Background

Teachable moments are described as "naturally occurring events thought to motivate individuals to spontaneously adopt risk-reducing health behaviours".1^(p.156) Pregnancy itself is argued as a prime teachable moment, with women undergoing a life transition while in frequent contact with healthcare professionals.2 For many, this may be their first encounter with healthcare services. The National Institute for Health and Care Excellence guidance on behavior change emphasizes the importance of intervening at "key life stages or times", including pregnancy.3 An underlying assumption is that behaviors change due to women both prioritizing fetal health and also responding to social norms on the acceptability of certain behaviors in pregnancy. A major question is whether or not there are similarities or differences across health behaviors in women's propensity for positive change.

Behaviors and conditions associated with a greater risk to the mother and fetus include maternal obesity and weight, smoking and alcohol use. These behaviors form the focus of this umbrella review, enabling comparison of a behavior/condition that is gaining in prevalence (maternal obesity), a behavior generally in decline (smoking) and a behavior where safe levels are contested (alcohol use). Maternal obesity is increasing in prevalence, carries associated health risks to mother and infant (including gestational diabetes, pre-eclampsia and preterm birth), and has implications for healthcare resources.4-7 Prevalence of smoking among pregnant women in high-income countries is decreasing; however, a different picture is evident among socially disadvantaged women and among women in the developing world.8-11 Health risks of smoking during pregnancy include spontaneous abortions, ectopic pregnancies, placenta previa and infants who are small for gestation age.12-14 Prevalence of alcohol use in pregnancy varies between countries and is associated with preterm birth, low birth weight and fetal alcohol spectrum disorders.10,15-18

National and international guidelines exist for weight management19-21 and smoking cessation.22-26 Guidance on alcohol consumption is variable, with advice ranging from abstinence to light consumption.23,27 However, the extent to which changed behaviors are adopted among pregnant women is unclear.28 Research into smoking during pregnancy shows a number of "spontaneous quitters", though many do not continue to abstain beyond pregnancy, arguably due to a newly defined self-concept and perceived social stigmatization.1,29 Weight management research shows some women make dietary changes due to prioritizing the health of their unborn baby.30 Others report pregnant women showing vigilance in making changes to a range of health behaviors under pressure from societal norms.31

There is clear emphasis on pregnant women as a target for intervention across different behavior types. Preliminary searching was undertaken for reviews addressing weight management, smoking cessation and alcohol reduction during pregnancy simultaneously to enable a "look across" effective interventions. Searches were carried out using PROSPERO, Epistemonikos, the JBI Database of Systematic Reviews and Implementation Reports, JBI registered titles and with Google Scholar. Retrieved reviews mainly focused on behavior specific interventions in pregnancy, with no attempt to synthesize across behaviors. Reviews considering multiple target behaviors focused on school children32 or on internet-based interventions among a general population.33 One review focused on the general population, with the inclusion of pregnant women for certain behaviors, yet little depth was given to this aspect.34 Another recent review considered the effectiveness of a range of lifestyle interventions among pre-conceptual women. A relatively short list of interventions with substantial evidence for effectiveness was reported.35 Perhaps most closely related to the focus of this umbrella review were two systematic reviews of reviews in the United Kingdom, which looked specifically at midwifery-based interventions.36,37 Alderdice et al.36 focused on interventions to improve maternal mental health and wellbeing, finding a lack of conclusive evidence around the role of midwives in maternal mental health interventions. McNeill et al.37 carried out a wider review of reviews considering a broad range of midwifery-based public health interventions, including (but not limited to) smoking cessation and weight management interventions. The review highlighted that a number of reviews presented statistically significant positive findings. However, there were also noted gaps in getting evidence into practice. These reviews differ from this proposal in that their focus was solely on midwifery-based interventions and their target behaviors were different. Furthermore, the searches for the McNeill et al.37 review of reviews were last carried out in 2010, and there is likely to be updated evidence available since then.

The increasing focus of multiple interventions and the development of a plethora of referral systems, pathways and guidelines can ultimately present a significant burden to women, healthcare professionals and services trying to manage complex pregnancies. An umbrella review focusing on effective interventions across these three different behavior types will enable enhanced understanding of mechanisms and which interventions are effective and for whom, as well as identification of similarities or differences across different target behaviors. Such information will be of value for the development and delivery of cost-effective interventions that have the potential to impact on short- and long-term health outcomes for women and children.

Inclusion criteria

Types of participants

The current umbrella review will consider existing research syntheses/reviews that include pregnant women at any gestational age. There will be no restrictions based on socio-demographic factors; such as age, ethnicity, parity, socio-economic status or disease status (e.g. gestational diabetes). The umbrella review will not include studies specifically focusing on preconception or postnatal periods.

Types of intervention(s)/phenomena of interest

The current umbrella review will consider research syntheses that evaluate interventions during pregnancy that are in one of the following public health areas of interest: weight management interventions, smoking cessation interventions and alcohol reduction interventions. In the case of weight management interventions, this will include interventions aiming to improve weight management-related behaviors (i.e. diet and physical activity) even if they do not specifically target weight management as an outcome.

Interventions that are targeted (e.g. to overweight women), as well as those delivered universally to all pregnant women will be included to provide as broad an overview as possible.

Outcomes

This review will explore the effectiveness of interventions in changing behaviors as the broad specified outcome. This will encompass both "explicit" behavior change outcomes and "proxy" behavior change outcomes, relating to the three target behaviors of the interventions. Examples of behavior change (explicit) and behavior change (proxy) outcomes for each target behavior are provided below. Please note that due to the broad nature of this review, this is not an exhaustive list, but is provided for illustrative purposes.

Target behavior of intervention - weight management (including diet and physical activity interventions):

* Behavior change (explicit): change in diet, change in physical activity levels and others.

* Behavior change (proxy).Determinants of behavior: intentions to change, self-efficacy and others.Health outcomes: gestational weight gain, gestational diabetes and others.

Target behavior of intervention - smoking cessation:

* Behavior change (explicit): quit rates, smoking reduction rates, quantity of cigarettes and others (self-report or biomedically validated that is urinary analysis or carbon monoxide levels).

* Behavior change (proxy).Determinants of behavior: intentions to change, self-efficacy, stage of change and others.Health outcomes: birth weight, preterm delivery and others.

Target behavior of intervention - alcohol reduction:

* Behavior change (explicit): abstinence, change in units of alcohol consumed and others.

* Behavior change (proxy).Determinants of behavior: intentions to change, self-efficacy and others.Health outcomes: birth weight, fetal alcohol syndrome and others.

The effectiveness data reported in the systematic reviews and included in the umbrella review will be grouped and summarized using a number of a priori decisions:

* A comparison of the effectiveness of all interventions in pregnancy will be carried out (i.e. not target behavior specific), with subgroup comparison of explicit and proxy measures of behavior change for all types of intervention.

* A comparison of the effectiveness of target behavior interventions will be carried out (i.e. weight management, smoking cessation and alcohol reduction), with subgroup comparison of explicit and proxy measures of behavior change for each target behavior.

* Where enough data are provided in the systematic reviews included in the umbrella review, further comparisons will be carried out exploring the effectiveness of interventions according to predefined subgroups listed below:Whether or not behavior change theory was explicitly used as the basis for the interventionMethods of intervention delivery (e.g. internet-based, peer-group delivered, individual patient/health professional, intensity of intervention etc.)Description of the intervention (e.g. educational, psychological, information provision, motivational interviewing etc.).

Types of studies

The review will aim to identify relevant international scientific evidence synthesized in systematic reviews that has been summarized narratively or via meta-analyses. The umbrella review will not include primary research studies or reviews where the primary sources of evidence are theoretical studies, qualitative data or opinion. Mixed methods reviews will not be excluded; however, due to the focus of this umbrella review on effectiveness, only quantitative elements of mixed methods reviews will be included.

Search strategy

The search strategy will aim to find both published and unpublished studies. A three-step search strategy will be utilized in this review. An initial limited search of the review databases - the Cochrane Library and the Database of Abstracts of Reviews of Effectiveness (DARE) - will been undertaken followed by analysis of the text words contained in the title and abstract, and of the index terms used to describe the article.

Initial keywords to be used will be:

Pregnancy OR Pregnant OR Maternal OR Maternity OR Antenatal

AND

Weight management OR Obesity, OR Body Mass Index OR Diet OR Physical Activity OR

Smoke OR, Smoking OR, Tobacco OR, Cigarette OR, Alcohol OR Drinking

The search strategy will be developed including identified keywords and index terms which will be tailored for each resource including key terms "systematic" and "meta-analysis". Where filters for review articles exist within databases, these will be applied.

A second search using all identified keywords and index terms will then be undertaken across several databases (refer to list below). Third, the reference list of all studies selected for critical appraisal will be screened for additional studies. Fourth, citation records of these eligible reviews will also be considered. Only reviews published in English will be considered for inclusion in this review. Reviews published since 2005 until the present date will be considered for inclusion in this umbrella review. This date has been selected following JBI guidance that research syntheses38 published in the last 10 years will yield primary research conducted 30+ years prior.

Where grey literature is identified, this does not have to be solely in published reviews: for example, reviews within government reports or student dissertations that are relevant will be included. If conference abstracts are identified, an attempt will be made to contact the authors for any peer reviewed papers or unpublished full reports due to limited information within the abstract itself.

The major repositories of systematic reviews to be searched include:

JBI Database of Systematic Reviews and Implementation Reports

The Cochrane Library (including Cochrane Database of Systematic Reviews)

DARE

PROSPERO

Epistemonikos

The following databases will also be searched:

EBSCOhost Medline

Ovid EMBASE

EBSCOhost CINAHL

EBSCOhost AMED

ProQuest ASSIA

Ovid Maternity and Infant Care

EBSCOhost PsycINFO, Science Direct

LILACS (http://lilacs.bvsalud.org/en/)

Social Care Online (http://www.scie-socialcareonline.org.uk/)

The search for unpublished studies will include:

Google Scholar

NICE Evidence Search (http://www.evidence.nhs.uk)

OpenGrey (http://www.opengrey.eu/)

The Grey Literature Report (http://www.greylit.org)

National Institute for Health Research (NIHR) Journals library

Health Technology Assessment Database (produced by the NIHR Centre for Reviews and Dissemination)

Ovid Health Management Information Centre Database (HMIC)

Cochrane Pregnancy and Childbirth Group (http://pregnancy.cochrane.org/)

All retrieved papers will be assessed for relevance using the title, abstract and index terms. Screening will be conducted by two reviewers. One reviewer will screen all of the abstracts for relevance. The other reviewer will screen only the excluded abstracts to ensure nothing of relevance has been missed. If relevance is unclear from the abstract, the full text will be accessed.

Reviews selected for retrieval will be then be appraised using predefined inclusion criteria developed for this review (Appendix I). This will ensure that only studies that meet the inclusion criteria of this umbrella review will be included. This will be carried out by two reviewers independently.

Assessment of methodological quality

Retrieved research syntheses that meet the inclusion criteria will then be critically appraised by two independent reviewers prior to inclusion in the review using standardized JBI critical appraisal instruments for Systematic Reviews and Research Syntheses and supporting guidance (Appendix II). The two reviewers will carry this process out for all included reviews. Reviewers will be blinded to each other's assessment, and assessments will be compared once they have both completed the appraisal. Any disagreements that arise between the reviewers will be resolved through discussion or with a third reviewer. Reviews will not be excluded based on the critical appraisal; however, the results of critical appraisal will be reported in narrative form and in relevant tables. This information will be used to critically consider the conclusions of included reviews and will inform the discussion of the umbrella review.

Data extraction

Data will be extracted from research syntheses included in the umbrella review using the JBI data extraction tool for Systematic Reviews and Research Syntheses (Appendix III) by two independent reviewers. All additions or modifications have been discussed in detail and submitted with the review protocol. In line with JBI guidance, a period of piloting has been carried out by three members of the review team (SD, NH and LA) to ensure consistency and the likelihood of relevant results being identified and detailed sufficiently. The review team are aware that these changes require approval for publication in the JBI Database of Systematic Reviews and Implementation Reports before being used by any reviewer.

The data extracted will be guided by the modified extraction tool and will include objectives of the review, type of review, participants, setting, scope of database searching, dates of the search, date range of included studies, the number/types of studies/country of origin of primary research studies in the included research synthesis, details of critical appraisal and quality assessment, method of synthesis/analysis, outcomes of significance to the umbrella review question and objectives and any comments or notes the umbrella review authors may have regarding included studies. Importantly, the RE-AIM framework (R: Reach, E: Effectiveness, A: Adoption, I: Implementation and M: Maintenance) will be used to guide data extraction and presentation of narrative information on the type of interventions that are most effective during pregnancy.39 This framework has previously been operationalized within an umbrella review of online behavior change interventions for multiple behaviors (including diet, physical activity, alcohol use, smoking and condom use).33 Within this pre-existing review, the final three constructs (AIM) were combined and defined as Use. This adaptation was due to these latter constructs of the framework primarily being intended for organizational level intervention (e.g. maternity services),39 therefore enabling a more useable definition at an individual level (e.g. pregnant women). It is anticipated that the data provided in the existing reviews will provide sufficient information on the RE constructs, with somewhat limited information on Use. However, this information will be extracted and considered within the umbrella review where available.

Reach

Data will be extracted and summarized to identify individual measures of participation and characteristics of participants as following:

* Total participant numbers: across all studies within the review, within intervention and control groups and in interventions considering behavior change explicit outcomes versus proxy measures of behavior change.

* Percentage of individuals who received an intervention compared to potential audience (participant rate).

* The recruitment strategy used and effectiveness of this strategy (e.g. recruitment and retention rates).

* Characteristics of participants and non-participants. In particular, considering the level of homogeneity/heterogeneity related to engagement with the intervention (e.g. socio-demographics of women recruited compared with those who decline participation, and for women retained compared with dropouts [maternal age, parity, socio-economic status and relationship status]).

* Barriers and facilitators to participation described (i.e. reasons for choosing/declining to participate or for dropout [such as time and cost]).

* Intervention delivered to a specific target group or universally to pregnant women.

Use

Information about Use will also be extracted from included systematic reviews according to the following topic areas:

* Fidelity to intervention delivery (extent to which program is delivered as intended/to which individuals engaged with the intervention as intended).

* Measures of adherence to the intervention.

* Dropout rates (considering both the intervention and control arms of study): number of withdrawals and characteristics of women retained compared with withdrawals (maternal age, parity, socio-economic status and relationship status).

* Maintenance: extent to which any behavior change is maintained in the long term (if studies included follow-up measures).

Data will be extracted via a custom built word table or an online survey tool (such as Bristol Online Survey [BOS] software, Bristol, UK). This will be available to all data extractors and will enable storage and collation of the data extracted.

Data summary

Due to the expected heterogeneity of the reviews, retrieved findings will be presented in a narrative form including tables and figures to aid in data presentation, where appropriate. A clear indication of any overlaps of original research studies in each of the included systematic reviews will be provided.

Tabular presentation of quantitative findings will be provided for interventions addressing each of the three target behaviors. For example, one table will summarize interventions addressing weight management, one will summarize interventions addressing smoking cessation and one will summarize interventions addressing alcohol consumption. In this instance, this will include Reach (e.g. participation rate, characteristics of participants informing this target behavior), Effectiveness on explicit outcomes and Effectiveness on proxy outcomes (e.g. overall effect estimates/other presented numerical data) and Use (e.g. loss to follow-up/participant retention). Heterogeneity of the results of the included reviews and limitations and recommendations reported in the reviews will also be included where possible. Data may be presented by subgroups where possible or where this is most intuitive, for example, weight management interventions may be presented by interventions that target physical activity, interventions that target dietary behavior and others.

A narrative summary will consider each of the target behaviors for intervention in turn. This will summarize the interventions' Reach and Use (where data exist) and then consider their effectiveness using explicit and proxy outcomes as two subgroups for analysis, with a further comparison of the similarities or differences between the two subgroups. For example, within weight management, the narrative summary will aim to consider:

* How effective are interventions that target/measure explicit behavior change measures (i.e. diet and physical activity levels)?

* How effective are the interventions that target/measure the proxy outcomes (i.e. intentions to change, pregnancy outcomes etc.)?

* Are there any similarities or differences between the two subgroups (explicit and proxy)?

These stages will be repeated for smoking cessation interventions and for alcohol reduction interventions.

Following this, the narrative will look across the three target behaviors of the interventions, providing a summary of key findings and a comparison of the similarities or differences across the reviews of the three different target behaviors. This will summarize Reach and Use of interventions and then consider effectiveness of interventions considering explicit outcomes and proxy outcomes. Following this, any similarities or differences between the two subgroups (explicit and proxy) would be considered.

A "Summary of Evidence" table will be produced to provide a simple visual indication of the findings of the review. This will include the target behavior of the intervention, a description of the intervention, the included research synthesis and a simple visual indication of the results which will be done via a "stop-light" indicator. A beneficial or effective intervention will be labeled green, a no-effect or no-difference in the investigated comparison (amber) and a detrimental intervention or less effect than a comparator (red).

Finally, the review team aims to involve a lay participant (e.g. a recent mother), a clinical expert (e.g. a public health midwife) and relevant stakeholder groups/government organizations (e.g. Public Health England) who will be provided with the progress, give feedback, discuss issues aroused during the systematic reviews and guide the translation of the research results into policy and practice.

Assessing confidence

The GRADE approach for assessing confidence in the quality of evidence will be used for this review to form an overall assessment of the quality of evidence for each intervention or phenomena of interest with the results presented in a summary of findings table created using GRADEPro.

Acknowledgements

The review team would like to acknowledge the input and support of Dr Sharon Hamilton, Director of Teesside Centre for Evidence Informed Practice: a Joanna Briggs Institute Centre of Excellence, and Caroline De Brun, the Centre's Information Scientist.

Appendix I: Screening tool based on the inclusion criteria of this umbrella review

Appendix II: Appraisal instrument

Appendix III: Data extraction instrument

References