Journal Logo

Research Article: Study Protocol Systematic Review

Non-pharmacological interventions for postpartum depression

A protocol for systematic review and network meta-analysis

Wang, Yang MDa,b; Li, Hui MD, PhDc; Peng, Wei MDd; Chen, Yalin MDa; Qiu, Mimi MDa; Wang, Jun MDa; Hao, Qinghong MDa; Tu, Yang MDa; Liu, Yunlu MSe,f,∗; Zhu, Tianmin MD, PhDa,∗

Author Information
doi: 10.1097/MD.0000000000021496
  • Open


1 Introduction

Postpartum depression(PPD) is the most common mental disorder for women following childbirth, which can deleteriously affect maternal behaviors and impair the cognitive, behavioral, emotional development of children.[1] Epidemiological studies have found that women with PPD were at high risk for suicidality,[2] and almost 1/3 women still reported depressive symptoms even 4 years after childbirth.[3] Besides, the global pooled incidence of PPD is estimated to be 17.7%, which is up to 38% in Chile.[4] As a consequence, PPD has been considered as a global public health issue.

Given the serious consequences of PPD for both mother and child and the high prevalence of PPD, effective treatment for this disease is of obvious importance. Drug therapy is the main treatment for major depressive disorder (MDD) in clinic and used cautiously during perinatal period.[5,6] Due to concern about infant exposure to antidepressant via breastmilk, many postpartum women are reluctant to start antidepressant therapies.[7] Moreover, PPD are often mixed with bipolar feature, which is usually hard to recognize, therefore inappropriate antidepressant therapies may lead to a deterioration of hypomanic symptoms and increase the risk of suicidality in mother.[8] As a consequence, non-pharmacological interventions have become many PPD patients’ preferences.

Methods of psychological/ psychosocial intervention, physical therapy, kinesitherapy, music therapy, acupuncture are the widely used non-drug approaches for PPD,[9] which have none of negative effects of pharmacotherapy mentioned above, therefore, women with PPD could continue breastfeeding. Cognitive behavioral therapy (CBT) and interpersonal psychotherapy (IPT) are 2 of the most common psychological/ psychosocial interventions with strong evidences on clinical efficacy for treating PPD.[10,11] Physical therapies were also proved effective for PPD, such as repetitive transcranial magnetic stimulation (rTMS) and light therapy. Repetitive transcranial magnetic stimulation, which was widely used in MDD, can significantly improve the PPD symptoms.[12] Furthermore, the efficacy of rTMS has been proved to last even for more than 6 months.[3] Light therapy is another effective physical treatment method, targeting disordered sleep and circadian rhythms, which are main risk factors of PPD.[13] Kinesitherapies, like aerobic exercise and yoga, are considered as the low-cost, freely available interventions for postpartum women with depressive symptoms, which have the potential to become more generally in PPD population.[14,15] Music therapy, demonstrated effective to MDD patients, was also found curative for PDD.[16] Additionally, as one of the most popular alternative medicines in the world, acupuncture therapy could alleviate the depressive and anxious symptoms of PPD women, as well as the estradiol levels.[17]

It can be clearly seen that there is a wide choice of non-pharmacological approaches for physicians, who need high quality evidence to back them to make the best decision in clinical practice. Therefore, to meet the increasing need for comparing the values of these non-drug interventions, a network meta-analysis should be conducted to fill in the gap for assessment of comparative efficacy and understand the relative merits of the various approaches, in the condition where the evidence is indirect.[18] The systematic review and network meta-analysis based on high quality randomized controlled trials (RCTs) aim to inform clinicians by comparing and ranking different non-pharmacological interventions for precise and safety treatment of women with PPD. Furthermore, by analyzing individual treatments or combinatorial usage, we hope the results of the study will provide a wider picture for summarizing and interpreting the existed evidence data.

2 Methods

2.1 Design and registration

This systematic review and network meta-analysis will evaluate the comparative effectiveness and safety of non-pharmacological interventions for PPD. This protocol is designed according to the guideline of Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P)[19] (Fig. 1) and registered on the international prospective register of systematic review (PROSPERO), with the registration number CRD42020166801. The findings of this study will be reported in line with the guideline of Preferred Reporting Items for Systematic Reviews and Network Meta-Analysis (PRISMA-NMA).[20]

Figure 1
Figure 1:
Flow chart of systematic review and network meta-analysis for non-pharmacological interventions for postpartum depression.

2.2 Ethics

Since the present study is a meta-analysis based on existed researches, formal ethical approval is not required.

2.3 Eligibility criteria

2.3.1 Type of studies

All RCTs of non-pharmacological interventions for PPD will be included in this study, regardless of blinding. The written language is limited in English or Chinese.

2.3.2 Type of participants

Participants were primiparous or multiparous postnatal women diagnosed with postpartum depression through a systematical clinical interview or a validated depression scoring system.

2.3.3 Type of interventions and comparators

We will include the RCTs evaluating the effectiveness of non-pharmacological treatments in treating PPD. Psychological/psychosocial interventions (including cognitive behavioral therapy (CBT) and IPT), physical therapies (including rTMS and light therapy), kinesitherapies (including aerobic exercise and yoga), music therapy, and acupuncture (including traditional acupuncture, or electro-acupuncture, or auricular acupuncture) will be involved in our study.

There are no limitations on the administration methods and duration of treatments. Studies combining non-pharmacological treatments and pharmacological treatments will be excluded, while studies with concomitant use of different non-pharmacological interventions will be included.

The control group will include usual care, waiting-list control, non-treatment control, and placebo, which will be regarded as a single node in network meta-analysis.

2.3.4 Outcome measures

The main outcome is the change in the depression score from baseline to endpoint, measured by validated depression scoring systems, such as Geriatric Depression Scale (GDS), Hamilton Depression Rating Scale (HAMD), Beck Depression Inventory (BDI), Self-Rating Depression Scale (SDS), Center for Epidemiologic Studies Depression Scale (CES-D), Edinburg Postpartum Depression Scale (EPDS), and Patient Health Questionnaire-9 (PHQ-9).

In order to examine the possible maintenance effects of non-pharmacological treatments, effect at the end of follow-up will also be evaluated as the secondary outcome.

2.3.5 Exclusion criteria

We will exclude the researches with quasi-random allocation or the studies without available data. Duplicate publication will also be excluded in the current study.

2.4 Search strategy

A comprehensive search of the following electronic bibliographic database resources will be attempted for the identification of English data:

  • PubMed
  • Scopus
  • The Cochrane Library
  • Web of Science

The following Chinese databases will also be searched for the identification of Chinese data

  • China National Knowledge Infrastructure (CNKI)
  • VIP Database
  • Wanfang Database
  • Chinese Biomedical Literature Database

Both English and Chinese databases were searched from their inception to March 15, 2020. In addition, we will also search the potential studies in reference lists of eligible studies, and the ongoing eligible trials in international trial registry websites.

The preliminary search strategy of PubMed is shown in Table 1, which will be adjusted in accordance with specific database.

Table 1
Table 1:
PubMed search strategy draft.

2.5 Study selection and data extraction

All databases mentioned above will be scanned using the pre-designed search strategy, and all retrieved articles will be imported into Endnote X7 (Thomson Reuters LLC, Philadelphia, USA). The duplicate reports will be deleted initially. Then 2 independent reviewers (YW and HL) will review the titles and abstracts of studies to select available RCTs in line with the eligibility criteria. Afterward, the full-text will be downloaded for further assessment. Finally, 2 reviewers (YW and HL) will independently extract data using a standardized extraction form, which includes:

  • (1) general information (e.g., authors, journal, year of publication, country, methods of randomization and blinding);
  • (2) participants (e.g., number, age, and severity of postpartum depression);
  • (3) interventions and controls (e.g., type of therapy, follow up period and control group);
  • (4) outcomes (the primary and additional outcomes).

Any disagreements will be resolved through discussion or consultation with a third reviewer (WP). If there is any missing data, we will try to contact the original author by email or telephone to obtain the complete data.

2.6 Assessment of risk of bias

Two independent reviewers (YC and MQ) will assess the risk of bias of included studies according to the Cochrane Handbook,[21] which includes random sequence generation, allocation concealment, blinding of participants, personnel and outcome assessors, incomplete outcome data, selective reporting, and other sources of bias. We will classify each domain into 3 levels: high, unclear or low risk of bias. Any variation in opinion between reviewers will be resolved by discussion or consultation with a third reviewer (WP).

2.7 Data synthesis and statistical methods

2.7.1 Pairwise and network meta-analysis

We will conduct the conventional pairwise meta-analysis for direct comparisons. The effect size of continuous variable data will be calculated with standardized mean difference (SMD), while the effect size of categorical variable data with risk ratio (RR). The related 95% confidence intervals (CIs) of SMD and RR will also be computed.

For indirect comparisons, a network meta-analysis is required to mix results to improve statistical efficacy. The network meta-analysis will be performed with the “netmeta”package of R software (R Foundation for Statistical Computing, Vienna, Austria), which based on frequentist method.[22] The direct and indirect comparison results will be presented as a network diagram.

2.7.2 Assessment of heterogeneity

Before the combination of effect size, the homogeneity will be analyzed by I2 statistic to check whether the results of individual studies are mergeable. When I2 ≤ 50%, the fixed effect model will be used to combine the effect size. Conversely, the random effect model will be used for meta-analysis after excluding the influence of apparent clinical heterogeneity.

2.7.3 Subgroup and sensitivity analysis

The subgroup analysis will be conducted to resolve the potential high heterogeneity across RCTs.

We will conduct sensitivity analysis to verify the robustness of the results. The research with unclear or high risk of bias will be exclude to check if the results would change.

2.7.4 Assessment of inconsistency

Consistency is a basic principle in network meta-analysis, which indicates the consistency between direct comparative result and indirect comparative results. As a loop is established among interventions, inconsistency will be evaluated by Z test.[23] The Z-value and its corresponding P-value will be calculated, and the inconsistencies exist between the different comparisons if the P-value < .05.

2.7.5 Publication bias

We will also evaluate potential publication bias using funnel plots and Egger test.[24]

2.8 Grading the quality of evidence

The evidence quality will be evaluated by 2 independent reviewers (JW and QH) using the Grading of Recommendations Assessment, Development and Evaluation (GRADE),[25] which classifies the quality of evidence as high, medium, low, and very low.

3 Discussion

The designation of this network meta-analysis protocol referred to the previous work.[5,26] Accumulating studies have provided options of using non-pharmacological interventions in PPD clinical practice. However, the best decision of non-pharmacological approaches is hard to make for physicians to treat woman with PPD, due to the lack of strong evidence base. The current clinical studies of verifying the efficacy of different non-pharmacological methods are rarely compared between each other directly. The network meta-analysis provides us a way to compare, rank, interpret the indirect evidence base of multiple non-drug therapies. We hope the results of the present study will help the clinicians to prescript a suitable single or mixed non-pharmacological treatment for women with PPD.[27]

Author contributions

Conceptualization: Yang Wang, Yunlu Liu, Tianmin Zhu.

Data curation: Yang Wang, Hui Li.

Formal analysis: Yalin Chen, Mimi Qiu.

Funding acquisition: Tianmin Zhu, Yunlu Liu.

Investigation: Yang Wang, Hui Li, Yalin Chen, Mimi Qiu, Jun Wang, Qinghong Hao, Yang Tu.

Methodology: Yang Wang, Wei Peng.

Supervision: Tianmin Zhu.

Writing – original draft: Yang Wang, Yunlu Liu.

Writing – review & editing: Yunlu Liu, Tianmin Zhu.


[1]. Field T. Postpartum depression effects on early interactions, parenting, and safety practices: a review. Infant Behav Dev 2010;33:1–6.
[2]. Glasser S, Levinson D, Gordon ES, et al. The tip of the iceberg: postpartum suicidality in Israel. Isr J Health Policy Res 2018;7:34.
[3]. Cox EQ, Killenberg S, Frische R, et al. Repetitive transcranial magnetic stimulation for the treatment of postpartum depression. J Affect Disord 2020;264:193–200.
[4]. Hahn-Holbrook J, Cornwell-Hinrichs T, Anaya I. Economic and health predictors of national postpartum depression prevalence: a systematic review, meta-analysis, and meta-regression of 291 studies from 56 countries. Front Psychiatry 2017;8:248.
[5]. Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis. Lancet 2018;391:1357–66.
[6]. Thomson M, Sharma V. Therapeutics of postpartum depression. Expert Rev Neurother 2017;17:495–507.
[7]. Chabrol H, Teissedre F, Armitage J, et al. Acceptability of psychotherapy and antidepressants for postnatal depression among newly delivered mothers. J Reprod Infant Psychol 2004;22:5–12.
[8]. Rihmer Z, Gonda X. Antidepressant-resistant depression and antidepressant-associated suicidal behaviour: the role of underlying bipolarity. Depress Res Treat 2011;2011:906462.
[9]. Dimidjian S, Goodman S. Nonpharmacologic intervention and prevention strategies for depression during pregnancy and the postpartum. Clin Obstet Gynecol 2009;52:498–515.
[10]. O’Hara MW, Stuart S, Gorman LL, et al. Efficacy of interpersonal psychotherapy for postpartum depression. Arch Gen Psychiatry 2000;57:1039–45.
[11]. Sockol LE. A systematic review of the efficacy of cognitive behavioral therapy for treating and preventing perinatal depression. J Affect Disord 2015;177:7–21.
[12]. Kim DR, Epperson N, Pare E, et al. An open label pilot study of transcranial magnetic stimulation for pregnant women with major depressive disorder. J Womens Health (Larchmt) 2011;20:255–61.
[13]. Swanson LM, Burgess HJ, Zollars J, et al. An open-label pilot study of a home wearable light therapy device for postpartum depression. Arch Womens Ment Health 2018;21:583–6.
[14]. Pritchett RV, Daley AJ, Jolly K. Does aerobic exercise reduce postpartum depressive symptoms? A systematic review and meta-analysis. Br J Gen Pract 2017;67:e684–91.
[15]. Buttner MM, Brock RL, O’Hara MW, et al. Efficacy of yoga for depressed postpartum women: a randomized controlled trial. Complement Ther Clin Pract 2015;21:94–100.
[16]. Daisy F, Rosie P. Could listening to music during pregnancy be protective against postnatal depression and poor wellbeing post birth? Longitudinal associations from a preliminary prospective cohort study. BMJ Open 2018;8:e021251.
[17]. Li S, Zhong W, Peng W, et al. Effectiveness of acupuncture in postpartum depression: a systematic review and meta-analysis. Acupunct Med 2018;36:295–301.
[18]. Higgins JP, Welton NJ. Network meta-analysis: a norm for comparative effectiveness. Lancet 2015;386:628–30.
[19]. Shamseer L, Moher D, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation. BMJ 2015;350:g7647.
[20]. Hutton B, Salanti G, Caldwell DM, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. Ann Intern Med 2015;162:777–84.
[21]. Higgins J, Thomas J, Chandler J, (eds). Cochrane Handbook for Systematic Reviews of Interventions version 6.0. The Cochrane Library.
[22]. Zhang C, Geng PL, Guo Y, et al. Application of netmeta package in R language to implement network meta-analysis. Chin J Evid Based Med 2014;14:625–30.
[23]. Higgins JP, Jackson D, Barrett JK, et al. Consistency and inconsistency in network meta-analysis: concepts and models for multi-arm studies. Res Synth Methods 2012;3:98–110.
[24]. Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005;58:882–93.
[25]. Schünemann H, Brożek J, Guyatt G, (eds). GRADE handbook for grading the quality of evidence and the strength of recommendations using the GRADE approach. The GRADE Working Group.
[26]. Furukawa TA, Salanti G, Atkinson LZ, et al. Comparative efficacy and acceptability of first-generation and second-generation antidepressants in the acute treatment of major depression: protocol for a network meta-analysis. BMJ Open 2016;6:e010919.
[27]. Salanti G. Indirect and mixed-treatment comparison, network, or multiple-treatments meta-analysis: many names, many benefits, many concerns for the next generation evidence synthesis tool. Res Synth Methods 2012;3:80–97.

network meta-analysis; non-pharmacological interventions; postpartum depression; protocol

Copyright © 2020 the Author(s). Published by Wolters Kluwer Health, Inc.