Introduction

Psychiatric disorders during the perinatal period greatly affect mothers and children . Untreated depression has been associated with preterm birth, low-birth weight , dysfunctional parent–child interactions, inconsistent safety practices , and deficits in social and cognitive skills in children .

Estimated rates of antenatal major depression range between 8.5 and 11.0%, and postpartum rates are between 6.5 and 12.9% . Perinatal depression is associated with demographic factors, such as low income and unemployment, but is also dramatically affected by levels of social and partner support . Consequently, these social factors have frequently been targeted with psychosocial treatments, such as interpersonal psychotherapy (IPT) , which are designed to address these and other common perinatal issues.

The use of psychotherapy during the puerperium is empirically supported. More data support the efficacy of psychotherapy than medication in the puerperium . Additionally, the use of psychotherapy is safe compared with malformations or prematurity. Of greatest importance when considering safety, however, is treating perinatal depression. Although this can be optimally treated without drugs, the priority is the treatment of depression to remission . Moreover, psychotherapy is frequently viewied as first-choice treatment by perinatal women, in part reflecting their perceptions of the risk of medication during pregnancy and breast feeding .

Psychological treatments for perinatal depression

A recent meta-analysis showed that psychological interventions for depression achieve higher effect sizes if they are adapted for used with specific populations . A number of trials have shown psychological treatments to be efficacious for perinatal depression; however, to date, only manual-based, professionally delivered psychotherapies have been well validated .

Four major meta-analyses have examined psychological treatments for use with perinatal depression. Sockol et al. included pregnant and postpartum women. Data were analysed from 27 studies (nine open trials, two quasi-randomised and 16 randomised-controlled trials); both pharmacological and psychological interventions were represented. Across all studies, post-treatment levels of depressive symptoms were significantly reduced (overall effect size 0.65). Individual therapy was significantly more efficacious than group therapy, and among all therapies, cognitive–behaviour therapy (CBT) and IPT had the greatest effect sizes compared with control conditions, with a slight but statistically insignificant advantage for IPT. Despite the large number of data points, head-to-head comparisons were insufficient to separate the relative effects of psychotherapy and medication reliably.

The only other meta-analysis that included antepartum and postpartum women was conducted by Bledsoe and Grote They showed that psychotherapy, particularly CBT (group and individual) and IPT (alone or in combination with medications), was effective in reducing depressive symptoms. Higher effect sizes were achieved for the treatment of postpartum depression compared with antepartum depression. Their meta-analysis did not compare the effects of treatments to control conditions.

Dennis and Hodnett included 10 trials that compared psychological or psychosocial interventions for postpartum depression to treatment-as-usual controls. Professionally conducted psychotherapy and community-based psychosocial interventions both achieved 30% greater reduction in depression symptoms than treatment-as-usual. The authors reported that most of the studies were limited by relatively small sample sizes and lack of methodological rigor.

Cuijpers et al. conducted a meta-analysis of 17 psychological treatment trials for postpartum depression. Compared with control conditions, psychological treatments led to moderate improvement (effect size 0.61). Those trials which compared psychosocial interventions with wait-list control groups had higher effect sizes than those which were compared with treatment-as-usual (effect size 0.96 v 0.41). Only small differences between the effect sizes for different psychotherapy modalities were found, suggesting that common non-specific factors may mediate improvement for all psychosocial interventions for postpartum depression. The investigators were unable to draw conclusions about the long-term effects of psychotherapy owing to the lack of data beyond 6–12 months postpartum. This limitation was also noted by Sockol et al. .

Many of the studies in the meta-analyses have significant limitations. For example, only about one-half required women to meet diagnostic criteria for major or minor depression. Subject inclusion criteria in many were based solely on Edinburgh Postnatal Depression Scale (EPDS) scores above cut-offs ranging from 9 to 12. Many were also limited because they had small sample sizes and were underpowered.

Four studies stand out for their methodologic rigor, including large samples, use of DSM criteria for major depression for study entry, and longer-term follow up . O’Hara et al. randomised 120 women to either 12 sessions of IPT or to a waiting list control condition (WLC) over 12 weeks. Participants who had completed IPT ( n = 48) had significant reductions in depressive symptoms, higher rates of recovery, and higher levels of social adjustment relative to the WLC condition ( n = 51 completers). Women in the WLC condition were treated with IPT immediately after their waiting period, so no controlled follow up could be conducted. Ninety-seven women from both groups who ultimately received IPT, however, were followed for 18 months after treatment . Among women who had recovered, 43% experienced a recurrence during the following 18 months. Among those who did not recover with treatment, 84% subsequently recovered during the follow-up period.

Cooper et al. randomised 193 women to four conditions: non-directive counselling, cognitive counselling, psychodynamic therapy, or routine primary care. Routine primary care was described as ‘the normal care provided by the primary health care team (i.e. general practitioners and health visitors) with no additional input (apart from assessment) from the research team.’ No information was provided about medication usage. The effect of therapy was significant, but no differences were observed between the active treatment conditions. Follow-up assessments were conducted at 4.5 months, 13.5 months, and 5 years. No differences were found between groups at any of these follow-up assessments. Treatment relative to usual care was not effective in preventing relapse or recurrence.

Milgrom et al. randomised 192 women to CBT-based group therapy, group counselling, individual counselling, or routine primary care. Interventions lasted 12 weeks and included three partner sessions. Similar to Cooper et al. , the investigators found a significant effect for treatment but no differences among interventions.

Morrell et al. conducted a pragmatic cluster-randomised trial in primary care. The investigators randomised 101 general practices to usual care or to CBT-based counselling or listening visits. Health visitors were trained to identify PPD and deliver the counselling interventions to 595 women who scored over 11 on the EPDS at 6 weeks postpartum. A significant effect was found for treatment at 6 and 12 months, but no differences were found between the two active treatments.

Psychological treatments for perinatal depression

A recent meta-analysis showed that psychological interventions for depression achieve higher effect sizes if they are adapted for used with specific populations . A number of trials have shown psychological treatments to be efficacious for perinatal depression; however, to date, only manual-based, professionally delivered psychotherapies have been well validated .

Four major meta-analyses have examined psychological treatments for use with perinatal depression. Sockol et al. included pregnant and postpartum women. Data were analysed from 27 studies (nine open trials, two quasi-randomised and 16 randomised-controlled trials); both pharmacological and psychological interventions were represented. Across all studies, post-treatment levels of depressive symptoms were significantly reduced (overall effect size 0.65). Individual therapy was significantly more efficacious than group therapy, and among all therapies, cognitive–behaviour therapy (CBT) and IPT had the greatest effect sizes compared with control conditions, with a slight but statistically insignificant advantage for IPT. Despite the large number of data points, head-to-head comparisons were insufficient to separate the relative effects of psychotherapy and medication reliably.

The only other meta-analysis that included antepartum and postpartum women was conducted by Bledsoe and Grote They showed that psychotherapy, particularly CBT (group and individual) and IPT (alone or in combination with medications), was effective in reducing depressive symptoms. Higher effect sizes were achieved for the treatment of postpartum depression compared with antepartum depression. Their meta-analysis did not compare the effects of treatments to control conditions.

Dennis and Hodnett included 10 trials that compared psychological or psychosocial interventions for postpartum depression to treatment-as-usual controls. Professionally conducted psychotherapy and community-based psychosocial interventions both achieved 30% greater reduction in depression symptoms than treatment-as-usual. The authors reported that most of the studies were limited by relatively small sample sizes and lack of methodological rigor.

Cuijpers et al. conducted a meta-analysis of 17 psychological treatment trials for postpartum depression. Compared with control conditions, psychological treatments led to moderate improvement (effect size 0.61). Those trials which compared psychosocial interventions with wait-list control groups had higher effect sizes than those which were compared with treatment-as-usual (effect size 0.96 v 0.41). Only small differences between the effect sizes for different psychotherapy modalities were found, suggesting that common non-specific factors may mediate improvement for all psychosocial interventions for postpartum depression. The investigators were unable to draw conclusions about the long-term effects of psychotherapy owing to the lack of data beyond 6–12 months postpartum. This limitation was also noted by Sockol et al. .

Many of the studies in the meta-analyses have significant limitations. For example, only about one-half required women to meet diagnostic criteria for major or minor depression. Subject inclusion criteria in many were based solely on Edinburgh Postnatal Depression Scale (EPDS) scores above cut-offs ranging from 9 to 12. Many were also limited because they had small sample sizes and were underpowered.

Four studies stand out for their methodologic rigor, including large samples, use of DSM criteria for major depression for study entry, and longer-term follow up . O’Hara et al. randomised 120 women to either 12 sessions of IPT or to a waiting list control condition (WLC) over 12 weeks. Participants who had completed IPT ( n = 48) had significant reductions in depressive symptoms, higher rates of recovery, and higher levels of social adjustment relative to the WLC condition ( n = 51 completers). Women in the WLC condition were treated with IPT immediately after their waiting period, so no controlled follow up could be conducted. Ninety-seven women from both groups who ultimately received IPT, however, were followed for 18 months after treatment . Among women who had recovered, 43% experienced a recurrence during the following 18 months. Among those who did not recover with treatment, 84% subsequently recovered during the follow-up period.

Cooper et al. randomised 193 women to four conditions: non-directive counselling, cognitive counselling, psychodynamic therapy, or routine primary care. Routine primary care was described as ‘the normal care provided by the primary health care team (i.e. general practitioners and health visitors) with no additional input (apart from assessment) from the research team.’ No information was provided about medication usage. The effect of therapy was significant, but no differences were observed between the active treatment conditions. Follow-up assessments were conducted at 4.5 months, 13.5 months, and 5 years. No differences were found between groups at any of these follow-up assessments. Treatment relative to usual care was not effective in preventing relapse or recurrence.

Milgrom et al. randomised 192 women to CBT-based group therapy, group counselling, individual counselling, or routine primary care. Interventions lasted 12 weeks and included three partner sessions. Similar to Cooper et al. , the investigators found a significant effect for treatment but no differences among interventions.

Morrell et al. conducted a pragmatic cluster-randomised trial in primary care. The investigators randomised 101 general practices to usual care or to CBT-based counselling or listening visits. Health visitors were trained to identify PPD and deliver the counselling interventions to 595 women who scored over 11 on the EPDS at 6 weeks postpartum. A significant effect was found for treatment at 6 and 12 months, but no differences were found between the two active treatments.

Psychological treatments for perinatal depression: compliance and treatment delivery

Engagement and retention in treatment are particularly challenging for pregnant women with depression . Non-compliance with psychiatric interventions often lead to less favourable outcomes and greater drop out . In addition to multiple financial and logistical barriers to care, lack of trust in the provider and stigma are critical obstacles in the treatment of perinatal depression .

One approach for improving the rates of retention and adherence is to target high-risk populations that are presumably more motivated to engage and comply with treatment. Several studies have examined adaptations of CBT and IPT for depression in pregnant women with low incomes , those living in inner-city or rural areas , high risk Latinas , and pregnant adolescents . Several studies of CBT and IPT-based treatments have also been conducted for women who are victims of violence or have lost a child .

In brief, CBT is based on the concept that one’s thoughts affect mood and functioning. Negative thoughts about self (e.g. ‘I’m a bad mother’), others (e.g. ‘my partner doesn’t care for me’), and the future (e.g. ‘my depression will never get better’) are targets of the treatment. Behavioural change focuses on increasing pleasurable activities. Interpersonal psychotherapy is based on the concept that social support and interpersonal problems affect mood and functioning. Interpersonal problems such as disputes (often with partner), transitions (literally those such as having a baby or going back to work), and grief and loss issues (such as miscarriages or other perinatal losses) are addressed. Behavioural change focuses on increasing social support and improving interpersonal communication.

New delivery modalities have also been developed, such as telephone-based IPT for postpartum depression , telecare , and internet-based therapy . Several modifications of IPT that include partners have also been pilot tested . Preliminary results have shown positive effects on depression outcomes for all of these modifications.

Psychological treatments for perinatal depression: psychotherapy during pregnancy

Four psychotherapy trials for depression during pregnancy have been conducted. Spinelli et al. randomly assigned 38 pregnant women with major depression to either 16 weeks of IPT or a parenting education control group. Participants who received IPT improved significantly compared with controls on the EPDS, Beck Depression Inventory (BDI), and Hamilton Depression Rating Scale (HAM-D) ( P < 0.03), and had substantially higher rates of recovery as defined by Clinical Global Impression score 2 or less (60% v 15.4%), or HAM-D score less than 6 (19.0% v 0.0%). A limitation of the study was the relatively high drop-out rate (32% of controls and 16% for the IPT group).

In a small open pilot study , seven pregnant women completed 12 weeks of IPT. At intake, all met criteria for major depression. None of the women received medications. A decrease in BDI scores was reported from 28.1 at intake to 3.2 at study completion ( P < 0.001), and mean HAM-D scores decreased from 19.8 to 8.4 ( P < 0.001). Only one of the seven met DSM-IV criteria for major depression at the end of treatment.

Two open-pilot trials examined IPT adapted for pregnant adolescents in group format . Twelve weekly group sessions of IPT were conducted within public schools in New York City. Participants attended an average of 8.8 sessions (SD = 1.7). The 14 participants in the first study were predominantly Hispanic girls; the second included 11 African–American girls. At treatment termination, participants in both studies had significant decreases in the BDI and EPDS scores (50% reduction in study 1, and 40% reduction in study 2), with strong effect sizes when pre- and post-treatment scores were compared (0.8–0.9). Treatment gains were maintained to 20 weeks postpartum.

To date, no treatment trials of CBT for acute antepartum depression have been conducted. Several, however, have used IPT or CBT delivered during pregnancy to prevent postpartum depression.

Zlotnick et al. investigated whether an IPT-based group intervention in pregnant women reduced the incidence of postpartum depression. The study included 37 women at risk for postpartum depression because of low socioeconomic status. They were randomly assigned to a four-session group IPT-based intervention or treatment as usual. At 3 months postpartum, none of 17 the women who received the IPT-based intervention developed major depression, whereas 33% ( n = 18) of the control group did.

Grote et al. investigated the effect of ‘culturally enhanced’ IPT during pregnancy for the prevention of postpartum depression. Fifty-three African–American pregnant women of low socioeconomic status were identified as ‘at risk’ if they screened positive for depression (EPDS > 12). They were randomly assigned to enhanced IPT ( n = 25) compared with enhanced usual care ( n = 28). The IPT consisted of one engagement session, eight acute IPT sessions before birth, and several maintenance IPT sessions up to 6 months postpartum. More participants completed the full course of active treatment (seven to eight sessions) than were retained in the usual care group ( n = 17 v n = 2). At 3 months after baseline, the IPT group had much higher response rates (defined as 50% EPDS improvement) than the control group (80% v 20%, P < 0.001). At 6 months postpartum, no women in the IPT group met criteria for major depression compared with 70% of the usual care group.

Le et al. conducted a randomised-controlled trial to evaluate the efficacy of CBT in preventing postpartum depression in high-risk Latinas. A total of 217 pregnant women were randomised to treatment as usual or 8 weeks of group CBT followed by three booster postpartum sessions. Women who received CBT had significantly lower BDI scores during late pregnancy, but the effects were not maintained postpartum.

Austin et al. examined the effectiveness of group CBT in at-risk pregnant women. Participants were recruited from within primary care; 191 were randomised to group CBT or a control condition ( n = 86). Eighty-nine women completed the full course of 6-weekly 2-h sessions. Attrition rate was high (52%). Depressive symptoms improved modestly in both groups; however, intention-to-treat analyses did not reveal significant differences between groups.

Cho et al. conducted a pilot study to examine the efficacy of brief antenatal CBT for the prevention of postpartum depression in women at risk. Broad screening of 1000 women took place in obstetrics and gynaecology clinics in Korea; 27 pregnant women with BDI scores greater than 16 were randomised to CBT ( n = 15) or a control condition ( n = 12). Compared with controls, women who received CBT were much less likely to experience clinically significant depressive symptoms 1 month postpartum (BDI > 16; 18.3% v 30%).