Study setting and participant recruitment

Data for this study were obtained from a single-blinded randomized controlled trial to study the effectiveness of the Healthy Mother—Happy Baby (HMHB) intervention to reduce anxiety among pregnant women (clinicaltrial.gov identifier: NCT03880032) [17]. The study recruited 1200 women from Holy Family Hospital (HFH), a public facility in Rawalpindi, Pakistan, between 16th April 2019 until 31 January 2022. HFH is located in Rawalpindi, Pakistan, is a large facility with around 900 beds, making it a major regional healthcare facility. Annually, around 5,000 births are delivered at HFH, evidence of its crucial role in maternal and neonatal care in the region where it serves a diverse population from urban, rural, and semi-urban areas. All participants were recruited by female research assistants in the outpatient Gynaecology and Obstetrics Department during their initial prenatal visit. Participants were followed up at six-weeks after birth.

Screening and inclusion criteria

The study employed three levels of inclusion/exclusion screening criteria during the enrolment process. In the first level, women had to be at ≤ 22 weeks' gestation, ≥ 18 years old, reside ≤ 20 km from HFH, and have a basic understanding of Urdu. Women who met these criteria and showed willingness to participate were asked to provide informed consent. At the second level of screening, potential participants were excluded if reporting life-threatening health conditions, such as active severe depression or suicidal ideation. Other exclusions included self-reported significant learning disabilities, a self-reported psychiatric disorder or ongoing psychiatric care, medical disorders or severe maternal morbidity requiring inpatient management, and ICU admission indicated by diagnosis (not solely for assessment purposes), past or current significant learning disabilities, past or current psychiatric disorders, medical disorders, or severe maternal morbidity. At the third level of screening, potential participants were assessed for the presence of at least mild anxiety using the Hospital Anxiety Depression Scale (HADS) screening tool. Those who scored ≥ 8 on the HADS anxiety sub-scale (indicating at least mild anxiety) were interviewed by trained assessors who conducted a Structured Clinical Interview for DSM IV Diagnoses (SCID) to rule out depression. Women who met the conditions for a major depressive episode (MDE) were not included. MDE was defined using a diagnostic semi-structured guide from the Structured Clinical Interview for the Diagnostic and Statistical Manual (SCID), which is based on American Psychiatric Association's Diagnostic and Statistical Manual for Mental Disorders (DSM). Assessment with this method is considered equivalent to a clinical diagnosis in line with the DSM criteria.

Randomization

Study participants were randomly assigned to either the intervention or control group using a pseudo random-number generator. The random sequence was assigned through permuted blocks of size 4, 8, 12, and 16. The assignment list was printed in order, with each assignment separately stored in opaque envelopes and numbered sequentially. Once an eligible individual consented to participate in the study, the research team proceeded by selecting the next available envelope to determine the individual's assignment into either the intervention or control arm. Throughout this process, the trial team, comprising the assessment team, principal investigators, and co-investigators, remained blinded to the allocation.

Study groups

Those randomized to the intervention group received the HMHB program, an intervention relying on principles of Cognitive Behavioural Therapy (CBT) that aimed to reduce symptoms of anxiety during pregnancy [18]. It was adapted from WHO endorsed psychosocial intervention for perinatal depression called the Thinking Healthy Programme [19]. The intervention was delivered by non-specialist providers (with a two-year bachelor’s degree and a two-year master’s degree in psychology but no clinical experience). They were trained on the intervention and received regular weekly group supervision over the trial period. HMHB was designed to target risk factors for anxiety that women experienced during pregnancy that were identified in our preliminary research [18]. It also integrated stress management techniques, such as breathing exercises. To make the intervention more culturally appropriate, personalized illustrations were used in order to facilitate guided discovery, behavioural activation, stress reduction, and convey essential health messages [18]. The sessions were supplemented by take-home exercises.

HMHB consisted of six core weekly sessions and up to six optional booster sessions (delivered as needed). The first five weekly one-on-one sessions were intended for early to mid-pregnancy. The final sixth core session was given in the third trimester of pregnancy. This session was aimed to help manage anxiety during late pregnancy, prepare for baby’s arrival, and navigate the early post-natal period. It highlighted the importance of breastfeeding and providing colostrum as a pre-lacteal feed instead of culturally common practices involving feeding infants honey or herbal tonics. It also encouraged family support for mothers to breastfeed.

The control group received enhanced usual care at the Gynaecology and Obstetrics Department. Usual care recommended at the study hospital typically involves up to eight visits for evaluating health status, discussing any concerns, and performing routine exams consistent with the stage of pregnancy. The care of women participating in HMHB was enhanced by reminders for study visits, expedited care (shorter wait times), as well as reimbursement for transportation to visits and for as many ultrasounds as were medically indicated at HFH during pregnancy.

Breastfeeding indicators

In line with WHO definition of exclusive breastfeeding, women who confirmed providing only colostrum/breastmilk within the first 24 h and reported no use of formula, Ghutti (traditional pre-lacteal feed), herbal water, tea, or other animal milk were considered to have engaged in ‘early exclusive breastfeeding’. Breastfeeding women who reported both breastmilk and at least one other nutritional source fell into the ‘early breastfeeding’ category. We also assessed breastfeeding at six weeks postpartum by asking mothers if they were breastfeeding and if they had given breast milk or any other nutrition to their infants to determine whether it was exclusive or non-exclusive. These women were categorized as having engaged in ‘recent exclusive breastfeeding’ if no other nutrition source was provided or ‘recent breastfeeding’ if receiving both breastmilk and other nutrients. Our indicators for 'recent' breastfeeding practices, namely 'recent exclusive breastfeeding' and 'recent breastfeeding', pertain to exclusive and non-exclusive breastfeeding within 24 h before the six-week postpartum assessment. We chose the term 'recent' because it covers both exclusively breastfed infants and those receiving some breastmilk, unlike the WHO definition of 'predominant breastfeeding', which focuses solely on the infant's main source of nourishment.

CovariatesThe hospital anxiety and depression scale

The Hospital Anxiety and Depression Scale (HADS) is a well-known instrument that includes 14 items rated on a 4-point scale that has been validated in numerous languages and settings [20, 21]. The HADS focuses on non-physical symptoms to screen for anxiety and depression, but does not include all of the diagnostic criteria of depression specified in the Diagnostic and Statistical Manual of Mental Disorders (DSM) [21]. For example, it does not include items on appetite, sleep and self-harm/suicidal thoughts [21]. It comprises two distinct subscales to assess anxiety and depression, each containing 7 items with scores ranging from 0 to 21. Typical symptom cut-offs are 0–7 (normal), 8–10 (mild), 11–14 (moderate), and 15–21 (severe). A cut-off of ≥ 8 was defined as the threshold for being ‘at risk’ for anxiety or depression. The Urdu adaptation of this scale has been previously modified for use in Pakistan and has been administered successfully [22] including in pregnant women [23], showing satisfactory reliability, validity, and high concurrence with the English version for use of the symptom threshold of ≥ 8 [24] when assessing antenatal anxiety and depression in Pakistan [25].

Perceived stress scale (PSS-10)

The PSS-10 is a validated global measure of perceived stress that consists of 10 items scored on a 5-point scale ranging from 0 to 4 (maximum score 40) [26]; higher scores indicate more stress. It has been adapted for use in Pakistan [27]. A score of ≥ 20 corresponds to high stress.

Multi-dimensional scale of social support (MSPSS)

The Multi-dimensional Scale of Perceived Social Support (MSPSS) is a 12-item measure of subjective availability of support (primarily emotional) which has been validated and successfully adapted to the Pakistani context [28]. Scores are on a 7-point scale (1 = very strongly disagree; 7 = very strongly agree), with higher scores indicating more support.

Statistical analyses

Descriptive statistics were performed to investigate postpartum symptoms of depression and anxiety in relation to breastfeeding. In women with completed measures of breastfeeding outcomes, we examined the differences at baseline between intervention and control arms to verify that randomization generated comparable arms. We used standard statistical comparisons, including Chi-square test for categorical factors and Student’s t-test for continuous factors, to determine the statistical significance of any differences between arms. All analyses followed the intent-to-treat (ITT) principle unless otherwise noted, comparing the four breastfeeding outcomes in the groups to which they were randomized.

We compared early breastfeeding, exclusive early breastfeeding, recent breastfeeding, and exclusive recent breastfeeding outcomes between arms, where ‘early’ was defined as first 24 h after childbirth and ‘recent’ was defined as within the last 24 h before the assessment that occurred approximately six weeks postpartum. Comparisons between arms following the principle of ITT were estimated with logistic regression. In addition to the intent-to-treat analysis among all participants, we also performed a stratified analysis to separately estimate the intervention effects for women who had mild anxiety levels (HADS anxiety score: 8–11) and the intervention effects for women who had moderate to severe anxiety levels (HADS anxiety score: 11–21) at enrolment.

We also conducted an analysis to examine breastfeeding outcomes in a subset of women randomized to the intervention arm who only included those that received all six intervention sessions (“intervention completers”). We adjusted for potential confounders including gestational age, depression and anxiety at enrolment, stress at enrolment measured with the Perceived Stress Scale (PSS-10), general social support and social support from family (both measured with the Multidimensional Scale of Social Support (MSPSS)), maternal age, child’s sex, whether first pregnancy or not and history of stillbirth or miscarriage. Selection of confounding factors for adjustment was based on prior knowledge of what was expected to influence intervention session receipt and by examining the baseline variables for those with six intervention sessions compared to those in the control arm. Given breastfeeding was a secondary outcome, this study was not specifically powered to detect differences related to breastfeeding outcomes. Rather, it was powered to detect a difference in the six-week postpartum mental health outcomes of participants between arms among 1,200 enrolled participants with a 30% expected dropout rate.

Finally, we evaluated a dose response relationship for the intervention considering the receipt of booster sessions. Using the Cochrane Armitage test for trend, we examined this relationship for the four types of breastfeeding considered, across three dose groups, 1) control (no intervention), 2) only core intervention sessions, and 3) core and booster sessions.

Ethics

Ethical approval for this research was received from the Institutional Review Boards of the Johns Hopkins Bloomberg School of Health (IRB No. 00009177; Approved April 2, 2019), the Human Development Research Foundation Ethics Committee (IRB/001/2017; Approval March 10, 2017), and Global Mental Health Data Safety and Monitoring Board appointed by the National Institute of Mental Health (NIMH) in the United States (No assigned approval number; Approved March 11, 2019). Prior to their involvement in this study, all participants provided written informed consent, indicating their willingness to take part in the research.