Effects of interventions for women and their partners to reduce or prevent stress and anxiety: A systematic review

Background: The period from conception to two years postpartum (the first 1000 days) represents a normative transitional period, which can be potentially stressful for some parents. Parental stress and anxiety adversely impacts psychological and physical health for parents and children. Aim: The aim of this review is to systematically examine effects of interventions for women and their partners to reduce or prevent stress and/or anxiety during the first 1000 days. Methods: MEDLINE, Embase, CINAHL, PsychINFO, and Maternity and Infant Care were searched from inception to March 2019. Randomised controlled trials examining intervention effects on parental stress and/or anxiety during first 1000 days were eligible for inclusion. Data were independently extracted by two reviewers and narratively synthesised. Findings: Fifteen interventions, reported in 16 studies, met inclusion criteria (n = 1911 participants). Overall, findings were inconsistent and the majority of trials demonstrated high risk of bias. Interventions were predominantly delivered to women during pregnancy and only two studies included fathers. There was some evidence that adapting interventions to the pre and postnatal periods provided benefits for stress and anxiety reduction, however there was limited evidence for individual intervention types or approaches. Conclusions: There is currently inconsistent evidence of what interventions are most effective for women during the first 1000 days and there is insufficient evidence for any interventions for male partners during this period. There is a clear need for rigorous development and examination of interventions developed specifically to reduce or prevent stress and/or anxiety across the first 1000 days. © 2020 Australian College of Midwives. Published by Elsevier Ltd. All rights reserved. Statement of Significance


What is Already Known
Some interventions (e.g. psychotherapy) demonstrate reductions in stress and anxiety but the role of interventions for women and partners across perinatal period has not been systematically examined.

What the Paper Adds
There is currently insufficient evidence for effectiveness of stress and anxiety reduction interventions directed at parents during the first 1000 days. There is a need to develop, implement and evaluate interventions to reduce or prevent parental stress and anxiety in the first 1000 days.

Introduction
Stress and anxiety during the period from conception and up to two years postpartum ('the first 1000 days') are experienced by up to 84% of women [1] and 13-14% [2] of men, with prevalence rates varying across trimesters of pregnancy and the postpartum period [3]. Stress and anxiety are highly correlated though distinct constructs [4] characterised by emotional, cognitive, behavioural, and physiological components including responses such as cortisol reactivity [5]. Stress is defined as a person's perceived discrepancy between the demands of external events or stimuli, and perceived resources to meet those demands [6]. Anxiety is defined as a perception of external stimuli as threatening [7], and can be experienced as state anxiety or trait anxiety. Though perinatal stress and anxiety levels are lower in high-income countries than low to middle-income countries [3], there is evidence of association between stress and anxiety, and a range of adverse outcomes in countries such as Australia [8], Canada [9], Norway [10], and France [11]. Stress and anxiety in the first 1000 days have been associated with increased risk of depression [12], impaired relationship functioning [8,12] poor parent-child bonding and attachment [13], and poor health outcomes [14]. Adverse child outcomes include impaired neurodevelopment [15] and motor development [16], increased risk of internalising disorders [17], and emotional, self-regulation difficulties [18], and health difficulties [16].
Parental stress and anxiety have predominantly been examined among mothers, rather than fathers [19]. This may be due to the potential for intrauterine programming of child outcomes via maternal stress, and lower levels of reported stress and anxiety for men than women [20]; though adverse parental and child outcomes are associated with paternal stress [21,22]. The first 1000 days is recognised as a highly stressful transitional period for both men and women [23,24] due to changing roles, and responsibilities [8]. Though not experienced by all men and women during this period, parental stress can be associated with parental health and obstetric issues, [20], social support [25,26], perceived self-efficacy [27], sociodemographic factors [20], and stressful life events [26] including bereavement [28]. Antecedents may differ between men and women [2,20,23] and may vary at different time-points across the first 1000 days [29,30]. As such, focusing on the prenatal period only may limit our understanding of the complexity of parental stress.
Interventions for prenatal and postpartum stress and/or anxiety to date include, but are not limited to, cognitive behavioural therapy (CBT), mindfulness, music, and exercise [31]. While intervention effects are inconsistent [31], approaches including psychosocial [32], psychotherapeutic [19], CBT [33], and mind-body interventions [19,33,34] demonstrate greater positive effects for women at risk for mental health issues and/or experiencing elevated prenatal stress or anxiety, than for women who are not at risk. For men, a lack of support and tailored stress and anxiety treatment options during adjustment to the transition to fatherhood have been identified [35]. CBT and group work, digital support interventions, and clinic-based antenatal childbirth education demonstrate benefits for fathers' perinatal anxiety and depression [35,36] although evidence is limited [36].
While a range of interventions has been used to address stress and anxiety during the first 1000 days, the effects of interventions designed specifically to target stress and/or anxiety during this period have yet to be systematically examined. The aim of this review is to systematically review the effects of interventions, for women and their partners during the first 1000 days, to reduce or prevent stress and/or anxiety.

Searches
This review was registered in PROSPERO (CRD42019126057) and is conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [37]. MEDLINE, Embase, and CINAHL, Psy-chINFO, and Maternity and Infant Care were searched from inception to March 2019. Reference lists of identified articles and of relevant reviews [31,34] were also searched.

Search terms
Search terms used were: (Pregnan* or antenatal or perinatal or maternal or prenatal or postpartum or antepartum or parenting or parent or parental or father or dad or mum or mother or mom or '1000 days' or paternal) AND (intervention* or strateg* or treatment or program or programme or policy or policies or education or prevention) AND (stress or distress or anxiety or mental health, psychological health) AND (RCT or randomised control* trial or randomized control* trial). See Supplementary file 1.

Eligibility criteria
Studies were eligible for inclusion if they were randomised controlled trials (RCTs) that: (1) included pregnant women, expectant fathers and/or partners during pregnancy, and/or parents in the first two years postpartum, from economically developed countries only (based on membership of the Organisation for Economic Co-operation and Development); (2) used a standard care or active control comparator group (e.g. included components and/or activities such as information provision or discussion of symptoms or concerns); (3) examined effects of nonpharmacological interventions developed to prevent or reduce stress and/or anxiety; (4) examined effects on parental stress and/ or anxiety. Only studies published in English, Dutch or Portuguese were eligible for inclusion.

Screening and data extraction
Two reviewers (KMS & CF), independently screened titles and abstracts against eligibility criteria and, subsequently, full texts for eligibility (KMS & CF); disagreements were discussed with a third reviewer and resolved by consensus. Data were extracted using a standardised data extraction form (Supplementary File 2).

Quality assessment
Risk of bias for each study was assessed by two reviewers (KMS & CF) using standard Cochrane risk of bias criteria [38]. The GRADE approach was used to assess the quality of the body of evidence for stress and anxiety outcomes. As all included studies were RCTs, they were initially rated as high quality, with quality downgraded for serious (one level) or very serious (two levels) limitations related to risk of bias, indirectness of evidence, inconsistency, imprecise estimates of effects, or potential publication bias [39]. Study quality was not used as an exclusion criterion and so no studies were excluded on the basis of quality. This is because adopting a more stringent approach to study selection by including RCT designs only, is suggested to ensure higher quality as per the GRADE criteria; it was thus deemed more appropriate to examine quality within identified RCTs than to exclude on this basis.

Data synthesis
It was intended that a meta-analysis be conducted in Review Manager 5.3 software using studies reporting crude and/or adjusted estimates. Due to substantial heterogeneity between studies in terms of interventions and outcomes reported, a meta-analysis was not considered to be appropriate. A narrative synthesis of intervention effects, characteristics and theoretical underpinnings, was instead conducted. As interventions were conducted predominantly in either the prenatal or postnatal periods, with three interventions delivered in both periods, findings are presented for prenatal, postnatal, and both pre and postnatal periods.

Results
Full texts of 71 articles were assessed against eligibility criteria. Sixteen studies, representing 15 interventions, met inclusion criteria for the current review (Fig. 1).

Study quality
Only one study was rated as low risk of bias [49], six studies were rated as moderate risk of bias [40][41][42][43][44]50]; nine studies were rated as high risk of bias [38,39,41,42,47,48,[51][52][53]. The most common sources of bias related to blinding; selective reporting bias in most studies was unclear as few studies were pre-registered or had published protocols. Though individual studies demonstrated varying quality, the quality of evidence for stress and anxiety outcomes overall was rated as low due to serious limitations regarding risk of bias, indirectness of evidence, inconsistency, imprecision of effect estimates and potential publication bias, as outlined by the GRADE criteria. See Supplementary file 3.

Study populations
Characteristics of the included studies are summarised in Table 1. Seven studies [40,42,[44][45][46][47][48] examined effects of prenatal interventions for women at risk of developing stress, anxiety and/or depression [45,48], or who already experienced elevated levels [40,42,44,46,47]. One study examined effects of an intervention for low-income pregnant women who were, for the most part, unmarried [41]. Postpartum interventions included women with anxiety and/or depression [43]; parents of preterm infants [49,50] or infants born with very low birth weight [51]. Four studies were conducted in populations who are not considered 'at-risk', herein referred to as 'general population' [52][53][54][55]. Two studies included fathers [49,50] and were conducted in the postnatal period; no study examined same sex partners (Table 1).

Intervention effectiveness
Effects of interventions were inconsistent across the reviewed studies; two interventions demonstrated reductions in stress or anxiety, eight studies demonstrated inconsistent effects, and five studies reported no effects for stress or anxiety. Only results for stress and anxiety are described in detail herein; findings for additional outcomes are summarised in Table 2.

Interventions demonstrating reductions in stress and/or anxiety
Two studies from the same research group examined effects of a brief unguided online cognitive behavioural intervention adapted for either pregnancy [40] or the postnatal period [43]. The intervention included women with or at risk of developing stress, anxiety and/or depression [40]. Intervention content was tailored to reflect symptoms and challenges in pregnancy [40] or postpartum [43] respectively. In the prenatal study, reductions were reported for generalised anxiety disorder, as measured by the Generalised Anxiety Disorder scale (GAD; p < 0.01, g = .76) and general psychological distress, measured by the Kessler-10 psychological distress scale (K-10; p < 0.01, g = .52). These reductions were observed in the third trimester relative to women receiving usual care. In the postnatal intervention there was a significant reduction in anxiety as measured by the GAD (p < 0.001), and distress, as measured by the K-10 at approximately 1-7 months postpartum.
Interventions demonstrating inconsistent effects for stress and/or anxiety Prenatal interventions. Five studies conducted during the prenatal period reported inconsistent intervention effects for stress and/or anxiety outcomes. One study examined effects of an adapted cognitive behavioural intervention for women experiencing elevated levels of anxiety or depression at approximately 16 weeks gestation [46,47]. The control conditions included usual care for women with elevated anxiety or depression. This study reported a significant reduction in the cortisol awakening response (CAR; p = 0.047) at 3 months postpartum [47]. No effect was found for cortisol area under the curve (AUC; p = 0.83), self-reported prenatal stress measured using the prenatal distress questionnaire (PDQ; p = 0.88) or general stress measured using the Perceived Stress Scale (PSS; p = 0.92) [47]. Another study examined effects of a prenatal mindfulness intervention on perceived stress and anxiety at a 3-month follow-up in a group of women with a history of mood concerns [48]. Timing of follow-up measurement ranged from approximately 7.5 months gestation to 1.5 months postpartum. The intervention demonstrated a reduction in anxiety, measured using the State Trait Anxiety Inventory (STAI) (p = 0.04, d = .85) but no effect for stress using the PSS (p = 0.35, d = .39) [48].
A third intervention included information provision on stress and stress management techniques, including breathing and muscle relaxation for a general pregnant population [53]. The intervention was delivered for 6 weeks from approximately 17 weeks gestation and was examined relative to a wait-list usual care control condition. At an average of 23 weeks gestation, a reduction in perceived stress measured with the PSS (mean difference À3.23 (95% CI: À4.29 to À0.29)). There was no change in state anxiety (mean difference À1.5 (95% CI: À2.7 to 1.7)) or trait anxiety (mean difference À2.29 (95% CI: À4.9 to 0.3)) as measured by the STAI. A fourth study included HRV biofeedback, which aims to increase HRV through paced breathing exercises using a small handheld heart rate measurement device [44]. The intervention took place as five weekly sessions and daily at home practice of progressively increasing duration per week with a sample of pregnant women experiencing stress. The control group comprised a wait-list condition of pregnant women experiencing stress. Relative to the control group, the intervention demonstrated reductions in anxiety (p = 0.001) but not stress (p = 0.19), both measured using the Depression Anxiety and Stress Scale (DASS). Outcomes were assessed at roughly 24 weeks gestation. The fifth prenatal study involved peer-mentoring, and was delivered in the 1st and 2nd trimesters to a sample of active duty women and wives or military service members [54]. The control condition comprised active duty women and wives of military service members receiving usual care. At approximately 30 weeks gestation the intervention demonstrated reductions in anxiety related to identification with motherhood (p = .049) and preparation for labour (p < 0.005). No effects were reported for anxiety related to acceptance or helplessness (p > 0.05), or for depression, self-esteem or resilience (p > 0.05); all outcomes were measured using The Lederman Prenatal Self-Evaluation Questionnaire-Short Form [56].
Postpartum interventions. One of the two postpartum interventions intervention [50] which demonstrated inconsistent effects was for parents of children in the neonatal intensive care unit (NICU) 10-15 days after delivery. The intervention included information provision and relaxation techniques, delivered as five 90-min sessions during the infant's NICU stay and an audio CD for at home use. The intervention was examined in comparison to an active control involving information provision regarding preterm infant care. Three months following infant discharge from the NICU, there was no effect on perceived stress overall, as measured by the PSS (p = 0.70); an increase in stress was reported for participants with high baseline stress levels (p < 0.001), higher education (p < 0.005) and lower income satisfaction (p < 0.005). No effects were reported for morning cortisol (p = 0.94), +30 min cortisol (p = 0.26) or bedtime cortisol (p = 0.26). There was also no effect for state anxiety (p = 0.52) but a significant reduction was reported for trait anxiety (p < 0.005), as measured by the STAI. Results in this study were not examined separately for mothers and fathers [50].
The second study examined effects of a mother-father-infant (triadic) attachment intervention program for parents of preterm infants [49]. The intervention focused on parent-infant interactions and parental understanding of infant development, to reduce stress and improve parent-infant relationships to enhance infant developmental outcomes. Using the Parenting Stress Index-Short Form (PSI-SF) to evaluate stress, there was no effect on parenting stress for mothers or fathers when the infant was approximately 3 months old. At 18 months, the intervention demonstrated reductions in global stress for mothers (p < 0.001) and fathers (p < 0.005); reduced parenting stress for mothers (p < 0.005) but not fathers (p = 0.07); and reduced parent-child stress for mothers (p < 0.001) and fathers (p < 0.005), relative to a usual care control condition of parents of preterm infants.
Prenatal and postpartum interventions. One study examined effects of a nurse community health worker home-visiting intervention with a group of low-income pregnant women, the majority of whom were unmarried [41]. At 15 months postpartum there was no effect for perceived stress, measured using the PSS, in comparison to Medicaid care as usual (p = 0.058). For women with low psychosocial resources, the intervention significantly reduced stress (p < 0.005); there was no effect based on baseline stress differences (p = 0.34) or when baseline stress and psychosocial resources were combined (p = 0.13). Another study examined the effects of a CBSM intervention in a general pregnant population [55]. At 6 months postpartum the intervention reduced stress, measured using a visual analogue scale, relative to usual care among a general pregnant population (p < 0.01); no differences were reported for morning, evening or average cortisol, or cortisol slope (all p > 0.05). At 18 months, there was no effect on stress, morning or evening cortisol, or cortisol slope (all p > 0.05). A significant reduction in average cortisol, based on samples collected 45 min after waking and at 8 pm at night, was reported (p < 0.05).

Interventions demonstrating no effects for stress and/or anxiety
Five studies reported no effects for stress and/or anxiety. These interventions may have demonstrated beneficial effects for other outcomes (Table 1) but as stress and anxiety are the outcomes of interest in the current review, these interventions are categorised as having no effect here. No studies reported negative effects of the interventions however. Four of these interventions were delivered in the prenatal period and one was delivered in the postpartum period.
One study examined the effects of a CBT intervention for pregnant women at risk for developing stress, anxiety and/or depression, on state and trait anxiety up to 4 months postpartum [45]. The intervention was delivered in the late 1st and early 2nd trimester. At 4 months postpartum there was no effect for anxiety (p > 0.05), as measured by the STAI, in comparison to a control group who were provided with verbal and written information on perinatal mental health [45]. A second study examined effects of an adapted cognitive behavioural intervention for women experiencing elevated levels of anxiety or depression at approximately 16 weeks gestation [46,47]. This study reported no effect for anxiety, measured using the STAI, at approximately 24 weeks gestation (p = 0.25) or up to 3 months postpartum (p = 0.53) in comparison to usual care received by women with elevated anxiety or depression [46]. There was no effect for anxiety sensitivity, using the anxiety sensitivity scale, at approximately 24 weeks gestation (p = 0.41) or up to 3 months postpartum (p = 0.14). A third CBT-focused study examined effects of a cognitive behavioural stress management (CBSM) intervention for low-income women up to 3 months postpartum, in comparison to a control condition involving care as usual and Information on locally available social services. A significant effect was observed for perceived stress, using the PSS, (p < 0.05) and diurnal cortisol slope (p < 0.05). Intervention condition participants with high baseline anxiety demonstrated steeper decreases than low anxiety participants (p < 0.05). There was no effect for area under the curve (AUC; p = 0.31) or the CAR (p = 0.07) but non-Latina intervention participants demonstrated greater CAR decline than non-Latina controls (p < 0.05). The final prenatal study examined effects of a mindfulness-based intervention in a general population of pregnant women, compared to a control condition involving activities such as identification of stressors and strengths (see Table 2) [52]. Six weeks following intervention cessation, at 30-36 weeks gestation, there was a non-significant effect for stress measured using the PSS (p = 0.82 d = 0.15) [52].
The postpartum study examined an intervention for mothers of very low birth weight infants in comparison to a control condition involving information provision and contacts with a 'care convener' [51]. The intervention was delivered when the child was a mean age of 33 days, and provided information on recognising and utilising strategies for managing one's own anxiety and information on recognising and responding to the infant's cues. At 6-8 weeks corrected infant age, there were no effects of the intervention for stress related to infant behaviour or appearance (p = 0.14) or parent role restriction (p = 0.76); these were assessed using the Parental Stressor Scale: Neonatal Intensive Care Unit [57]. There was also no effect for anxiety measured by the STAI (p = 0.28) or post-traumatic stress disorder (PTSD) measured by the Perinatal PTSD Questionnaire (p = 0.54).

Theory use in interventions
Nine studies explicitly reported using theory to inform the intervention. Of the seven studies [40,[45][46][47]50,51,53] not explicitly reporting using theory, interventions such as those based on CBT [40,[45][46][47] and HRV biofeedback [44], are derived from theory based approaches. Reported theories and models varied between interventions and no theories or models were used in multiple interventions; all reported theories and models were used in one intervention only (See Table 2). Reported theories included the co-emergence model of behaviour reinforcement [58]; mindfulness-based cognitive behavioural model; the Ecological Stress Theoretical Framework [59,60]; Social Learning Theory [61], Lewinsohn's approach to mood management [62]; and attachment theory [63].

Discussion
This review identified 15 interventions, in 16 studies, developed to reduce or prevent stress and/or anxiety during the first 1000 days. Three main findings emerged from the current review. Firstly, findings were inconsistent, with most studies at high risk of bias and just one intervention reducing stress and anxiety in both periods. Secondly, interventions were highly heterogeneous with variation in intervention type, timing of delivery, target populations, and outcome measures. Thirdly, the majority of interventions were delivered prenatally and/or to vulnerable populations, with few delivered postnatally or to male partners, and none for same sex partners.

Intervention effects
Intervention outcomes were inconsistent overall, with effects differing for different intervention types. The most commonly used intervention approach in reviewed studies was CBT. While the current review found insufficient support for CBT interventions to reduce parental stress or anxiety overall, one online cognitive behavioural intervention that was adapted for use in the prenatal and postpartum periods demonstrated reductions in anxiety and distress. This highlights some potential usefulness for online CBT tailored for use across the first 1000 days and previous reviews have identified CBT-based interventions as effective treatments for perinatal distress [34,35]. Recent reviews have also highlighted some beneficial effects of mindfulness during the perinatal period [64] but the current review does not provide support for effects on stress and anxiety in the first 1000 days. This discrepancy may be due to a broader range of interventions (i.e. those not explicitly designed for stress and/or anxiety) and time-periods (i.e. prenatal or postpartum only) included in previous reviews. This review included only those interventions specifically developed to reduce or prevent stress and/or anxiety and so additional components included in broader mindfulness or CBT interventions may have contributed to previously observed effects. Psychoeducation was included in all reviewed interventions, which is unsurprising as there is consistent evidence that parents want clear, consistent information about aspects of their pregnancy and early parenting [65][66][67]. The importance of psychoeducation for stress and anxiety prevention/reduction has previously also been noted [31,36,68]. Findings of the current review further support the importance of psychoeducation in that studies utilising active control conditions, particularly those based on information provision, were more likely to report no effect [45,51,52].

Methodological considerations of reviewed studies
A number of methodological issues may influence observed findings of the current review. For instance, use of appropriate theory is essential to guide development of appropriate and effective interventions [68] but few interventions in the current review explicitly reported being based on theory; some interventions were based on theoretical approaches such as CBT. Social-ecological approaches to stress, as included in one reviewed study [41], provide useful approaches to understand and target stress and anxiety at multiple levels across the first 1000 days. Greater consideration and inclusion of appropriate theoretical underpinnings for future interventions is needed. High risk of bias in the majority of reviewed studies, and low quality of examined outcomes across trials, is also problematic in the current review. For instance, the use of small sample sizes, lack of trial protocols or pre-registration, and poor reporting of confidence intervals and effect sizes limit confidence in observed findings. Further, variability across studies in conceptualisations and definitions of stress and anxiety potentially impacts intervention development and outcome measurement.
Timing of outcome measurement is also problematic in some studies because varying times of outcome assessment can introduce variability in intervention outcomes [3]. In addition, few studies examined effects of interventions on potential mediators or mechanisms of change. While this may be due to small sample sizes limiting statistical power, it limits conclusions that can be made about direct or indirect effects of interventions. Similarly, timing of intervention delivery is an important consideration in evaluation and interpretation of intervention effects. The majority of interventions included in the current review were delivered in the prenatal period. Considerably fewer interventions were delivered in the post-partum period, or across the first 1000 days. Prenatal and post-partum stress and anxiety can differ in terms of antecedents and potential outcomes [69]. As the expected mechanisms underlying treatment effects are likely to differ between pregnancy and postpartum, it is logical that separate interventions are developed for each period. However, as the first 1000 days represents a potentially stressful transitional period with stress and anxiety during pregnancy influencing experiences in the postpartum [70], provision of supports across this period is appropriate and necessary. In the current review, the only intervention adapted for both the prenatal and postnatal periods [40,43] demonstrated reductions in anxiety and distress in both periods. This represents a beneficial tailored approach across the first 1000 days that should be considered in future research.

Study populations included in reviewed studies
The study populations included in the reviewed studies can also influence likelihood of effects and the majority of women and their partners in the current review were at increased risk of adverse mental health outcomes. Previous reviews have reported greater benefits for treatment than preventive interventions [34]. Inclusion of participants experiencing, or at risk for, psychological distress may increase the likelihood of detecting intervention effects; particularly as large sample sizes are needed to detect changes in stress and anxiety symptoms in general populations for which baseline levels are not elevated. In line with this, interventions conducted with general pregnant populations in the current review demonstrated inconsistent findings [53][54][55]. As such the findings of the current review are more applicable to women at high-risk of mental health issues, limiting generalizability to women at low risk. It is important to note however that general populations may include individuals with stress and anxiety symptoms that could become exacerbated as pregnancy progresses and/or in early parenthood due to the stressful transitional nature of this period [23,24]. Robust longitudinal examinations of interventions that provide skills to manage stress and anxiety across the first 1000 days are needed to better determine effects of preventive, in addition to reduction-focused, interventions.
Inclusion of partners in future stress and/or anxiety reduction interventions in the first 1000 days should also be considered. Though fathers tend to report lower stress and anxiety than mothers [20], paternal stress is associated with maternal stress [27], perceived social support [24], and child development [71]. Only two reviewed interventions included fathers, and these were delivered postpartum to fathers of preterm infants with inconsistent effects [49,50]. This review cannot therefore provide support for interventions for fathers in the first 1000 days. Future interventions including fathers are needed to investigate potential direct and indirect effects via mechanisms, such as social support, for parents and children.

Limitations of the review
The inclusion criteria used in the current review was designed to identify interventions developed specifically to reduce or prevent stress and/or anxiety. Therefore, not all interventions that have been used to reduce or prevent stress and/or depression have been included. However, the findings highlight the limited number of interventions that are explicitly stress and/or anxiety focused. A range of diverse intervention types and populations were included in the identified papers, which limited evidence synthesis. However, it appears that evidence for effective interventions that are appropriate for use with both men and women across the first 1000 days is scant.

Conclusion
The findings of the current review highlight insufficient evidence for the effectiveness of stress and anxiety interventions for women and their partners from conception to two years postpartum. Future interventions should include psychoeducation and may benefit from inclusion of partners, though evidence from this review is limited. Further research on preventive interventions may also be useful as this review provides more evidence for intervention effects in vulnerable populations. While stress and anxiety prevention and reduction should not be a 'one-size fits-all' approach, the prevalence of stress and anxiety for mothers and fathers across the first 1000 days warrants a reappraisal of for whom stress and anxiety interventions and approaches are developed and implemented. Evidence from this review indicates that interventions in both prenatal and postpartum periods may result in reductions in stress and anxiety; with evidence from one study highlighting benefits of tailoring interventions across these periods [40,43]. Development and methodologically robust examination of interventions specifically targeting stress and anxiety for use across, or tailored to, the prenatal and postnatal periods is essential.

Funding
Dr Matvienko-Sikar is supported by a Health Research Board ARPP Grant (HRB-ARPP-A011). Dr Flannery is supported by a Health Research Board ICE Grant (HRB ICE-2015-1026). The funders had no role in the conceptualisation, conduct or reporting of this research.

Ethical statement
This manuscript reports a systematic review of previously published studies. As such, an ethical statement is not applicable.

Conflicts of interest
Karen Matvienko-Sikar, Caragh Flannery, Sarah Redsell, Catherine Hayes, Anja Huizink and Patricia M Kearney declare that they have no