Search strategy and selection criteria
Rapid evidence assessment (REA) methodology was utilised to systematically review the literature for each of the four outcome areas. The REA approach applies rigorous methods for locating, appraising and synthesising the evidence to provide structure, balance and transparency of a practice, but the methodology places restrictions in search criteria due to the breadth of evidence . We searched the following electronic databases with a limited date range of January 2006 to March 2016:
- Cochrane Central Register of Controlled Trials (searched DATE)
- Medline (searched DATE)
- PsycINFO (searched DATE)
- CINAHL (searched DATE)
- PubMed (searched DATE)
Grey literature with a priority focus on reports from government agencies, and quality reports from reputable stakeholders fitting the review scope were also searched. International literature, in English only, that focused on research from high-income countries, populations and settings was included. Books and book chapters were excluded. An individual search strategy was performed for each outcome area rather than a single over-arching search strategy across all four areas. This gave a better reflection of the flow of studies for each topic at each stage of screening for eligibility. The search was updated in August 2019. The search criteria for each of the topics are included in Additional file 1.
Eligibility criteria (PICOS format)
Interventions delivered to parent(s) and/or children during the first two weeks to five years of the child’s life were eligible. The minimum child age was set to two weeks to exclude interventions delivered in the first few days after childbirth when the parent/child is potentially still under hospital care. However, studies recruiting in hospitals within two weeks of birth were eligible. The limit was set to 5 years to ensure interventions were offered primarily to pre-schoolers, in keeping with the evidence that the early years are central for future development.
Studies focusing exclusively on the following populations were not eligible as they were delivered in completely different health care settings:
- Parent(s) and/or children from low-income countries, populations and settings
- Parent(s) and/or children with a clinical diagnosis of an emotional, behavioural or conduct disorder (e.g. anxiety disorder, ADHD)
- Parent(s) and/or children with specific disabilities, illnesses or comorbidities (e.g. diabetes, cancer)
- Unique environmental circumstances (e.g. refugee, war torn/disaster zone, military families, homeless)
The current review aimed to identify universal and targeted interventions (selective and indicated) as defined by the United States Institute of Medicine . Studies were excluded if it could not be determined whether the intervention was universal or targeted. Tertiary interventions (e.g. interventions that reduce disability, enhance rehabilitation and prevent relapses and recurrences of the illness) and/or interventions delivered in a tertiary setting were not eligible. Brief interventions were defined as interventions that did not exceed 4 sessions in keeping with previous systematic review literature [18, 19]. Interventions that included 4 or less sessions but stipulated that parents follow a specific regime outside of the sessions were excluded as (i) intervention fidelity may vary dramatically within participant groups, and (ii) ability to adhere to a schedule may impact parental confidence. These tight inclusion criteria ensured that eligible studies could be adapted for delivery within existing universal child health service structures where only a handful of visits are achievable . No restrictions were placed on the length of time of the intervention sessions.
Interventions delivered by any healthcare practitioner, family member or peer were eligible for inclusion, provided they were deliverable within a UCHS platform. For example, an intervention where clinical psychologists delivered cognitive-behavioural techniques within a tertiary setting would not be eligible but if the same psychologist delivered the same techniques as part of a well-child care program then the intervention could be considered eligible. Telephone-, digital- and internet-based and in-person interventions were all eligible for inclusion if they were delivered in a finite and structured format. Interventions that were not session-based and allowed continual access to support were excluded, for example, online forums where mothers could speak with peers or practitioners at their convenience. Interventions which involved screening but no structured, session-based response for women exceeding screening instrument thresholds were also excluded. Interventions delivered in any setting (e.g. home, community, healthcare) were eligible except for interventions targeting outcomes relating to ‘home learning environment’. Due to the confounding influence of nursery/pre-school/community groups in fostering similar outcomes and the focus of universal services being on the family unit, we stipulated an additional inclusion criterion that infants had to receive the interventions targeting home learning environment within their home. This permitted (i) interventions provided to parents outside of the home but to be delivered to the infant in the home and (ii) interventions delivered directly to the child by intervention provider (e.g. healthcare practitioner). Lastly, pharmacological interventions were not considered eligible due to their lack of suitably to a universal child services’ platform.
iii. Comparison groups
Studies with the following comparison groups were eligible:
- Usual care pathways, wait-list or no-intervention comparison control groups
- Leaflet-based information.
Follow up assessments where there was not an equivalent control group comparison would not be reported.
Given the wide and diverse number of family and child outcomes that UCHS could influence, a shortlist of priority areas was decided upon based on the application of a priority-setting matrix and consultation with experts from within the field . The methodology ensured that the areas chosen were of significant (1) prevalence, (2) impact on families, (3) international relevance, and (4) likelihood of being delivered via a UCHS platform in a brief format. The four priority areas selected were:
- Child social and emotional wellbeing (Child SEWB) – including bonding and attachment
- Infant sleep disorders
- Home learning environment
- Parental mental health
Children with low social and emotional wellbeing (SEWB) are at an increased risk of learning difficulties, academic underachievement, and mental health disorders [22, 23]. Infant sleep duration and quality can have lasting impact on a child’s behavioural, cognitive and physical development without early intervention , and increases the likelihood of postnatal depression in mothers from 10% to 45% . The home learning environment is a key determinant of child development. Children who grow up in a poor home learning environment with sub-optimal stimulation have lower levels of educational achievement when they leave school and lower employment levels in adulthood [26, 27]. One in five children has a parent with a mental health disorder . Poor parental mental health is known to increase the risk of social and behavioural problems in childhood and adolescence and increase the child’s risk of developing mental health problems as they get older [29, 30]. Definitions of the priority areas and examples of the outcomes that could be used to measure effects in these areas are presented in Table 1.
Interventions may have collected outcome data relating to several areas, but each intervention was categorised as focusing on a single outcome area according to the primary outcome or recruited population. The purpose of categorising interventions under a primary outcome area was to see whether uptake may have been influenced by the ‘offer’ of the intervention. For example, if an intervention invited families with infant sleep problems but measured sleep as a primary outcome and parental mental health as a secondary outcome, it was categorised as an infant sleep intervention.
v. Study design
Any study with a comparison group, including randomised controlled trials (RCTs) and quasi-randomised trials were eligible. All other trial designs without an established comparison group were excluded. Systematic reviews were excluded but were searched for relevant studies. Only studies with outcome data collected at least one month after intervention delivery were eligible.
Table 1. Definition of outcomes
Outcomes of interest
Child social and emotional wellbeing
Interventions designed to improve, promote and optimise child behavioural outcomes, positive social and/or emotional wellbeing and reduce mental illness in children.
- Externalizing behavioural problems (e.g. oppositional defiance, antisocial behaviour, and aggression)
- Internalising behaviour problems (e.g. anxiety, depression)
- Infant attachment behaviour
Infant sleep disorders
Behavioural and/or education interventions that aiming to preventing or improve sleep problems.
- Difficulties falling or staying asleep
- Excessive total sleep time
- Night waking
- Settling problems
Home learning environment
Interventions that aim to improve the home learning environment of children by promote positive intellectual and social development in the child.
- Any relevant cognitive areas (i.e. literacy, pre-literacy, numeracy, pre-numeracy, language and communication, and/or general cognitive functioning).
- Frequency of reading, attitudes towards reading
- Literacy scores
- Language ability
Parent mental health
Interventions that aim to (i) to prevent mental illness and promote positive mental health in parents or (ii) improve outcomes of existing mental health problems.
- Rates of diagnoses of mental health disorders (e.g. anxiety, depression)
- Self-report on mental health symptom scales (e.g. anxiety, depression)
Selection of studies
Data was managed using EPPI-Reviewer 4 software, which is EPPI-Centre’s comprehensive online software tool for research synthesis. Search results for each topic were filtered for duplicates and imported into EPPI-Reviewer 4 software for screening against inclusion/exclusion criteria based on title and abstract. Full-text versions of remaining eligible studies were retrieved and imported to EPPI-Reviewer 4, for full-text screening. Twenty percent of studies were also screened by a second reviewer at the full-text screening stage, to ensure consistency across the project. Consistency of 100% between reviewers was required before studies were accepted for inclusion, and discrepancies were resolved by discussion between reviewers to achieve this. Eligible studies remaining after this final screening were included for review and subject to data extraction.
Data extraction and analysis
Data from the individual studies were extracted in a consistent format using a form developed for this review. Information extracted for each intervention included details on:
- Approach (universal, selected, indicated)
- Content (what format did the intervention take and what were they targeting)
- Mode of delivery (e.g. telephone, in-person, internet)
- Intensity (number of sessions, length of sessions)
- Provider (who delivered the intervention to participants)
- Effectiveness (outcome data)
- Engagement (recruitment and attrition data)
- Adherence (to what extent did patients complete all the intervention components)
To determine the length of an intervention, the endpoint was defined as the final time participants received intervention content from the intervention provider. Intervention contacts solely for data collection or for following up on participants without new content were not classed as intervention sessions.
Self-report data and observer-reported outcome data (e.g. video-coded behaviour assessment) were extracted. Outcome data not relating to our four outcome areas were not extracted. Data from intention-to-treat analyses were used where reported. Due to variation in the wide range of outcome measures used (both in terms of the outcome areas and/or the instruments used to assess the outcomes), it was not possible to conduct a meta-analysis and results were reported using narrative synthesis of findings.
The National Institute for Clinical Excellence (NICE) quality appraisal checklist for quantitative studies was used to assess study quality (http://www.nice.org.uk/). This checklist considers the appropriateness of the theoretical approach, study design, data collection, trustworthiness, analysis, relevancy of the findings and ethics. Studies received one of the following three potential quality scores:
- ++ (Low risk of bias): All or most of the checklist criteria are fulfilled; where they have not been fulfilled, the conclusions are very unlikely to alter.
- + (Medium risk of bias): Some of the checklist criteria are fulfilled, where they have not been fulfilled, or not adequately described, the conclusions are unlikely to alter.
- - (High risk of bias): Few or no checklist criteria are fulfilled and the conclusions are likely or very likely to alter.
Studies were not excluded based on quality but this information was used to consider the conclusions of included studies, and for the interpretation when findings across studies differed. The quality appraisal was used for deciding which interventions may be most suitable for recommending as ‘best bet interventions’. Two trained researchers appraised the quality of each study.