Background

Medical research continually reinforces the health benefits of human milk for infants. What is novel, however, is a growing body of evidence that shows how important it is to breastfeed in order to maintain good physical and mental health of recent mothers [1, 2]. Much of the debate on interventions to promote successful breastfeeding has focused on milk expression methods, peer community-based support, professional education, maternity care practices, leave policies, workplace regulations, or social marketing [3,4,5]. Overall, public health interventions have been effective in increasing the proportion of breastfeeding initiation [6, 7]. However, more efficient resources are essential to achieve an improvement in breastfeeding exclusivity and maintenance [8] among mothers who need extra support.

In 1858, Marcé published the first paper devoted entirely to puerperal mental illness. He reported that 33% of women developed depression in the lactational period [9]. Current research confirms that up to 5.5% of women use antidepressant drugs during the perinatal period [10]. The proportion of pregnant or lactating women that are prescribed antidepressants varies by geographical location. Across Europe, there are marked differences in the prescription of antidepressants to women during and after pregnancy, from 3.3% in Italy, to 9.6% in Wales. On the other hand, prescription of antidepressants during pregnancy is lower in Europe than in the USA, with reported percentages between 5.6 and 10.2%. In all databases, antidepressant use was at its lowest during the second and third trimesters of pregnancy, but by six months post-pregnancy, the rate of antidepressant use had returned to pre-pregnancy levels [11]. It is worth emphasizing that systematic reviews fail to draw any clear conclusions about the effectiveness of antidepressants for the prevention of postnatal depression [12]. No less important is that antidepressant prescription independently predicts exclusive formula feeding shortly after birth [13].

It is widely accepted that perinatal mood disorders (PMD) affect an estimated 20% of breastfeeding mothers [14]. Additionally, there is strong evidence for an association between higher levels of maternal depressive symptomatology and shorter breastfeeding duration [15,16,17,18,19,20,21,22,23]. Even when cessation of breastfeeding is due to inherited metabolic disorders, mothers who are obliged to replace breastfeeding with special infant formulas experience the highest degrees of stress [24]. Another issue under consideration, less obvious but no less important, is to what extent breastfeeding prevents mood disorders. It has been shown that the rising levels of inflammation markers during the third trimester of gestation constitute a risk factor for PMD [25, 26]. Breastfeeding may intervene to counterbalance this situation via down-regulation of both stress and inflammatory response systems [27,28,29,30].

Notwithstanding the above, breastfeeding support through psychotherapy interventions has been scarcely explored. The aim of the present study is to critically assess the effectiveness of psychological programmes (i.e. psychotherapy, relaxation, and stress-releasing techniques) on the breastfeeding success of recent mothers. Early references unveiled a multifaceted relationship between stress and breastfeeding. Severe stress may result in a negative effect on the process of lactation, both behaviourally and biologically [31, 32]. At its best, psychological support might relaunch a virtuous cycle involving the prophylactic role of breastfeeding in reducing maternal psychological distress.

Method

Data sources

An extensive literature search was performed up to the week starting January 15th, 2020. The primary endpoint was the assessment of breastfeeding success in mothers participating in psychotherapeutic interventions aimed to provide support to participants in enduring the difficulties associated with childbirth. For the purpose of this review, psychological therapy interventions include a wide range of interventions that target cognition, motivation and behaviour. The approach was to explore quantitative and qualitative parameters. It should be noted that this is a narrative review rather than a systematic review. Quantitative parameters included breastfeeding initiation, duration, and composition. Qualitative parameters recorded the evaluation of maternal perceptions on breastfeeding success. Databases searched were: Cochrane Library, PubMed, EBSCOhost, and PsycINFO for papers published since 1980.

Table 1 outlines the search strategies and key terms used. During the search, keywords for breastfeeding and for psychological interventions were considered primary, and were either combined to each of the other keywords individually or used in various combinations.

Table 1 Search terms used to identify existing literature reporting psychological interventions to support breastfeeding, 15 January 2020
Full size table

The search yielded 239 articles, out of which 20 articles were considered relevant for inclusion in this review (Fig. 1). First, titles and abstracts of articles from databases were screened and identified for eligibility. Selected articles were evaluated independently by two reviewers. When supporting data were not available, recommendations were made based upon the combined opinions of more than two authors. We recognize that inconsistencies can certainly occur during the searching stage. We describe what was done with the literature once it was identified, in order to assess and bias: selected studies met the following four conditions: (1) to include trials that do not substantially overlap, the main outcome of searching more than one database is that differences in indexing across databases increase the chances of retrieving relevant papers; (2) to have matching scopes, even though literature review is a cascading process of summing up materials about a topic in order to fully match the scope and innovation in a specific field; (3) to have good methodological quality; and (4) were published from 1989 to 2019 [33].

Fig. 1
figure 1

PRISMA flow diagram 1c

Full size image

Although the hierarchy of evidence favours quantitative methods, mixed research approaches are becoming increasingly acceptable in assembling evidence. There is a place for rigorous qualitative research or case series that include in-depth analyses of few patients in their natural clinical setting.

Study selection

Studies were eligible if they collected data relating to breastfeeding behaviour, outcomes or biological sequelae (ie. breast milk composition), and focused on psychological interventions in the peripartum (beginning sometime during the final month of pregnancy through about 5 months postpartum) [34]. Given the high variability in breastfeeding definitions, and for the purpose of this review, breastfeeding is defined as the provision of breastmilk to an infant by any means. Studies of mothers with premature or very low birthweight infants were also deemed eligible.

Conversely, studies of infants with cleft palate, cerebral palsy, gastrointestinal disorders, food allergy or other specific medical problems known to affect feeding were excluded due to the high risk of confounding. Additional exclusion criteria for this review were protocols, reviews, commentaries, and letters to the Editor.

We divided all studies included in four different sections: (1) randomized controlled trials; (2) quasi-experimental trials; (3) pre-post design studies; and (4) case series that reported more than one case report of psychological interventions to support breastfeeding. We only included single-center study data.

Pre-registration

This study is pre-registered. A PDF document containing the AsPredicted Protocol ‘Perinatal psychological interventions to promote breastfeeding’ (#40903) is publicly available [35]. Question 1 of the AsPredicted document shows that some data had been already collected at the time of pre-registration. In this case, preregistration involves specifying the study design, and takes place before the data are analysed.

Results

Study characteristics

Studies included were published between 1989 and 2019. Samples were derived from 11 countries (Australia, Canada, China, Croatia, Denmark, India, Iran, Malaysia, Pakistan, United Kingdom, and the United States). Twenty studies were finally included in the present critical analysis and are summarized here. Each study is briefly commented on, highlighting aspects of study design and results. Out of the included studies, 11 were randomized controlled trials (RCT), four were quasi-experimental, three were case-series reports, and the remaining two studies used a pre-post design. The total number of mothers included was 3136. Interventions were analysed, regardless of whether these programs were initiated antenatally or postnatally. The studies were clustered in four groups composed of two to three comparable psychological interventions.

Nine studies of the 11 included RCTs, revealed an association between breastfeeding and psychological interventions on maternal mood. On the other hand, the results of a RCT on interpersonal psychotherapy [36] and another on self-hypnosis [37] did not suggest that mothers on therapy may be more likely to breastfeed. Out of four quasi-experimental studies, the relationship between breastfeeding and mothers’ relaxation interventions was demonstrated in two studies with a pre-test-post-test design, another did not report an association between psychological therapy and breastfeeding outcomes, and another could not be merited because measurements do not apply to it. The rest were case series.

In six studies, the control group received no extra attention other than the usual perinatal care. Five of those studies reported significant differences between control and intervention groups. Conversely, in nine studies, either mothers were randomly allocated to intervention or control sessions, or mothers in experimental and control groups received the same breastfeeding support and guidance during the trial. Both groups reported equal satisfaction levels with regards to health care offered to protect breastfeeding. Out of nine studies, eight reported significant differences between experimental and control groups.

Most papers deal with universal interventions targeting a population not at increased risk for perinatal mental health conditions; only two articles reviewed are devoted to selective psychological interventions for women perceived to be at risk of perinatal depression [36, 38].

Two main different psychotherapeutic approaches have been evaluated. On the one hand, approximately half of the papers have evaluated the effectiveness of psychotherapies that are performed by psychiatrists and psychologists [36, 39, 40], and to a lesser extent by nurse therapists, social workers, or specialised counsellors under close supervision [37, 38, 41,42,43] (i.e. cognitive-behavioural therapy, interpersonal psychotherapy, group therapy, or hypnosis). In the remaining studies, a reference is made to research projects whose staff has received brief training in applying methods that have already been tested, and are suitable for healthcare or prevention [44,45,46,47,48,49,50,51,52,53,54,55] (manual relaxation, meditation relaxation techniques, music therapy or compassion-focused therapy).

Results are similar for interventions delivered by specialised versus non-specialised professionals. Psychological interventions delivered by specialists achieved significant results in seven out of 10 cases, and also in six out of seven interventions by non-specialists.

Study descriptions have been tabulated (Table 2), along with method quality and risk of bias (Table 3).

Table 2 Studies’ description. Characteristics of women and infants for whom a psychological intervention to support breastfeeding has been tested. Results reported according to study design
Full size table
Table 3 Interventions’ description and study limitations
Full size table

Cognitive behaviour therapy programs

Cognitive behavioral therapy (CBT) refers to a group of interventions that combine cognitive and emotion-focused techniques, with the objective of replacing unreal and negative beliefs with more precise and positive thoughts. This type of therapy was used in a study in rural Pakistan, where various sessions of CBT were carried out by trained health workersThis intervention doubled the rate of exclusive breastfeeding at six months of life [42]. Similarly, a project in India reports that a single session of CBT was effective in significantly improving how well the infant latches onto the breast in the immediate neonatal period [41].

Interpersonal psychotherapy

Interpersonal psychotherapy (IPT) is based on attachment theory [56, 57]. Our review identified three studies investigating the effects of IPT on breastfeeding outcomes. One study from the New York State Psychiatric Institute, examined the comparative effectiveness of IPT and a parenting education program (PEP), in promoting breastfeeding duration among pregnant women who met DSM-IV criteria for a major depressive episode [36]. Irrespective of the fact that IPT or PEP was administered, 76% of the women were breastfeeding at the fourth postpartum week. These researchers point out that their breastfeeding rates post-intervention exceed consistent findings in the literature. The other two studies assessed the effectiveness of IPT group intervention, with regards to breastfeeding success. One of them, the ROSE program consisted of four group sessions for 3–5 pregnant women at high risk of perinatal depression [38]. Though women in the ROSE program had similar breastfeeding initiation rates, as compared to women on standard care, they maintained breastfeeding for longer. Finally, in a study on women visiting a maternity clinic in Tehran, women in the intervention group took part in three supportive psychotherapy sessions and, thereafter, the rate of breastfeeding increased significantly.

Hypnosis

Hypnosis has previously been found to provide pregnant women with skills to successfully manage labour [58,59,60]. A Danish RCT compared the breastfeeding outcomes of three groups of nulliparous healthy women: a self-hypnosis group, a mindfulness relaxation group, and a usual-care group [37]. Concerning self-hypnosis, there was no difference in establishing breastfeeding or breastfeeding duration between the intervention and the two control groups.

Relaxation techniques

Various relaxation methods share a long history of clinical practice and research. Such therapies have the potential to both add quality to breastfeeding matters, and to alleviate stress during the postpartum period [61, 62]. We analysed 13 studies that examined the effect of these interventions on breastfeeding practice. One study was devoted to compassion-focused relaxation therapy [55], and another isolated study combined pharmacotherapy with active mental imaging relaxation techniques [49]. Three papers explore manual relaxation as an integrative psychotherapy approach [44,45,46]. Four studies address the influence of various environmental sensory stimulations on milk production [43, 50, 51, 53]. Finally, auditory-mediated exercises that guide the body to relax and control breathing, thus alleviating stress and promoting breastfeeding were analysed in five research papers [40, 47, 48, 52, 54].

Compassion-focused therapy is a system of psychotherapy designed to help individuals being kind to one-self and being mindful, in order to reduce high levels of self-criticism [63]. An Australian study sought to evaluate the potential utility of a package of online resources designed to improve self-compassion for mothers within 24 months postpartum [55]. Mean total scores for the Maternal Breastfeeding Evaluation Scale [64] indicated improvement from pre- to post-intervention.

A case series illustrates that women with tetraplegia can sustain breastfeeding for extended periods [49]. Active mental imaging and relaxation efforts on the part of the mother were required to facilitate let-down in two cases. Later on, one of these cases needed a nasal spray of oxytocin after receiving oxycodone for pain relief. The third tetraplegic mother was successful in her attempts to breastfeed without additional relaxation or drug therapy.

Three different manual relaxation papers have been evaluated [44,45,46]. A case series [46], and a clinical trial using a pretest–posttest design has examined reflexology [44]. Reflexology is a treatment in which pressing some areas (hands, feet or ears) can cause deep relaxation and result in the secretion of hormones. In a case series of three mothers after an obstructed labour, reflexology three times a week was shown to increase milk secretion. Similarly, a trial showed that the increase in milk volume from day 1 to day 5 after the reflexology intervention was significantly higher in the intervention group, compared to the control group. When it comes to back massage, a quasi-experimental study was conducted to assess the effectiveness of this relaxation technique in post-feeds weight gain, among primiparous mothers of full-term neonates [45]. The intervention started immediately after birth. It was found that the mean frequency of micturition and post-feed weight gain was higher among infants in the study group.

Four studies investigated the association between specific environmental sensory stimulations and breastfeeding success [43, 50, 51, 53]. In one study, participants entered the Snoezelen room with an unsettled baby and breastfeeding issues aggravated by maternal tiredness [43]. The main features of the room are the wheel projection that slowly rotates to display patterns on the wall, a tropical fish tank, music and aromatherapy. Eleven women shared their experience using the Snoezelen room during the early stages of their breastfeeding, 80% of them achieved exclusive breastfeeding and all of them indicated they would recommend its use to other breastfeeding mothers.

Two studies on music therapy involved mothers of preterm infants. In one of them [50], the control group received only the standard support, whilst the other three experimental groups received mp3 players with a recording of a spoken progressive muscle relaxation protocol, accompanied or not by selections of lullabies for guitar or by a series of images of the mother’s infant. In particular, the intervention that included a slideshow of images of the mother’s child improved breast milk quantity more than the others. The other study was carried out on mothers of preterm babies when they went to the NICU to express breast milk [51]. Mothers’ salivary cortisol level showed a significant decrease, and mean volume of expressed breast milk showed a significant increase after music therapy. Another study recruited 60 women in-hospital who had just delivered their first child, and had the intention to breastfeed. The independent variable was music therapy paired with relaxation techniques. Mothers who received a longer music therapy session ended up breastfeeding for a longer amount of time and their perceptions of their experience were more positive [53].

Finally, five studies examined the effect of auditory-mediated mind guidance on breastfeeding practice [40, 47, 48, 52, 54]. The first results on promoting breastfeeding through these techniques dates back 30 years. Mothers in the intervention group were instructed to listen every day to a 20-min audio that consisted of a progressive relaxation exercise followed by a guided imagery section. Approximately one week after enrolment, the mean volume of expressed milk in the intervention group was 63% greater than the mean volume in the no intervention group [54]. Another study sought to clarify the effect of stress on breast milk humoral immunomodulation. A convenience sample of healthy breastfeeding women was recruited within 48 h of delivery. The women were assigned to one of the following three study groups: relaxation intervention (group 1), attention control (group 2), and no intervention (group 3). Milk samples were tested for sIgA five times, until six weeks postpartum. The sIgA levels in breast milk samples from women in each of the intervention groups were not significantly different. However, this paper presents the intriguing results that sIgA levels in group 1 varied with relaxation success: the women who rated very good and poor at relaxation, had levels of 0.30 and 0.67 g/L, respectively. This is opposite to the concept that cortisol may suppress the immunoglobulin production by plasma cells in the breast [48]. The unique research on autogenic relaxation was a randomized prospective study that analysed data of breastfed infants from the second month of age until the age of six months [41]. In the experimental group, 47 out of 50 mothers (94%) fully breastfed their child during the first six months of life, while those in the control group were 17 (34%) [41]. In 2019, two groups of researchers unveiled details of how relaxation techniques affect breastfeeding outcomes. A Chinese within-subject study on 20 primiparous mothers who were breastfeeding compared basic physiological parameters from five different approaches to relaxation: meditation tape (RM), music tape (M), relaxation lighting (L), combined RM + L, combined M + L, and control session without intervention [47]. Heart rate (HR), blood pressure (BP), fingertip temperature, and the let-down breastfeeding reflex were assessed before and after each session. RM resulted in the greatest change in BP and HR, and four participants experienced milk let-down (ejection reflex) during the session. This was not found for the remaining four treatment or control groups.

The last study recruited healthy first-time mothers during their third trimester of gestation [52]. At two weeks postpartum, mothers and exclusively breastfed full-term infants were randomly assigned to receive relaxation therapy, or to the control group. Mothers in the relaxation group received a modified audio-guided imagery protocol designed for breastfeeding mothers to listen to while breastfeeding, during each home visit. They were also encouraged to listen to it as frequently as they found useful, throughout the 12 weeks trial. Analysis with a linear model showed that infants from the relaxation group had a mean milk intake at 12 weeks that was 226.5 g/d higher than the intake of infants in the control group.

Discussion

Summary of evidence

The purpose of this qualitative review is to examine the relationship between perinatal psychological interventions and breastfeeding outcomes. Out of the 20 studies included, three studies reported no relationship between these variables: relaxation techniques do not increase milk sIgA levels, and a brief course of self-hypnosis, or individual sessions of interpersonal psychotherapy have no effect on breastfeeding duration. Five studies with small sample sizes (n < 30) may have lacked sufficient power. Importantly, all studies on mothers of preterm infants provide strong evidence in favour of the potential that simple relaxation interventions have for improving the success of milk expression. These cheap techniques may help mothers who are especially stressed during the postpartum period and at a high risk of breastfeeding failure.

According to the levels of evidence developed by the Centre for Evidence Based Medicine for treatment [65], all psychotherapy-based interventions represent level-I or level-II evidence (large RCT trials with clear cut results or small RCT with unclear results, respectively). Based on the quality of evidence to support or oppose these outcomes, psychotherapy appears to be beneficial for these mother-infant dyads, but patient preference should have a substantial role in choosing an appropriate practice.

As for research on relaxation techniques, five studies represented level-I evidence, five level-II evidence, and three level-IV evidence. Based on the quality of evidence and consistent findings, clinicians should recommend relaxation techniques in support of breastfeeding but should remain alert to new information.

Discussion of findings

Research on psychological interventions intended to support breastfeeding practices has largely been neglected over the years. As far as we know, only three very recent reviews have focused on the effects of behaviour-change or stress-releasing techniques on breastfeeding success, or breast milk composition [66,67,68]. A 2018 critical analysis has assessed the effectiveness of body and mind stress-releasing techniques on the breastfeeding success of mothers of healthy neonates [66]; nine studies were included in this analysis. Researchers conclude that cognitive-behavioural counselling seems effective in improving breastfeeding initiation and duration, that manual relaxation techniques seem effective in promoting breastfeeding initiation, and that mind guidance seems also effective in promoting breastfeeding duration. A 2018 systematic review [67] has identified four studies investigating the effects of relaxation therapy on breast milk yield and composition. Among them, two RCTs [50, 54] found that relaxation therapy significantly increased milk yield. In terms of milk composition, one RCT [50] reported a significant increase in fat content in the breast milk of mothers in the intervention group. A 2019 meta-analysis [68] aimed to investigate the effectiveness of socio-psychological interventions in promoting breastfeeding initiation, duration or exclusivity. Reviewers found 20 studies that employed behavioural change techniques, with a majority of interventions falling into the category of Education and Support. After adjusting for publication bias, socio-psychological interventions reviewed did not increase rates of any breastfeeding across the postpartum period.

To the best of our knowledge, our review is the first one to comprise all psychological interventions intended to assist mothers in need of extra support to breastfeed. Psychotherapy and relaxation techniques are necessary to deal with some cases that are not adequately addressed by peer-support programmes or public health initiatives, implemented to create environments to support breastfeeding. Our findings are in line with the effectiveness of previously reported interventions offering consistent psychological support in the peripartum period, despite the inherent differences in psychological mechanisms of action or manner of psychological well-being promotion.

Given the proven interaction between the reproductive and stress systems [69, 70], there is no doubt of the impact of stress-releasing techniques on lactation. However, most interventions included in this review did not theorise specifically on how the interventions were intended to work. This is all part of the majority of successful breastfeeding interventions whose mechanism of action remains unknown [6, 8, 71].

Limitations

Systematic reviews have increasingly replaced traditional narrative reviews to update the best evidence for the most basic and clinical questions. The major advantage of systematic reviews is that they are based on the findings of comprehensive literature searches in all available resources, avoiding subjective selection bias, while narrative reviews can provide the authors’ experiential perspectives in focused topics.

This narrative review is not focused on a specific question. It is conceived as a descriptive overview of a topic. As concepts can be expressed in different ways, combinations of keywords and subject headings have been used to look at all the evidence. If a study was left out, it was by mere chance and not because of the study’s results. This kind of search strategy has a higher sensitivity, it is used to minimize bias, but often lowers the precision (relevance or accuracy) of the results. Additionally, there are several search tactics for which there is no consensus. Variation in practice around such issues as limits, searching for observational studies, for outcomes and comparators, persists over an extended period of time [72, 73].

While the small number of studies included in this review reflects the current status of this novel approach to breastfeeding support, it also reduces the strength of the respective recommendations, due to the difficulty in obtaining pooled data. Some studies had a small sample size, which may not provide sufficient power to detect differences in outcomes. Despite employing broad eligibility criteria to include interventions, inclusion of published studies may have introduced some bias towards studies with positive findings. The lack of a clear breastfeeding definition decreased the comparability of research results. Current definitions for breastfeeding have been questioned. Clinicians and statisticians need accurate definitions for breastfeeding. Many studies provided varying cut-offs and most studies evaluated infant-feeding methods at various times throughout the postpartum period; heterogeneity in measurements limits their comparability. The majority of the reviewed studies do not distinguish between exclusive and partially breastfed infants. We elected to accept all definitions of “exclusive breastfeeding” as provided by the different study authors. Another limitation was that several studies incorporated self-report measures, hence there may have been underreporting of the use of nonhuman milk because previous research suggests that preventive infant-feeding practices tend to be overestimated by self-report measures. In addition, with lengthy time-periods involved, retrospective studies may have some degree of recall bias.

This review acknowledges that there is a particular instance where mixed research methods may contribute in assembling evidence from quantitative studies and in-depth analyses of few patients.

Although the same subjects were not expected to be included, multi-centre studies have been excluded because a cluster effect might occur; in exchange, there is a risk of losing larger studies.

Finally, generalizability of results may have also been limited because only half of studies controlled for potential confounding variables such as maternal age, education or family income in their multivariate analyses.

Conclusions

The studies included in this review show that psychotherapy as well as stress-releasing interventions, seem to support breastfeeding success. Conversely, this review suggests that self-hypnosis does not lead to improvements in breastfeeding outcomes. Concerning population groups, the strongest evidence was for an effect in increasing milk volume expressed by mothers of preterm infants. Future research is needed to evaluate the impact of psychological interventions on breastfeeding duration, to enable high-quality evidence to be implemented into practice.