cant other to one of the sessions. The tasks of IPC included identifying symptoms of depression and relating these to interpersonal problems, identifying protective factors and vulnerabilities, agreeing the focus and identifying who will assist in promoting recovery. IPC is structured around one of four focal areas; grief, role transitions, role disputes or interpersonal sensitivities. The focal area is collaboratively chosen to address the primary interpersonal context in which depression occurs and common strategies focusing on improving communication and processing emotions are combined with focus speci c interventions e.g. mourning the loss of a signi cant role or identifying key issues and different expectations in a dispute. The individual's social network is actively engaged and guided to understand and contribute to recovery from depression. meeting criteria for severe depression (ICD-10) according to CIS-R, and those who had received CBT or IPT within the last 6 months.

Recruitment
This took place between 1st January 2019 and 30th September 2019. We aimed to compare the feasibility of two different recruitment methods: Method 1: Midwife booking clinics. Women who screened positive on a two-item depression case nding instrument (Whooley et al 1997) used routinely by midwives at their rst appointment with the midwife (around 12 weeks of pregnancy) and who consented to be contacted were referred to the research team.

Method 2:
Women at the ultrasound scan clinics (12 and 20 week scans) were also given study information, screening questions (EPDS score with 10 or above a positive screen) and a consent to contact form from either a research assistant or administrative assistant staff.
Eligible women were then asked to complete a face-to-face assessment with a research assistant to establish eligibility and obtain consent for the study and collect baseline data. Partners were recruited in person at this baseline visit, online or by mail to complete measurements of their mood. Randomisation will be carried out remotely by Bristol Randomised Trials Collaboration randomisation service.
Randomisation was 1:1 and strati ed by recruiting centre and minimised by parity (with random block sizes). Follow-up data at 12 weeks post randomisation were collected online.
Following randomisation women received a routine clinical assessment with the IAPT service and those who were considered to be clinically appropriate for a low intensity intervention with the IAPT service, were allocated to a PWP for either IPC or perinatal-speci c CBT. Both treatments were low intensity individual psychological interventions of up to six sessions each between 30-45 minutes long.

Assessments
The assessment measures at baseline measured mood, relationship quality and quality of life. The CIS-R a computerised structured diagnostic interview (Lewis et al 1992) was completed at baseline only. This provided and ICD-10 diagnosis and con rmed eligibility to the study (mild or moderate depression). The These measures were completed at baseline and 12 weeks post randomisation with the exception of the CIS-R which was completed at baseline only.
All 12-week measures were collected on-line with participants prompted via e mail messages. In addition, at 12-weeks, data were collected from the jMHWs on the number of sessions attended, number of sessions with a partner, whether therapy was considered by the jMHW complete, if individuals were provided with a referral for additional treatment at a higher intensity within the service. At the end of the study, supervisors rated each jMHW using a four-item checklist assessing their ability to follow the IPC model.

Data Analyses
As this is a feasibility study, it was not powered for statistical testing of any difference between the interventions. The analyses focus on reporting data that will be used for planning and assessing the feasibility of the full scale trial. Feasibility measures reported include the number and proportions of eligible, recruited, randomised, started treatment, completed treatment, lost to follow-up, completing follow-up and completing individual outcome measures. The primary outcome was the number of eligible women successfully recruited to the point of randomisation. We set a recruitment target of 60 subjects from 338 potentially eligible women (17.8%) during the nine-month recruitment period giving a 95% con dence interval for recruitment of (13.9%, 22.3%).
Descriptive statistics for clinical and health economic outcome measures considered for the main trial are reported according to allocation to IPC or CBT .

Nested qualitative study
We assessed the acceptability of the recruitment method, intervention and study design through a series of in-depth interviews with participants and staff. Women who received therapy (both IPC and CBT) were interviewed and those who dropped out of the process before therapy. A few partners; the jMHWs who were trained in IPC; and the managers from both services were also interviewed. Interview topic guides were informed by the research literature, team discussions and input from our PPI groups. Interviews were recorded, transcribed verbatim by a professional transcription service and anonymised. Thematic analysis of the data was an ongoing and iterative process using NVivo software to organise and code the transcripts .[ QSR International Pty Ltd] Codes and themes were developed and discussed within the qualitative research team at regular intervals during both data collection and analysis, to achieve consensus ( Braun and Clarke 2006).

Recruitment of services and training
The local IAPT services in Bristol and Exeter agreed to participate and 12 jMHWs were allocated randomly, six to IPC training and six to perinatal-speci c CBT training.
Recruitment of women during pregnancy and outcome data collection A total of 1173 potentially eligible women were approached during the 9-month recruitment period with 1128 through the scanning clinics and 45 through midwives. (see Consort diagram Fig. 1) Of the 606 who completed the EPDS, 237 (39%) had an EPDS score of 10 or more. Of these 237 potentially eligible women, 106 (44%) were assessed at baseline and of these 52 (49%) were eligible, meeting all criteria for the study including a diagnosis of mild or moderate depression according to the CIS-R; they were randomised to IPC or perinatal-speci c CBT. Of the 52 randomised, 42 (81%) provided outcome data at 12 weeks post randomisation.
The baseline characteristics are shown in Table 1. Women were recruited at a median of 15 weeks of pregnancy and had a mean age of 31.4 and 54% were primiparous. It is notable that there was a selection bias according to educational level with 73%, having a degree.
As very few partners were recruited to the study (n = 13), data are omitted for simplicity.

Outcomes
The outcomes relevant to the feasibility of the trial are shown in table 2. In summary, most recruits came though scanning clinics compared to midwife booking clinics. Of the 26 randomised to IPC, 2 started CBT as timing was more convenient to them. Of the 17 who started IPC, 12 (71%) completed ve or more sessions and were considered to have completed treatment. Of the 52 women randomised, two women, one from each study arm, were stepped up to a higher intensity intervention after completing treatment.
Health economic measures Response rates for all economic measures were high enough to be con dent of collecting the necessary data needed to undertake an economic evaluation alongside a future de nitive trial.
Both groups recorded an improvement in health-related quality of life (HRQoL) as measured by ReQoL-10 a measure developed to assess the quality of life of people with different mental health conditions. The mean increase in ReQoL-10 score was 1.9 (s.d. 5.5) for the IPC group and 2.9 (s.d. 5.4) for the CBT group.
Both groups recorded a drop in HRQoL scores as measured by EQ-5D-5L the NICE recommended QoL measure for economic evaluation which captures both mental and physical attributes of QoL. The mean change in EQ-5D-5L utility scores was − 0.038 (s.d. 0.16) for the IPC group and − 0.031 (s.d. 0.087) for the CBT group. Scores fell due to the natural decline in the physical health domains as gestation increases, however this fall was lessened by increased scores in the mental health domain.
The Resource Use Questionnaire (RUQ) was costed providing a total cost for both treatment and nontreatment related costs. The total cost per patient including treatment for IPC was £774 (s.d. £1032) and for CBT £539 (s.d. £492).
Acceptability from qualitative interviews: Twenty-three women were interviewed for the nested qualitative study: 19 who received therapy and four who dropped out before therapy. Three partners were interviewed; six jMHWs who were trained in IPC; and the managers from both services.
Overall, all participants and partners interviewed found the study screening and recruitment process highly acceptable including the completion of the outcome measures online.
Several women found it quite di cult to understand their emotions in early pregnancy with many nding it an anxious time. Those who had no previous experience of mental health problems, reported being reluctant to discuss such issues with a midwife at their booking appointment, so "this study was a lifeline" for them.
Scanning clinics were felt to be a good place to read the study lea et, with time in the waiting room to think about it or discuss with a partner. It gave them "space to acknowledge their low mood". It was also a more con dential way of admitting to low mood than talking to a midwife.
Most of those interviewed understood the concept of randomisation; a few expressed an expectation that the allocated therapy would be the "right one for them". The recruitment appointment also enabled them to "acknowledge that things were not right" for them.
Psychological services "fast-tracked" pregnant women (ie. saw them within 4 weeks of referral), and many were pleased with how quickly and e ciently their sessions were organised. However, women were reluctant to take time off work to attend sessions and combined with limited provision of 'out of hours' sessions and restricted venue options, they sometimes had to wait for several weeks before starting their treatment course.
Most women liked seeing their therapist face-to-face, and for some the commitment of having an appointment to attend outside the home, motivated attendance. Those who received telephone-only therapy reported that a lack of rapport meant that it took longer to 'open up' or 'get to know' their therapist and feel relaxed. Most agreed that offering a combination of face-to-face and telephone sessions would be the most acceptable method of service provision. There were no video sessions at the time of the study.
Women found both treatments focused on practical issues, offered 'tools for life' and they appreciated being given exercises to complete and handouts that they could refer to later.
Women described IPC as being "helpful","practical" and empowered them to "ask for help". They particularly mentioned goal setting, mapping their circle of support and communication skills with partner as being very useful. CBT had a more structured and task-based therapist-led approach which was appreciated by some women and they were generally happy with it.

Partners
There were many reasons why partners did not attend IPC sessions as intended. jMHWs were not used to conducting joint sessions and felt uncomfortable with the idea, but also women did not always feel it was relevant to ask them, whether due to work issues, needing them to provide childcare or just not being happy with the idea themselves. However, those men who were interviewed saw the intervention as positive and helpful particularly for relationships.
At rst, some jMHWs felt under-prepared and anxious when delivering IPC therapy, which was a new treatment to them. However, they felt that IPC was a very appropriate therapy for pregnant women especially concerning partners and relationships, and found the IPC approach 'more human'. JMHWs reported that it was sometimes di cult to keep sessions to their allotted length, but felt they learnt much from it and gained in con dence throughout the study (see Table 4). Scanning clinics: 'Space to acknowledge low mood' That was a good time for me actually to have it, yeah, and actually it's quite good if you're giving something while you're in a waiting area because it's something to do while you're waiting rather than when you're rushing around, so it was quite a good time to look at it. (1034, site A, primip) …it might not have been something I spoke to my midwife about, I might not have….if I hadn't been directly asked I might not have answered those questions in that respect. This was a con dential way […] that I could do it in my own time to refSlect in how I was feeling' (1021, site A, multip) The recruitment appointment: 'Recognising things are not right'.
It felt good to just talk about how I had been feeling with someone, actually. I know that you were doing it for the purpose of nding out which help I needed, but even sometimes just sharing your thoughts with someone can help. I have to admit I was nervous about any implications, I was worried whether I would come across as a bad mum or whether it would lead to any intervention from any government bodies, but overall it felt good to be able to share how I was feeling with someone. I think face to face is obviously better, you've got your eye contact and it just feels different I think face to face, but I am sure telephone will be more helpful than not having it. But I would say face to face is preferable, just to have that rapport as well with the therapist I think helps. (1034, site A,  …knowing who is in the support group and [name] re ected on that actually, and there was a chart that she gave me where I could write who it was, my relationship with them, and what good they bring me, and can I rely on them for emotional and physical support, that was really helpful to go through, to know who I had and who I could rely on. (2005, site B, primip, IPC) Junior Mental Health Worker views "IPC approach is 'more human'' There was freedom and a different focus [in IPC therapy], focus still on depression but there was focus on relationship that it's not really the main thing in CBT, and I think a lot of clients that I worked with found that helpful and having space to talk about things a bit more freely it seemed like it was helpful. (jMHW, Site A) Different [from CBT]… I think the relationship aspect of it was really helpful, who's in the networks and who is going to support you and… what are you going to do to help… what skills can we develop to help you get that… express your needs, I nd that so helpful, especially for pregnant women, and also new babies as well (jMHW site B) Discussion This is the rst study to assess the feasibility of a trial and acceptability of IPC for antenatal depression.
We found that the trial was both feasible and acceptable.
We found that recruitment using the routine pathway through midwives is not feasible but recruitment through scanning clinics is both feasible and acceptable to women. This raises important questions about the identi cation of less severe mental health problems in maternity services. Currently, midwives ask the Whooley questions face-to-face to identify women who may have antenatal depression (Nice 2014). This screening method has high sensitivity (95%) and modest speci city (65%) when used as a self-report tool for antenatal depression (Bosanquet et al 2015). However, using these questions face-toface has a much lower sensitivity with lower depression detection (7.5%) when used in midwife run booking clinics and thus low referral rates in this study. This low rate of detection was also reported in another recent UK study (10%, Howard et al 2018). Anecdotal evidence suggests that women are not always aware that they can be referred for further help following a positive Whooley screen (Darwin et al 2015). In contrast women were willing to complete a depression scale whilst waiting in the scanning clinic and around 20% of those who completed the scale scored above the threshold indicating at least mild depression. This rate is consistent with other studies using the EPDS (Pearson et al 2018).
Therefore, we will recruit from scanning clinics in a full-scale trial of IPC.
IPC is a recently developed brief intervention derived from a more intensive treatment, Interpersonal therapy (IPT). It focuses on the importance of relationships in the onset and maintenance of depression and addresses issues that are reported by women to be particularly important during pregnancy. Although a trial of the more intensive IPT has shown good effect for postnatal depression when delivered by telephone by specially trained nurses (Dennis et al 2020), we have shown here that IPC, can be delivered in routine psychological treatment services. We have also shown that the intervention can be taught to junior mental health practitioners without extensive previous clinical training following three days training and with supervision of cases. Although there was some anxiety about learning a new intervention, they found it very relevant to the perinatal context and quickly gained in con dence in delivering IPC throughout the 9-month course of treatment provision. This was echoed by the women who received the intervention who found particular aspects of the intervention helpful. Although very few partners completed questionnaires or attended sessions, those partners who were interviewed, saw the intervention as positive and helpful particularly for relationships.

Limitations
There was considerable drop out between randomisation and delivering the intervention with only 65% of those randomised starting a course of treatment. The main reason for this was that the psychological treatment service was unable to contact those who were referred. There is a practical question about how persistent this attempt to contact women should be. There is concern in psychological treatment services that those who do not respond to correspondence or telephone contact are unmotivated and therefore unlikely to attend or bene t from a psychological intervention. However, depression is demotivating and there is evidence that an assertive approach may be needed in some groups to engage them in treatment (Santiago et al 2013). A more assertive approach may be needed along with the prioritisation of perinatal women to the service to improve treatment up take.
We were not powered to detect a difference in outcome, and there was no evidence of difference between the two interventions. It is worth noting that the therapists in the comparison group received speci c perinatal training at the time of the intervention and a more pragmatic real-world design would compare IPC to standard usual care in psychological treatment services.

Conclusion
A trial of IPC taught to junior mental health workers in routine psychological treatment services in the UK is feasible and acceptable to women and staff. The progression criteria were met (see on-line supplement) and a full scale trial is now needed to evaluate the effectiveness and cost effectiveness of IPC for mild or moderate antenatal depression. Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing Interests
One of the authors RL developed the intervention and as such could not be considered entirely independent. No other competing interests.