Psychosocial and psychological interventions for treating postpartum depression
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The primary objective of this review is to assess the effects of psychosocial and psychological interventions compared with usual postpartum care in the reduction of depressive symptomatology.
Secondary objectives are to examine:
(1) the effectiveness of specific types of psychosocial interventions (e.g. non‐directive counselling, support groups);
(2) the effectiveness of specific types of psychological interventions (e.g. interpersonal psychotherapy, cognitive behavioural therapy);
(3) the effects of intervention mode (e.g. individual versus group‐based interventions);
(4) the effects of sample selection criteria (e.g. targeting women with clinically diagnosed depression versus self‐reported depressive symptomatology).
Background
Postpartum mood disorders are a common form of maternal morbidity following delivery (Stocky 2000). These affective disorders range in severity from the mild and transient "baby blues" experienced by 50% to 80% of women to postpartum psychosis, a serious condition which affects less than 1% of mothers and usually requires hospitalisation (Evins 1997). Among these disorders is postpartum depression, a non‐psychotic depressive episode that begins or extends into the first year postpartum. This condition often exhibits the disabling symptoms of uneasiness, irritability, confusion and forgetfulness, anhedonia, fatigue, insomnia, anxiety, guilt, inability to cope, and thoughts of suicide. Frequently exacerbating these symptoms are low self‐esteem, lack of confidence, and unrealistic expectations of motherhood. The development of postpartum depression is greatest in the first 12 weeks postpartum with duration frequently dependent on severity (Cox 1993). Some residual depressive symptoms are common up to a year after delivery (Cooper 1998).
Postpartum depression is a major health issue for many women from diverse cultures (Affonso 2000). Longitudinal and epidemiological studies have yielded varying prevalence rates, ranging from 3% to more than 25% of women in the first year following delivery; these rates fluctuate due to sampling, timing of assessment, differing diagnostic criteria (major or minor depression), and whether the studies were retrospective (low rates) or prospective (6‐ to 10‐fold higher). Frequently cited estimates range between 10% to 15% and a meta‐analysis of 58 studies reported the prevalence of postpartum depression to be 13% (O'Hara 1996). It is noteworthy that the absolute difference in estimates between self‐report assessments of depressive symptoms, such as the commonly used Edinburgh Postnatal Depression Scale (Cox 1987b) (which does not diagnose postpartum depression), and standardized diagnostic interviews (which do diagnose postpartum depression) was small.
This morbidity has well documented public health consequences for the mother, child, and family. While women who have suffered from postpartum depression are twice as likely to experience future episodes of depression over a five‐year period (Cooper 1995), infants and children are particularly vulnerable. Postpartum depression can cause impaired maternal‐infant interactions (Murray 1996) and negative perceptions of infant behaviour (Mayberry 1993), which have been linked to attachment insecurity (Hipwell 2000; Murray 1992), cognitive developmental delay (Cogill 1986; Hipwell 2000) and social/interaction difficulties (Cummings 1994; Murray 1999). Infants as young as three months of age have been shown to ably detect their mothers' mood and to modify their own responses accordingly (Cohn 1983). While cognitive skills (Whiffen 1989), expressive language development (Cox 1987a), and attention (Breznitz 1988) have been negatively influenced by postpartum depression, it has also been reported that children of depressed mothers are two to five times more likely to develop long‐term behavioural problems (Beck 1999; Orvaschel 1988). Child neglect/abuse (Buist 1998) and marital stress resulting in separation or divorce (Boyce 1994) are other reported outcomes. Maternal and infant mortality are rare but real consequences of postpartum depression.
The aetiology of postpartum depression remains unclear with little evidence to support a biological basis (Cooper 1998; O'Hara 1997). Despite considerable research, no single causative factor has been isolated and a multifactorial etiology has been suggested. However, consistent findings suggest the importance of psychosocial variables (Beck 2001; Cooper 1998; O'Hara 1997). In particular, stressful life events (Bernazzani 1997; O'Hara 1991), marital conflict (Bernazzani 1997; O'Hara 1991; O'Hara 1986), and the lack of social support (Bernazzani 1997; Brugha 1998; Chen 1999; Cooper 1998; O'Hara 1986; Small 1994; Stein 1989; Stuchbery 1998) have been found to significantly increase the risk of postpartum depression. The saliency of social support was especially highlighted in a predictive study of several thousand women, in which mothers who lacked social support were approximately two times more likely to develop postpartum depression than mothers with sufficient support (Cooper 1996).
Antidepressant medication, cognitive behavioural therapy (CBT), and interpersonal psychotherapy (IPT) have been validated as effective treatments for general depression. However, mothers are often reluctant to take antidepressants due to concerns about breast milk transmission or potential side‐effects (Dennis in press). Although there is evidence that antidepressants are relatively safe for breastfed infants and a Cochrane review has been completed examining the effectiveness of antidepressant medication for the treatment of postpartum depression (Hoffbrand 2001), the American Academy of Pediatrics classifies most antidepressants as drugs whose effect on nursing infants is unknown but may be of concern. Given these considerations, it is important that non‐pharmacologic interventions be evaluated for use with postpartum women. Based on the preceding risk factors, a variety of psychosocial and psychological interventions have been developed to prevent postpartum depression (Dennis 2004a) and a Cochrane systematic review was completed (Dennis 2005). The subgroup analyses results of this review suggested that preventive interventions were more likely to be effective if they were individually based, initiated in the postpartum period, and targeted 'at risk' women. Research now indicates that psychosocial (e.g. non‐directive counselling, Holden 1989) and psychological (e.g. interpersonal psychotherapy O'Hara 2000; cognitive behavioural therapy, Appleby 1997) interventions may also be used to treat postpartum depression (Dennis 2004b). This review will examine the effectiveness of psychosocial and psychological interventions in the treatment of postpartum depression where treatment is defined as any intervention initiated among mothers after 2 weeks postpartum who have been identified with depressive symtpomatology.
Objectives
The primary objective of this review is to assess the effects of psychosocial and psychological interventions compared with usual postpartum care in the reduction of depressive symptomatology.
Secondary objectives are to examine:
(1) the effectiveness of specific types of psychosocial interventions (e.g. non‐directive counselling, support groups);
(2) the effectiveness of specific types of psychological interventions (e.g. interpersonal psychotherapy, cognitive behavioural therapy);
(3) the effects of intervention mode (e.g. individual versus group‐based interventions);
(4) the effects of sample selection criteria (e.g. targeting women with clinically diagnosed depression versus self‐reported depressive symptomatology).
Methods
Criteria for considering studies for this review
Types of studies
All published, unpublished and ongoing randomised controlled trials and quasi‐randomised trials of psychosocial or psychological interventions in which the primary or secondary aim is a reduction in depressive symptomatology.
Types of participants
Women with depressive symptomatology that developed within the first 12 months postpartum who were identified either via self‐report measures (e.g., Edinburgh Postnatal Depression Scale (EPDS) (Cox 1987b), Beck Depression Inventory, (BDI) (Beck 1961) or a diagnostic interview (e.g. Structured Clinical Interview for DSM‐IV, (SCID) (Spitzer 1992)).
Types of interventions
Any form of standard or usual care compared to a variety of non‐pharmaceutical interventions (including psychoeducational strategies, cognitive behavioural therapy, interpersonal psychotherapy, non‐directive counseling, various supportive interactions, and tangible assistance), delivered via telephone, home or clinic visits, or individual or group sessions in the postpartum by a health professional or lay person. Psychosocial interventions are unstructured, non‐manualised and not theoretically underpinned.
Treatment comparions
The following treatent comparisons will be made:
(1) Psychosocial interventions (e.g. non‐directive counselling, support groups) versus standard care or usual care
(2) Psychological interventions (e.g. interpersonal psychotherapy, cognitive behavioural therapy) versus standard care or usual care
(3) Psychosocial interventions versus psychological interventions
Types of outcome measures
Primary outcome measure
The primary outcome measure in this review is Postpartum depression (as variously defined and measured by trialists e.g. SCID, BDI and EPDS).
Secondary outcome measures
1. Maternal outcomes
(a) Maternal mortality and serious morbidity including self‐harm, suicide attempts
(b) Health service utilization including outpatient and inpatient use of psychiatric unit, other health services
(c) Maternal‐infant attachment
(d) Maternal attitudes towards motherhood
(e) Anxiety (e.g. State Trait Anxiety Scale, (Spielberger 1970)).
(f) Stress
(g) Maternal confidence
(h) Maternal competence
(i) Self‐esteem
(j) General health
(k) Dissatisfaction with intervention
(l) Perceived social support
2. Infant outcomes
(a) Breastfeeding duration (variously defined)
(b) Breastfeeding level (exclusive, almost exclusive, high, partial, token, bottle‐feeding)
(c) Infant health parameters including immunization, accidental injury, non accidental injury
(d) Infant developmental assessments (variously defined)
(e) Child abuse and/or neglect
(f) Neonatal/infant mortality
(g) Neonatal/infant morbidity
(h) Quality of mothering (variously defined)
3. Family outcomes
(a) Marital discord
(b) Marital separation/divorce
Search methods for identification of studies
We will search the The Cochrane Collaboration Depression Anxiety and Neurosis group trials registers (CCDANCTR‐Studies and CCDANCTR‐References). This is a specialised register that contains more than 9,000 records on trials comparing treatment options that are within the scope of the CCDAN. The register is updated regularly adding the results on searches of The Cochrane Library, CINAHL, EMBASE, LILACS, MEDLINE, National Research Register, PsycINFO, PSYNDEX and SIGLE. Also, quarterly systematic screening of relevant journals and conference proceedings takes place (for information on the full search strategies, visit http://web1.iop.kcl.ac.uk/IoP/ccdan/index.htm.) Using these registers, the following search strategy will be used:
Diagnosis = "Depression, Postpartum" or "Depression, Antenatal"
or
Diagnosis = Depress* and Comorbid Diagnosis= Pregnan*
CCDANCTR‐References (for those references not yet coded into studies) will also be searched using the following search strategy
Keyword = Depress*
and
Freetext = postpartum or post‐partum or postnatal or post‐natal or puerper* or antenatal or ante‐natal or pregnan*
or
Freetext = baby‐blue*
In addition we will search the Cochrane Pregnancy and Childbirth Group trials register, the Cochrane Central Register of Controlled Trials, MEDLINE (1966 to 2006), EMBASE (1980 to 2006) and CINAHL (1982 to 2006) using various combinations of the terms postpartum/postnatal depression. We will also scan secondary references and contact experts in the field to identify other published or unpublished trials.
Data collection and analysis
Selection of trials
Titles and abstracts of the electronic searches will be reviewed by the primary reviewer. Trials under consideration will be evaluated by both reviewers independently for methodological quality and appropriateness for inclusion. In the case of uncertainties regarding the appropriateness for inclusion, resolution will be established through discussion and consensus.
Methodological quality assessment
Quality assessment of the trials that met the eligibility criteria will be carried out by both reviewers using the following criteria:
(1) randomization: A = adequate, B = unclear, C = inadequate, D = no information;
(2) blinding of outcome assessor: yes, no, inadequate, no information;
(3) blinding of caregivers not providing the intervention: yes, no, inadequate,
(4) blinding of outcome assessment: yes, no, inadequate or no information;
(5) completeness of follow up data (including any differential loss of participants from each group): A = < 3% of participants excluded, B = 3% to 9.9% of participants excluded, C = 10% to 19.9% excluded, D = 20% or more excluded, E = unclear;
(6) analysis of participants in the same groups randomised (intent‐to‐treat analysis).
Both reviewers will assign a rating to each trial; results will be compared and differences discussed until agreement is obtained. Reasons for exclusion of any apparently eligible trial will be clearly described.
Data extraction
Data will be extracted independently from trial reports by both reviewers using a standardized data extraction form. Wherever necessary, unpublished or missing data will be requested from the trial contact author. In addition, data will be sought to allow an 'intent‐to‐treat' analysis. Data will be entered into RevMan 4.2.8 by one reviewer and double data entry will be completed by the other reviewer.
Data synthesis
Trials using different treatment strategies will be analysed separately and the results combined only if there is no reason to think that they differ in relevant ways. While the primary meta‐analysis will be based on the occurrence of postpartum depression or not (however measured by trialists), several depression rating scales or cut‐off points will be incorporated. To address the potential measurement differences, direct comparisons between trials using the same rating scale and cut‐off will be made using the weighted mean difference (WMD). Where trials use different ways of measuring the same continuous outcome, the standardised mean difference will be used (SMD). The random effects model will be employed for combining data. Meta‐analyses will be performed using relative risks as the measure of effect size for binary outcomes, and mean differences for continuous outcome measures, both with 95% confidence intervals. An assessment will be made of the extent to which there are between‐study differences, including variations in the population or intervention.
Random effect meta‐analysis will be used for combining study data. Where high levels of heterogeneity are found (I2 > 50%), they will be explored by sensitivity analyses, excluding the trials most susceptible to bias based on the following quality assessment: (1) those with unclear allocation concealment (B); (2) high levels of post‐randomisation losses or exclusions (D); or (3) unblinded outcome assessment or blinding of outcome assessment uncertain.
Management of skewed continuous data
The Cochrane Handbook states that with the more common positive skewness, presentation of a geometric mean with its 95% confidence interval is equivalent to an analysis of a log transformation of the data. However, log‐transformed and untransformed data can not be mixed in a meta‐analysis. The Handbook also states that skewness is not necessarily a problem for meta‐analyses in Revman if the sample sizes in the individual studies are large. If we have a small sample size for the specific meta‐analysis and skewed data we will complete the analysis and state that caution should be used when interpreting the results and provide the rationale.
Missing data
We will conduct intention to treat (ITT) analyses where there are missing data for dichotomous outcomes, and completer analyses where there are missing data for continuous outcomes.
Subgroup analyses
Four a priori subgroup analyses are planned:
(i) the effectiveness of specific types of psychosocial interventions (e.g. support groups, non‐directive counselling);
(ii) the effectiveness of specific types of psychological interventions (e.g. interpersonal psychotherapy, cognitive behavioural therapy);
(iii) the effects of intervention mode (e.g. individual versus group‐based interventions);
(iv) the effects of sample selection criteria (e.g. women selected based on clinically diagnosed depression versus self‐reported depression).
These sub‐group analyses can also be used where appropriate to test for clinical heterogeneity.
Publication bias
The existence of publication bias will be examined graphically by use of a funnel plot.
