Cognitive-Behavioural therapy and interpersonal psychotherapy for the treatment of post-natal depression: a narrative review

Post-natal depression: Definition, clinical features, risk factors, and effects

Post-partum period has been associated with mood disturbances since Hippocrates’ times [1]. It is reported that PND started being officially used in psychiatric manuals in 1994 [2]. It was “officially objectified” in the 1950s [3]. PND was given some “diagnostic criteria” at that time as practitioners began to notice that some women, after giving birth, would experience a psychological pattern with depressive characteristics.

In the earlier days, science was trying to understand PND in the context of causes. These varied from the socio-economic background of the individual to unplanned pregnancy [4]. In more recent years, clinicians have gained a better understanding of PND. They give emphasis to the combination of risk factors rather than single causes [5]. Risk factors can vary from low socio-economic background, family history of depression or personal history of mental health issues, low social support, smoking habits, sexuality issues, and immigration issues [6].

The Diagnostic and Statistical Manual of Mental Disorders - fourth edition (DSM-IV) [7] initially linked PND with major depression with post-partum onset within 4 weeks of birth. However, a study for the treatment of PND [8] highlights the variations amongst studies which define the onset of PND, from the first month up to the end of the first year following the infant’s birth.

The Diagnostic and Statistical Manual of Mental Disorders, DSM-5 [9] places PND in the category of unspecified depressive disorders where the main symptoms can cause significant distress or impairment on various levels in the individual’s life. PND includes different sub-categories called specifiers, amongst which are the peri-partum onset. This refers to the onset of depression during pregnancy or postpartum for the time following the birth of the child. According to the same manual, a large number of postpartum major depressive episodes begin during pregnancy, thus they are also called peri-partum episodes. These episodes range from mild to severe, with or without psychotic features. The individual may also experience hallucinations or delusions.

A study by Hewitt et al. [10] describes depression as a very serious mental health problem with important consequences on a societal level. In this study they estimate that depression will become the second-highest health problem by 2020. According to the same authors, PND is considered to be a very important category of depression with often serious consequences. It can affect both the mother and the infant, as well as the immediate and/or the extended family. PND can have long-lasting effects on the development of the infant on a cognitive and emotional level, including attachment issues amongst others [11].

PND can cause significant distress or impairment on various levels in the individual’s life, e.g. lack of motivation, affected mood, sleep and appetite issues, lack of concentration, rumination, unintentional or intentional suicidal ideation, or psychotic phenomena such as hallucinations or delusions.

PND affects 13% of women [12], and between 4 and 25% in men in the first 2 months after the baby is born [13]. Other research suggests that the proportion of mothers who suffer from PND is one in seven [14]. According to the same article, the mentality of organisations and health providers is changing, especially in the United States where there is a shift towards more systematic screening of mothers-to-be or young mothers who might experience symptoms of depression.

Psychological treatments of PND

Regarding treatment for PND, Rudlin lists its main therapeutic approaches [15]. They vary from medication, home visits, education, phone contact, one-to-one counselling, group therapy, and self-help resources such as books. CBT, together with interpersonal psychotherapy (IPT) are considered two efficacious non-pharmacological treatments for PND [16].

A meta-analysis [17] examined how effective psychological treatments are for PND in primary care. It was found that psychological interventions such as CBT and IPT, along with counselling, psychodynamic therapy and support groups can be very effective in reducing the symptoms of depression up to 6 months post-intervention.

CBT’s main focus is identifying distorted negative thinking patterns. It emphasises the link between thoughts, feelings and behaviour. Dalby [18] highlights Albert Ellis’s theory of irrational thinking patterns and how they could trigger emotional disturbance to the individual.

A common characteristic of people who suffer from depression is their tendency to experience automatic thoughts, usually of negative content. Beck [19] highlights that the negative automatic thoughts usually carry negative meaning in relation to the notion of the past or the future, about the individual themselves, and/or the world around the person.

CBT helps the individual to understand that identifying their own distorted negative thinking patterns allows them an opportunity to change them. By changing their thinking, the individual can change how they view and feel about themselves, and ultimately, change their behaviour [20]. CBT integrates many approaches in clinical practice such as problem solving, modelling, and cognitive restructuring, amongst others [21].

IPT focuses on four areas in the person’s life: grief about someone’s own sense of self or changes within their relationships, changes in roles, unresolved disagreements in interpersonal relationships, and a lack of life events. It focuses on strengthening the relationships of the individual, on increasing social support, and improves communication [22].

Conversely, Barlow [23] refers to the negative effects, or no effects of psychological interventions for various disorders, including trauma and addictions. Dimidjian and Hollon [24] talk about the adverse effect of psychotherapy, but this is yet to be researched adequately. There is no real agreement in the scientific community about ways to investigate and identify harmful psychological interventions. A review [25] published by the World Health Organisation (WHO) for psychological interventions on depression discusses the under-investigated but very real possibility of a negative effect of psychological therapies on depression. Some of those effects include the symptomatology of the individual becoming exacerbated or the individual experiencing a “relapse”.

Lambert [26] makes the point that psychological treatments for depression and other disorders have shown to be effective overall. Positive outcomes depend on the patients’ characteristics, but also the therapists’ “actions” or “inactions”. Lambert also highlights the reality of negative clinical outcomes for patients who experience depression. Lambert does, however, identify ways to minimise clinical negative impact and maximise positive outcomes which can be achieved through “measuring, monitoring, and tracking client treatment response with standardised scales”.

A meta-analysis by Cuijpers et al. [27], which we did not include in our initial search, investigates the effect of psychological treatments for PND. It was found that CBT, IPT, counselling and social support have an overall positive effect on PND, but they were less effective than what they have been on other psychological disorders. There was no real difference in therapeutic outcome between different psychological therapies. The same study also concluded that medication and electroconvulsive therapy can have higher effect size for PND than psychological treatments but that needs to be further investigated. In addition, it was found that the initial positive therapeutic effect of psychological treatments on PND could not be confirmed 6months or longer post-intervention. However, the authors of this meta-analysis highlight that some of these findings need to be interpreted with caution due to the small number of studies included, and that the quality of the studies was not the highest.

Virtual reality: Definition and its advantages

A question worthy of investigating in relation to the treatments of PND is whether they can be improved for better clinical outcomes using other treatments, such as virtual reality (VR).

VR is defined as “a way for humans to visualise, manipulate, and interact with computers, and it can be viewed as an advanced form of human-computer interface that allows the user to interact immersed in more intuitive and naturalistic fashion” [28].

VR promotes a sense of presence for the user in an environment which is computer based. According to Turner and Casey [29], VR can enhance the therapeutic effectiveness of psychological interventions. It expands beyond the strict boundaries of technology. VR is seen as a form of communication. It comprises elements such as an experience, visualisation and interaction [30].

Some of VR’s advantages in research and practice is that it can act as a powerful and effective tool which can complement traditional therapies such as CBT [31]. It is a form of therapy which enhances sense of control and raises self-efficacy. It uses technological means to help the individual. It is 3-dimensional and interactive. The VR user has the ability to “explore and engage” within the virtual environment.

VR can be delivered in a safe and controlled way [28]. It can be affordable, easy to access, and the therapist themselves can have control of how, what and when it is to be applied, which creates a sense of safety for the user [32]. One of VR’s main features and characteristics is that it can empower the individual, a very basic and essential ingredient in order for therapy to occur. The combination of CBT with VR can have a tremendous positive impact [29].

VR or virtual reality exposure therapy (VRET) has been used to explore a large number of topics, from stress, anxiety, phobias, acute pain, body image disturbances, eating disorders, training of children in spatial and navigation learning skills, functional skills [28], post-traumatic stress disorder (PTSD) [33–35], fear of heights and also fear of flying [36, 37].

Improvement of traditional treatments for PND

There appears to be a gap in the literature relating to the combination of VR with traditional therapies for the treatment of PND. A brief literature review so far identified only one study on the efficacy of VR on depression. A pilot study conducted by Falconer et al. investigated the concept of compassion and self-criticism in a virtual environment [38]. They examined whether compassion could be taught to subjects in a virtual environment. The study revealed that its participants, all adults with depression, were able to practice compassion both as a life-sized avatar and as a child avatar interacting with one another through the process of embodiment. Although this particular study was limited, being a small group (n = 15), and having no control group, its results were promising. Results indicated that most of the study’s participants had become more compassionate and less self-critical one-month post-intervention.

Our research group is exploring the utility of using VR in the treatment of PND, the final aim being to improve the treatment that mothers with PND receive. This could potentially be beneficial for the health and well-being of mothers, their families, and society in general. The investigation of the combination of CBT and VR could possibly provide a better treatment for PND from a clinical point of view, which could save on resources including time and money spent at an organisational level.

In order to explore the possibility of combining psychological therapies with VR for the treatment of PND, we decided to review past and current published literature on traditional therapies for PND. We wanted to find out what therapies work best, and under what circumstances. Our investigation researched other reviews, within the same clinical subject area. However, our review differs from other reviews, whether systematic or narrative reviews, in three methodological aspects. It focuses purely on the treatment of PND, rather than prevention, or prevention and treatment of PND. A second difference is that this review investigates mainly CBT as treatment for PND. A third advantage of this review is around population characteristics. We focused our research mainly on the post-partum clinical population and not on other types such as the ante-natal population. However, there was one exception where the clinical population was in the last trimester of their pregnancy in the beginning of the study, but it became post-partum at a later stage. We believe this three research characteristics help this research project make a clear contribution to the literature.

Objectives

We formulated the design of this review based on the working hypothesis that CBT is a successful treatment for various psychological disorders, amongst them PND. It is a therapeutic approach which is scientific based. It can follow a clinical protocol, where its clinical methods can be replicated. Its clinical efficacy can be tested and measured. We hypothesised that CBT is the most widely used and efficacious treatment for depression and PND.

We searched for specific parameters which we believe contribute to the efficacy of CBT. We wanted to pay particular attention to the types of participants, especially the ones who had been diagnosed with PND through a structured clinical interview. Another parameter was around the types of interventions or treatment components of CBT for PND. We searched for specific aspects of the CBT approach, in particular, cognitive restructuring, goal setting, and problem-solving.

In order to start this line of research and to design our PND intervention protocol, supported with VR, our first aim is to review the scientific literature relating to the most effective CBT treatments for PND. Then to identify the parameters that make those treatments effective. It will also investigate whether VR has previously been used as a treatment for PND.

Quality of studies

The quality of the studies varied from not strong to very good, with most to be considered fair.

This was based on the randomisation process, sample size, heterogeneity of methods, use of instruments, treatment protocol, generalisability and statistical significance of results, follow ups, and limitations of each study. Most studies reported bias except five studies for which it was unclear [16, 49, 55, 57, 58].

Treatment focus

Fourteen studies focused solely on the treatment of PND, one on the treatment of ante-natal depression and PND [56], and one on prevention and treatment of PND [50].

Almost all studies, except one [16], focused mostly on depressive symptomatology of the mother as a primary outcome measure.

Population studied

In the 15 studies the population, which in total were 2758, were either diagnosed with depression or had identified themselves as depressed. More specifically in six of the 15 studies the population were post-partum women who had been diagnosed through a clinical interview based on the DSM-IV [16, 49, 57, 59–61]. In one of them, the population were 16 years and older [56]. In the remaining 9 studies the participants would mostly identify with depressive symptomatology, mostly through interview based questionnaires such as HAM-D, or self-report questionnaires such as EPDS. In 2 out of the 9 studies the populations were “newly delivered mothers” [62, 63], and in one study, they were newly delivered mothers with low income [64].

Control groups

The control groups were made up of participants who would usually receive typical primary care, or they were on a waiting list. However, two out of the fifteen studies in the review did not have a control group [49, 65]. For example, the study by Appleby et al. [49] included four study groups which all received some kind of intervention. The study by Milgrom et al. [65] included three groups which all had some type of intervention.

In addition, it was not clear what the control conditions were for two other studies [58, 62]. For example, in the study by Honey et al. [62] the control conditions were routine primary care administered by health visitors, and in the study by Bennet et al. [58] the control conditions were standard primary care with a health visitor. However it was not clear whether, in either studies, the routine primary care involved GP visits, medication, both, or none.

In the remaining 11 studies, the control conditions were as follows: waiting list [16], health visitors contacting participants, and defining their postnatal care with the use of questionnaires and referring them to their general practitioners [55], enhanced routine care with regular weekly visits in the last month before birth, 1 month post birth and monthly visits for the next 9 months by routinely trained health workers who received regular supervision but they were not specialised in CBT [56], routine care in the form of clinic visits [50], home visits which focused on “child health and development, nurturing mother-child relationship, maternal health and self-sufficiency”, along with receiving treatment in the community [59], routine primary care which “would be typically provided by the primary health care team such as the general practitioner and health visitors with no additional input from the research team” [60], antidepressant medication received by control group subjects in a hospital outpatient program [61], “standard care with 6 weekly clinic visits lasting 20 to 60 minutes” [57], health nurses who would manage case by case the participants and refer them to other services where appropriate [63], antidepressant medication, brief psychotherapeutic interventions, GP consult, or referral to specialty services [64], and a single session with a midwife or obstetrician focusing on debrief [66].

Delivery of the interventions

In terms of the delivery of the interventions, 12 were individual-based and home visits, 2 were group-based [58, 64], and 1 study was carried out on an individual basis and was also group based [63]. Most interventions were delivered in the homes of the participants. One study was delivered at home and in a public hospital. One study provides no data relating to delivery of the intervention.

Location of the studies

Five studies took place in the United Kingdom, three studies in Australia, two studies in the United States, one study in France, one study in Canada, one in Pakistan, one study in Chile, and one study in Sweden.

Number of sessions and content of the interventions

The interventions in all of the studies varied in terms of the number of clinical hours and number of sessions. The majority of studies provided the number of sessions and number of clinical hours. However, 4 of the 15 studies [49, 56, 57, 60] provided only the number of sessions and not the number of clinical hours. The total number of clinical hours was approximately 168.5, with 1648 people having been provided with at least one of those interventions. The average number of clinical hours for each participant was 9.78. The average treatment period was 12.1 weeks.

Intervention providers

The intervention providers varied from nurses, psychologists, GPs, health visitors, and midwives. Almost all, except two studies [16, 50], included follow-ups varying from 1 week post-partum to 5 years following the birth of the child. The average period of follow-ups was 6.14 months.

Clinical trials using VR for the treatment of PND

We found no clinical trials that used VR for the treatment of PND. However, we found three studies in total [40, 67, 68] which used some form of technology. Two of them [67, 68] used video recordings, mostly for supervision purposes. The third study [40] was internet based for the delivery of CBT.

What CBT psychological treatments are effective for PND?

All fifteen studies included in our review used CBT as the main treatment for PND. However, there were 9 studies which compared CBT to other treatments such as non-directive counselling, psychodynamic and primary care, and 3 studies where CBT was combined with non-directive counselling, psychodynamic and primary care.

It appears that CBT can be viewed as a large clinical territory with many different techniques for the treatment of the same mental health issue. For example, in one study, CBT emphasised psychoeducation [64] while in other studies CBT focused on challenging negative thoughts and dysfunctional beliefs [55, 57]. In another study CBT was part of a wider community based program [64]. In the study by Rojas, the “purity” of the CBT approach was questionable [64].

We measured 7 components of CBT that were used for the treatment of PND, which were psychoeducation, cognitive restructuring, problem-solving, behaviour management, goal setting and goal achieving, stress management, and relaxation (Table 2).

The two most frequent used CBT interventions were found to be psychoeducation and challenging negative thoughts and beliefs with 9 studies in total having employed both at different times.

The second most frequent CBT intervention that was used was problem solving, while the third most frequently used CBT interventions were goal setting, behaviour management, and stress management.

What are the parameters that make those treatments having a successful clinical outcome?

This review shows the general outcome is that CBT as a therapeutic intervention is effective for the treatment of PND. It has an advantage over primary care for reducing depressive symptomatology in the post-partum period. However, the severity of PND varied in the included studies in our review, with most of the participants to be in the mild to moderate range. Thus it is unknown whether the same positive clinical results of CBT would be obtained for the more severe range of depression.

Another issue that was identified in relation to the effectiveness of CBT was the significance of the results. In five out of fifteen studies [60, 62, 63, 65, 66] it was found that their results on CBT’s effectiveness were positive overall but not statistically significant.

It is important to mention that although our review focused on CBT there were studies that compared CBT with other psychotherapeutic approaches that also showed effectiveness. It was found that non-directive counselling and psychodynamic approaches also had a positive effect on the reduction of symptoms of PND. To be more precise in the study conducted by Cooper et al. [60] which examined CBT, non-directive counselling and psychodynamic therapy, and compared them with typical primary care, it was found that all three interventions were effective. CBT was more effective on cognitive focus, behavioural tasks, and organisation. Nonetheless, the psychodynamic approach was more effective on relationships. It was also superior to the other two interventions in terms of depressive symptomatology according to structured interviews, especially up until the fifth month post-intervention. However, past that point, the initial therapeutic effect of all three interventions had started diminishing. From 9 months onwards up until 5 years post-intervention, the initial therapeutic effect was virtually non-existent.

Other studies have found similar results, whether the outcome is measured by self-reported measures or by a mental health professional conducting a clinical interview. CBT can be an effective treatment for PND in the short-term but its clinical effect long-term is questionable [65]. The same study measured the combination of CBT with counselling delivered either by a psychologist or a nurse. They found that the two approaches and a third one which was GP management, mainly through medication, had a good overall effect in the treatment of PND. We can conclude that CBT is an effective approach that can be delivered by various mental health professionals of different backgrounds, or even delivered by non-experts, such as lay people, or health visitors [56, 58].

It can also be concluded that CBT can be delivered in a flexible manner, whether through a home visit, or in a public place such as a hospital or clinic [65, 66]. There does, however, appear to be a preference toward home visits as it is believed to be more convenient and accessible to post-partum women with depression.

In addition to the treatment type and location CBT can be delivered in a brief manner which can have good therapeutic outcomes. According to Wiklund et al. [66] there is some evidence that brief CBT can benefit PND in the mild to moderate range. However it is worth noting that the population of that study were women who had not been diagnosed with depression but who only experienced signs of depression.

From the studies of this review we can conclude that CBT can also be delivered on an individual level or in a group [58, 63]. It seems that the individual delivery of the intervention is preferential and maybe the potential of the group intervention of CBT has not been adequately investigated. We can safely assume though that group intervention might have some advantages over individual treatment such as utilising less resources. At the same time it might be a disadvantage in terms of participants who might be more reluctant to participate initially in a group setting, where its participants share the same psychological experiences. This possibly underlines stigma on a societal level but this needs to be further investigated.

The studies of this review also show that CBT’s positive therapeutic outcome on PND is not impacted by the socioeconomic status of the population. Post-partum women with depression can benefit equally whether they are from a lower-middle income country or high income country. This review includes studies from different countries and continents and are categorised differently according to the GDP per capita, e.g. Australia vs. Chile [64, 65].

Another conclusion of this review is around the outcome measures. Most studies in this review measured individual depressive symptomatology in the post-partum period. However, one study [16] focused on the mother’s depression and measured it as a primary outcome. The authors included additional information, e.g. mother-infant relationship. It is worth mentioning here that the relationship between mother-infant was not in relation to the newborn baby, but rather on the already existing children in the family.

An issue that was raised was in relation to the target disorder and the self-reported measures which were used by the participants of some of the studies. The self-reported measures indicate depressive methodology but they do not necessarily ensure a formal diagnosis of depression. There was a lack of clinical diagnostic protocols in some of the studies.

A difficulty that one of the studies highlights was in relation to CBT. CBT was considered to be time consuming and highly demanding in terms of resource intervention [55]. Treating PND with populations of low socioeconomic characteristics, or in countries of low to middle income would prove to be a challenging task. Providing treatment to depressed mothers overall has been proven to be a challenging task on its own, especially in relation to the delivery of intervention. It becomes clear that home visits are a preferable way of delivery over GP practice or a public hospital for varying reasons. A young mother, or a mother with a high-demanding household might find it difficult to transport herself outside her home for a number of reasons, including a lack of transportation, money or time.

Another issue that was highlighted in our research was around the stigma of mental health issues. One study highlights the difficulties a clinical population, or mental health professionals might encounter in rural areas or low-income countries [56]. One way that the authors were able to deal with this issue was to support the idea that the CBT intervention was part of a larger community health program. Another way was that they promoted the idea of the infant’s health and well-being as a priority.

Some studies had a high attrition rate up to 30% but not all of them. It was reported that some participants did not complete the treatment. There was also a variation in terms of their length, and number of phases in the follow up process. All studies except two [16, 50] had follow ups with variations in frequency the follow ups were conducted for the majority of the studies and the span of the time period over which this happened, e.g. 2 months versus 5 years.

Almost all studies included different criteria in relation to population, e.g. primiparous vs. pregnant women vs. post-partum women, with different socioeconomic backgrounds and from different countries. We also did not have enough or adequate demographic information for all the included studies, which would allow us further analysis and conclusion reaching in our review process.

Has VR been used in the treatment of PND?

From our investigation, we found no clinical trials that had used VR as a form of treatment for PND. However, we found three studies which used technological means in clinical practice. One study investigated the efficacy of Toddler-Parent Psychotherapy (TPP), for the improvement of parents’ and infants’ mental health [67]. The technological means that were employed during this study were videotapes which had recorded the interaction between the mother and the infant. However, the videotapes were used for supervision purposes to ensure the “fidelity of the intervention” itself.

Another study investigated whether CBT could be delivered through the internet [40]. They used a specific program called MumMoodBooster. However, this did not include any VR elements.

A third study used video recordings to assess the interaction between mothers and infants [68]. The intervention used in the study was infant massage in a support group.

We have identified the lack of clinical trials using VR for the treatment of PND as a gap in the literature. For this reason we are planning to conduct two clinical studies following this review. The first one will be a pilot study which will measure the feasibility of using VR for the treatment of PND. This will act as a prerequisite which will guide us to conduct a clinical trial where we will investigate what effect, if any, VR has on traditional therapies for PND.

Limitations

As a narrative review, this study has some limitations. It is missing some of the qualities and advantages of the methodologies that a systematic review and/or a meta-analysis can provide. This review does not give strong emphasis on the methodological flaws of the initial studies included in it. It also does not provide any statistical analysis of the data of the included studies which could highlight issues around variations in individual studies, heterogeneity, or effect size. In addition this review does not assess risk of bias in a systematic way. It does not use any specific tool to assess bias in regards to allocation concealment, blinding of participants, incomplete outcome data, and selective reporting. This review reports on the quality of studies and the reporting of bias within the studies but rather in a brief and not in-depth way.

Another limitation was around the strict inclusion/exclusion criteria we used, and as a result, the total number of studies included was limited. We excluded studies of different type, such as pilot studies, observational studies, and case studies. We also excluded grey literature such as unpublished data, which otherwise might have offered a different or richer perspective.

We excluded studies which had a different target disorder. In most cases, we included studies only for the treatment of PND. We did not touch upon the prevention of PND, prevention or treatment of ante-natal depression, management of ante-natal or post-natal anxiety or distress.

We decided not to include any studies which did not measure as a primary outcome, the depressive symptomatology of the participants. For example we excluded studies which had different outcome measures such as mother-infant relationship. Although, it is known that the mother’s emotional state can affect the infant in different aspects of their lives at a later stage, such as their cognitive or emotional development. However, for the purpose of this review, we decided to exclude any studies that did not have the depressive symptomatology of the participants as the main or primary outcome measure, which is limiting in itself.