School-based self-management interventions for asthma among primary school children: a systematic review

A Cochrane review of school-based asthma interventions (combining all ages) found improved health outcomes. Self-management skills, however, vary according to age. We assessed effectiveness of primary school-based self-management interventions and identified components associated with successful programmes in children aged 6–12 years. We updated the Cochrane search (March 2020) and included the Global Health database. Two reviewers screened, assessed risk-of-bias and extracted data. We included 23 studies (10,682 participants); four at low risk-of-bias. Twelve studies reported at least one positive result for an outcome of interest. All 12 positive studies reported parental involvement in the intervention, compared to two-thirds of ineffective studies. In 10 of the 12 positive studies, parental involvement was substantial (e.g. attending sessions; phone/video communication) rather than being provided with written information. School-based self-management intervention can improve health outcomes and substantial parental involvement in school-based programmes seemed important for positive outcomes among primary school children.


INTRODUCTION
Asthma, the commonest long-term condition among children, causes significant morbidity and mortality globally 1 . Asthma guidelines recommend supported self-management to improve asthma control and reduce the use of urgent healthcare services [2][3][4] . Supported self-management, which includes discussion about selfmanagement and provision of a personalised asthma action plan supported by regular asthma review, can be delivered effectively in diverse cultural and demographic groups 5,6 . School-based asthma self-management interventions have been reported to improve asthma control and reduce school absenteeism and asthma exacerbations [7][8][9][10][11] . However, most systematic reviews analysed combined data from primary and secondary schools (5-18 years) [7][8][9][10] . One scoping review conducted in 2014 focused on primary school children, but the aim was to identify research gaps rather than assess outcomes 11 . The Cochrane review (Harris, 2019) used meta-analyses to assess intervention effectiveness and qualitative comparative analysis to examine the components of successful implementations 7 . The authors identified a number of components as being important: theoretical underpinning, parental involvement, child satisfaction and conducting the intervention during lesson time. However, the Cochrane review included interventions directed at children and adolescents (5-18 years), and did not distinguish the components associated with effective interventions in primary school children, which may differ from adolescents 7 . Educational intervention needs to be age-appropriate as primary school children will have less autonomy and capability to self-manage asthma compared to adolescents 12 . Thus, we aimed to review the effectiveness of school-based self-management interventions for primary school children with asthma and to examine the components associated with successful programmes. Figure 1 illustrates the article selection process using the PRISMA diagram. We included 23 studies; 16 studies from the Cochrane review [13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] , five studies from the updated database search [29][30][31][32][33] and two studies from the pre-publication update 34,35 . The total number of participants was 10,682. Some studies did not report numbers in each group so we cannot provide number by allocation 13,14,24 . We contacted all authors for information not reported in the papers, and nine (39%) responded 13,14,25,27,29,30,[32][33][34] .

Effectiveness of interventions
The effect of the interventions on each outcome of interest is detailed in Supplementary Table 1, with an explanation of how the direction of the effect was interpreted and the overall effect of the study assessed. Twelve studies (two at low risk-ofbias) were assessed as having an overall positive (beneficial) effect [13][14][15][16][17]21,22,29,30,[33][34][35] and eleven studies (two at low risk-ofbias) as having no effect [18][19][20][23][24][25][26][27][28]31,32 . No study was categorised as harmful or mixed effect. The Harvest plot (Fig. 2) illustrates the effect of varying degrees of parental involvement on school absenteeism, asthma control and urgent healthcare use.
Association of CFIR sub-domains and effectiveness Tables 1 and 2 are summary matrices comparing use of the 12 CFIR sub-domains in studies with overall positive or no effect (See Supplementary Table 4 for more detail). The number of CFIR subdomains used varied widely (2 to 12) and was similar in the studies with positive/no effect.

DISCUSSION
We identified 23 studies (four at low RoB) that evaluated the effectiveness of school-based asthma self-management intervention among primary school children. Twelve of the studies were categorised as being overall positive, though individual outcomes varied; no study reported overall negative impact. The number of CFIR sub-domains addressed varied between studies, but the only component that seemed to be associated with positive outcomes was substantial parental involvement. This was particularly apparent in studies at low RoB.
We found substantial parental involvement to be a crucial component of a school-based asthma self-management intervention among primary school children. Reviews that included interventions targeted at teenagers, in whom parental influence might be expected to be less important, have reached similar conclusions 7,8 . Parental involvement was also found to be important in other school-based interventions for obesity prevention studies 36,37 , self-management of mental health/disorders 38 , and academic enhancement 39 .
However, we did not find other components of interventions (theory-driven, conducted during lesson time, and child satisfaction) to be essential for successful intervention, as was found in the Cochrane review 7 . The differences in the findings were most probably due to a difference in the age group of the children as the Cochrane review included studies among older school children. Our review defined fun, interactive delivery of intervention, as a strategy promoting child satisfaction and engagement, whereas the Cochrane review examined measurement of child satisfaction, an evaluation used mainly in studies targeting adolescents 7 . Primary school children had good participation rates when the sessions were conducted during school hours including during recess, in contrast to adolescents who were less willing to devote their free time including during recess 7,13,16 . Social cognitive and Orem self-care theories were the most used   theories, adapted from adults which focuses on self-efficacy and skills of individuals [40][41][42] . These theories may be suitable for interventions targeting parents and adolescents, but may not be age-appropriate for primary school children with limited decisionmaking abilities and independent self-management skills 12,43 .
Primary school years are a critical time for children as they spend increasing time away from their parents and begin to learn asthma self-management for themselves 44,45 . Six-year-old children can express opinions, typically reflecting their parents' actions and views 44 . Over primary school years, they learn from their own experiences and gain the confidence to make independently decisions 44,45 . Although involving parents to support and empower their children's self-management behaviour is a key concept in the clinical management of children 2,3 , direct parental involvement was not always included in school-based intervention among primary school children 27,28,32 . A key challenge for involving parents is the difficulty of engaging them to attend session(s) delivered in school 24,46 . With the ease of modern telecommunication, alternative methods of engagement such as the use of telephone calls or video sessions could be explored as a convenient alternative to enable substantial parental involvement in the intervention 31,47 .
Although parental involvement is important, an aim of a schoolbased intervention is to shift the focus of self-management education from parents to children 48,49 . Studies in this review included up to eight educational sessions for children compared to only one to two sessions for parents 13,27,30 . A recent schoolbased health intervention has recommended the socio-ecological theory where children are the primary focus of an intervention that also involves the children's social network, e.g. parents, teachers, friends and the school plan/policy 48,50 . Schools could be an ideal setting for this approach, smoothing children's transition to independent self-management by being located in the child's environment and including parents as part of the children's social network [51][52][53] . Schools also provide a platform for interactive fun groups activities and peer support for children with similar conditions, which could reduce stigma and support selfmanagement practices 13,32 .
The effectiveness of self-management also depends on access and adherence to evidence-based treatments such as controller asthma medications, which is conventionally delivered in healthcare settings 2,5 . 'Access to healthcare', however, was a sub-domain least likely to be addressed in the studies included in this review. Although most US-based studies were conducted among minority deprived populations, in whom poor health outcomes may be due to the large disparities in healthcare provision 54 , only five studies reported the access of the children to effective controller medication 15,24,25,27,30 . Even in countries with universal health coverage, such as Canada and United Kingdom, equitable access to high quality healthcare for children cannot be assumed 55 . In low-and middle-income countries, socio-cultural beliefs, physical inaccessibility and lack of education and information are extremely common barriers to healthcare despite universal health coverage 56,57 . Similar barriers are widely described in the US 30,31,34 . Encouragingly, bridging school-based education with the children's healthcare providers has been a core component of recent school-based interventions 53,58 .
A strength of this review is that we used comprehensive search terms similar to the Cochrane review and searched seven relevant databases. Two reviewers conducted full text screening and data collection was duplicated. A pre-publication update was performed to ensure the findings was up to date this review. This review has some limitations. Despite a rigorous search strategy, it is possible that we may miss some studies. The screening of title and abstract was conducted by one reviewer, but good agreement resulted after training. Only two studies were conducted in low-and middle-income countries and many studies (15/23) were conducted in the US, reducing generalisability of the review. The included studies were variable in methodologies, instrumentation and data analysis. However, three low RoB studies coincided with the findings and some variability was illustrated in the Harvest plot with the other details described in Supplementary  Table 1. Poor reporting of interventions was a challenge and we may have overlooked some intervention components that were not explicitly described. We contacted all the authors to reduce the number of missing information and obtained 39% responses.
A multi-level intervention focusing on the children and involving their social network could provide a useful self-management interventions framework for primary school children and their parents. Specifically, there is a gap in our current understanding of school-based self-management education in younger children in low-and middle-income countries. Future research needs to focus on implementation strategies and effectiveness using this framework. Partnership between schools, parents and healthcare services could create a pragmatic and effective school plan/policy to improve asthma control among children.
School-based self-management interventions for asthma among primary education children can improve asthma outcomes and reduce absenteeism. Parental participation is an important component in this age group, but other features highlighted in secondary school interventions proved less relevant, perhaps reflecting the greater role of parents in younger children.

METHODS
This systematic review follows Cochrane methodology 59 , and PRISMA reporting standards. The protocol is registered with the PROSPERO database (registration number: CRD42019131955).

Study eligibility criteria
We used a Population, Intervention, Comparator/Control, Outcomes and Study Design (PICOS) strategy to define eligible studies (Table 3) 60 , using definitions similar to the Cochrane review 3,7,61 . Self-management intervention was defined as the active transfer of information to children with asthma to enhance their self-management skills; this was interpreted with reference to components of self-management recommended by global guidelines (Table 3) 2,3 . In line with the Cochrane review, we included nonrandomised trials to capture a broader range of studies and thence components used.

Outcomes of interest
We chose three outcomes of interest (school absenteeism and two health outcomes -asthma control and urgent use of healthcare services) to reflect the impact on children with poorly controlled asthma 2,7,61 .

Search strategy
The details of the search terms and databases used are in Supplementary  Table 5. The Cochrane review conducted searches in August 2017 using search terms developed by the Cochrane Airway Information Specialist in 23 electronic databases from 1995 onwards and included 55 papers 7 . Using the same search terms, with no language limitations, we updated the search in February 2019 in six-core databases (CENTRAL, MEDLINE, Embase, PsycINFO, CINAHL, AMED) 7 . In addition, we searched the Global Health database using similar search terms without date limits to include studies from low-and middle-income countries. We included all studies identified in the review that met our eligibility criteria (principally excluding those not delivered to primary school children). We checked the reference list and undertook forward citation of studies in the Cochrane review conducted among primary school children 62 .
A pre-publication update was conducted on 17th March 2020 using forward citation of the Cochrane review (published 28 January 2019) 7 and all the studies included in this review 62 .

Study selection and data extraction
We imported the list of articles from the electronic databases into Endnote software (version 7) to facilitate screening, de-duplication and overall management of the results. SNR and JS independently screened a random SN Ramdzan et al. selection of 10% of the titles and abstracts 5 . A 96.3% agreement was achieved prior to discussion, which reached total agreement after clarification of the screening criteria. SNR then completed title and abstract screening. Both reviewers independently conducted full-text screening (which included all the studies in the Cochrane review and those satisfying title and abstract screening), met to discuss discrepancies and decided on the final included papers. Supplementary Table 6 lists studies excluded from this review. A modified Cochrane data extraction form was used for duplicate data extraction (SNR and JS) 63 . SNR contacted authors for missing data by email and any further information received was added to the data extraction forms 59 .
At all stages, any discrepancies not resolved by discussion between the two reviewers were arbitrated by the study team (HP, KEM, LSM, SC).

Risk of bias of included studies
We used the Cochrane Effective Practice and Organisation of Care (EPOC) Risk of Bias (RoB) tool 64 to categorise risk into low, high and unclear risk in nine domains, which were then used to generate an overall assessment of the RoB for each study. The Cochrane EPOC RoB tool applies to randomised trials and non-randomised trials 64 . Studies with at least one high-risk domain were summarised as high risk; studies with no high-risk domains but at least one unclear domain were summarised as unclear risk and studies at low risk in all domains were summarised as low risk 64 .

Data handling
The Consolidated Framework for Implementation Research (CFIR) is a comprehensive framework that systematically identifies factors (subdomains) that influence the effectiveness of implementation in multilevel interventions 65 . Supplementary Table 3 outlines the 12 CFIR subdomains. We used CFIR sub-domains to identify context and components in each study (e.g., intervention characteristics, features of the setting and strategies for implementation) that might influence effectiveness of the interventions 66,67 .
We used a structured approach to divide the studies into four categories according to the change in the outcomes of interest 68 . This was a two-step process.
First, we determined the direction of effect in each of the three outcomes of interest (school absenteeism; asthma control; urgent use of healthcare service) for each included study. In some studies, several measures mapped to each outcome of interest: for example, emergency room visits and hospitalisation are both measures of unscheduled care potentially with conflicting findings. The rules at the top of Supplementary Table 1 define how we prioritised outcomes defined as 'primary' in the included study, outcomes measured with a validated instrument, and results that were clinically as well as statistically significant. The table then describes how the decision process was applied for each outcome of interest in each study.
Second, we categorised the overall effect of the intervention in each study as positive, negative, no effect or mixed effects, as follows:

Data synthesis
Our preliminary scoping suggested that the studies would be heterogenous in terms of context, components delivered and study design, so we undertook a narrative analysis. We used a Harvest plot 69 (coded to indicate number of participants, RoB and follow-up duration) to illustrate the effectiveness of the interventions on the three outcomes of interest for each study. A Harvest plot graphically displays not only outcomes but also the weight of the evidence in complex and diverse studies by illustrating selected methodological criteria 69 . We used a matrix to examine the association of the CIFR sub-domains with the overall effectiveness of the interventions. Supplementary Table 4 lists the CFIR sub-domains and how we interpreted them in our analysis.

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.

DATA AVAILABILITY
All data that support the findings of this systematic review are already in the public domain. Table 3. PICO study strategy and definition of terminology.

Participant/ population
Children with asthma aged 6-12 years

Intervention
School-based self-management education intervention.
Definition as active transfer of information to enhance self-management of asthma containing at least one of the corecomponents of self-management education 2,3 : • A basic explanation about asthma, triggers and the factors that influence control • Training about correct inhalation technique • Information on the importance of the child's adherence to the prescribed medication regimen • Written asthma action plan Children with asthma had to be the primary target for the intervention, though others (such as peers without asthma, parents, school staff) could also be included.
Comparator(s) Standard care or other (non-asthma, or not related to self-management or delayed intervention) education intervention or none Outcomes School absenteeism or/and asthma control or/and urgent use of healthcare service The definition of the three categories of outcomes of interest were guided by the American Thoracic Society/European Respiratory Society statement 61 : 1. School absenteeism: Number of days a participant was absent from school (priority due to asthma).
2. Asthma control: Clinical level of asthma control based on symptoms and capability to perform daily activities measured using asthma symptoms questionnaire/asthma diary with/without objective validation of asthma control, e.g. peak flows or lung function test.
3. Urgent use of healthcare service: Number of an unscheduled visit to a general practitioner and/or emergency department due to asthma, and the number of days of hospitalisation due to asthma.
Setting School (primary, elementary or middle school)