A systematic review of school-based weight-related interventions in the Gulf Cooperation Council countries

A systematic search was conducted in November 2022 using the databases Medline, Scopus, and ProQuest, which were chosen due to their comprehensive coverage of medical and health sciences literature to retrieve all relevant peer-reviewed studies published in English. The search was conducted without date restrictions to capture the full extent of research conducted in this area, with the search being completed on 17 November 2022.

Combinations of Medical Subject Heading (MeSH) terms and keywords were used under five common concepts: (1) population [‘adolescence’ or ‘teen’ or ‘youth’ or ‘child’ or ‘student’], (2) setting [‘school’ or ‘school-based’], (3) interventions [‘intervention’ or ‘initiative’ or ‘program’ or ‘project’ or ‘physical’ or ‘exercise’ or ‘sedentary’ or ‘diet’ or ‘nutrition’ or ‘behaviour’], (4) outcomes [‘obesity’ or ‘weight’ or ‘body mass index’ or ‘BMI’], and (5) geographical location [‘Gulf’ or “Saudi’ or ‘Emirates’ or ‘Kuwait’ or ‘Bahrain’ or ‘Oman’ or ‘Qatar’ or ‘KSA’ or ‘UAE’ or ‘Arab’]. Google Scholar was searched for grey literature. All identified records were imported into EndNote version 9 (Clarivate Plc, Philadelphia, United States and London, United Kingdom).

In the initial screening phase, after the removal of duplicates, two authors independently reviewed the titles and abstracts of the remaining studies to exclude irrelevant records. To minimise potential bias and enhance the decision-making process, any discrepancies between the authors were resolved by a third author. Backward citation tracking was also used to identify any additional studies.

As illustrated in Fig. 1, the initial search of the databases yielded 1303 records. After removing 414 duplicates, 889 records were screened based on their titles and abstracts, excluding 846 and leaving 43 potential records. These records were further screened based on their full-texts, and an additional three records were added through cross-referencing, yielding 46 studies. Based on the full text examination, 11 studies met the inclusion and exclusion criteria: 8 peer-reviewed studies [35‐42] and three doctoral theses [43‐45] were included in this systematic review.

The full texts of all potential studies were assessed for their eligibility to be included in this review if they (1) were school-based interventions conducted in any of the GCC countries; (2) targeted school-aged students (5–18 years old) during school hours; (3) included at least one weight-related lifestyle intervention (physical activity, sedentary behaviour or diet); and (4) had at least one anthropometric measure of body weight or weight-linked lifestyle factor (physical activity, sedentary behaviour or diet) as a primary or secondary outcome of interest. However, a study was excluded if it was not conducted in any of the GCC countries, was not school-based, was not an intervention study, or if the outcome was not related to weight or weight-related lifestyle factors. A systematic review of correlates of, and interventions for weight and weight-related behaviours among adults in the Gulf countries [32] only identified seven interventions, four of these did not have control groups. Therefore, this review also included pre-post school-based intervention studies.

Data from each included study was extracted into a table that included study title, authors, publication year, country, study aim(s), design, participant characteristics, sample size, school characteristics, intervention components, duration, and outcomes. Due to the heterogeneity of the studies’ characteristics, intervention components and outcome measures, meta-analysis was not suitable and hence a narrative synthesis was conducted. The tabulated data were analysed to explore the study design, intervention components and outcomes of interest, and limitations of the studies. Driven by the “Behaviour Change Wheel” (BCW) [46], the intervention components included nine intervention functions, which are education; persuasion; incentivisation; coercion; training; enablement; modelling; environmental restructuring; and restrictions. Outcomes of interest were measures of students’ anthropometrics, diet, physical activity, and sedentary behaviour. For each particular outcome, the interventions were broadly categorised as effective, not effective, or mixed effects. An intervention was considered ‘effective’ if there was a statistically significant improvement concerning a particular outcome. An intervention was regarded as ‘not effective’ if the results showed a non-significant improvement or no change at all. If the results were mixed among a particular outcome (e.g. a significant reduction in the prevalence of obesity, but no improvement in the prevalence of overweight), the intervention was considered mixed in terms of the effectiveness.

Two authors independently assessed the quality of the included studies using the Effective Public Health Practice Project (EPHPP) quality assessment tool, which has six key quality components, including selection bias, study design, confounders, blinding, data collection methods, and withdrawals or dropouts [47, 48]. Each individual component was independently rated as strong, moderate, or weak. An overall quality assessment for each study was rated strong, moderate, or weak accordingly. The overall quality of each study was assigned ‘strong’ if at least 4 of 6 quality criteria were rated strong and no criterion was rated weak; assigned ‘moderate’ when only one quality criterion was rated weak; and assigned ‘weak’ if two or more quality criteria were rated weak (Table 3). To ensure consistency, the authors crosschecked the quality of the included studies. Discrepancies between the reviewers’ ratings was discussed until a consensus was achieved.

As illustrated in Table 1, six studies were conducted in Saudi Arabia [35, 37, 39, 41, 44, 45], two in Kuwait [36, 43], and one each in Qatar [38], the UAE [40] and Oman [42]. All included studies were published between 2017 and 2022 except the one by Shama and Abdou [42] in Oman, which was published in 2009. Out of the eleven studies, six were randomised controlled trials [35‐37, 39, 41, 43], four were pre-post studies [38, 40, 44, 45] and one used a static group comparison design [42]. Six studies were purely quantitative [35, 39‐42, 45], and two adopted a mixed methods design [43, 44].

The sample sizes ranged from 28 in an RCT [35] to 3,967 participants across two schools in a pre-post study [40]. The age of the participants ranged from five to nineteen years. In six studies, the study participants were girls only [35, 37, 39, 43‐45], while four studies had male and female students [36, 38, 40, 42] and one study had only male participants [41]. In six studies, the participants were students from intermediate and/or secondary schools [35, 37, 39, 42‐44], two studies had primary school students [36, 38], one had students from primary, intermediate, and secondary schools [41], while the school education stage was not clearly reported in two studies [40, 45]. The number of schools participating in the studies ranged from one [38, 43‐45] to thirty [42]. The interventions were either only conducted at public schools [36‐39, 43‐45], only at private schools [40], at public and private schools [36], while two studies did not report the school type [41, 42]. The study in Oman by Shama and Abdou [42] compared conventional schools with health-promoting schools, which have three main components: health education, health services, and a healthy environment. Intervention components and outcomes of the included eleven studies are detailed in Table 2.

The intervention duration ranged between one month [35] and one year [42]. However, it was not clearly reported in one study [44] and not reported at all in another study [36]. To change participants’ behaviour as per the BCW, intervention components in our systematic review were reported under nine intervention functions, including education; persuasion; incentivisation; coercion; training; enablement; modelling; environmental restructuring; and restrictions.

Participants in all included studies [35‐41, 43‐45] were educated on various topics related to nutrition, physical activity, or health, including the study by Shama and Abdou [42], where health education was a component of the health-promoting schools initiative. Different methods of educational interventions were implemented by providing factual nutritional information and advice (e.g. nutritional posters and cards, recipe cards) to the students and their families [38], instructions for changing nutritional and physical activity behaviours (e.g. increasing fruit and vegetable intake and reducing the intake of sugar sweetened beverages, receiving information about the function of the pedometer) [35, 36], health education and awareness workshops and sessions on diet, physical activity, and obesity risk factors [37, 40, 41, 43, 44], and counselling sessions [38, 39].

Out of the eleven studies, only two [35, 45] used the persuasion function to change participants’ behaviours. There were changes in the perceptions of the intervention group in one study [35], and positive attitudes were reported in the other study [45]. Similarly, incentivisations were reported in two studies [36, 38]. In Allafi’s study [36], the FB + R group was provided with rewards in the form of stickers for achieving the step count milestone. However, in Choudhury’s study [38], participants received stamps in a book when they chose a healthy option in the cafeteria and were awarded a badge at the end, to motivate and reinforce healthy eating behaviours.

Interventions via training were reported in six studies [36, 37, 39, 42, 44, 45], where the participants were engaged in various practical activities related to nutrition, physical activity, and body image perception (see Table 2). Environmental restructuring was reported in three studies in the forms of food labelling and promotion of healthy diets at the school cafeteria [40] and redesigning the school canteen, which included posters and leaflets about healthy diets [38] and was a component of the health-promoting school initiative in the study conducted in Oman by Shama and Abdou [42].

Five studies [37, 40, 42, 44, 45] reported changes in participants’ behaviours through partnerships with stakeholders, including students’ parents [40, 42, 44, 45] or by enabling various teaching aids such as PowerPoint presentations, booklets, games, papers and cards, school boards, group discussions, and stickers [37]. The restriction function of the BCW to change students’ behaviours was reported only in the study conducted by Al-Mughamisi [44], where restricting access to unhealthy food was applied as a part of the environmental changes.

In addition to targeting the school students, three interventions [38, 40, 44] also targeted the teachers and other school staff, while another five studies involved the students’ parents [35‐37, 39, 44].

The outcomes of interest were weight-related measures, dietary behaviour, physical activity, and sedentary behaviour. All studies reported weight-related outcomes, such as a change in BMI, BMI-for-age z-score (BAZ), the prevalence of overweight or obesity, body weight perception, body fat percentage, body fat mass, waist circumference, or waist-to-hip ratio, except one study [44]. Most of the included studies [35, 38, 39, 41‐45] reported changes related to dietary intake and/or behaviour. Eight studies reported outcomes related to physical activity [35‐39, 41‐43, 45], and five reported sedentary behaviour outcomes [35, 37, 41, 43, 45].

Based on the Effective Public Health Practice Project (EPHPP) quality assessment tool, the quality of two studies was rated as ‘moderate’ [41, 43], and ‘weak’ for the other nine [35‐40, 42, 44, 45] (Table 3). The study by Hussein [40] was rated ‘weak’ in all six components of the assessment tool. All included studies were rated ‘weak’ in blinding, except for the study by Hefni [45]. Other ‘weak’ ratings were mainly due to selection bias [36, 40, 42], study design [35, 38, 40, 42, 44, 45], confounders [35‐40, 42, 44, 45], data collection methods [35, 36, 40, 44], withdrawal and dropouts [35‐37, 40, 42, 44].

The included studies were limited in their study designs in terms of sampling errors and participant allocation. Six of the included studies were randomized controlled trials (RCTs) [35‐37, 39, 41, 43], which are generally considered robust for intervention evaluations [49], including school-based weight-related interventions [20]. However, several methodological weaknesses were noted among these studies. For instance, in one RCT [35], the sample size was notably small (n = 28), leading to concerns about statistical power. Additionally, this study and another RCT [39] had unequal numbers of participants in the intervention and control groups, potentially affecting the balance and comparability of these groups. These issues, beside others, such as unreported confounders and variations in study designs, collectively suggest that the overall strength of the study designs was weak or moderate at best. Such limitations should be carefully considered when interpreting the results of these interventions. The RCT by Al-Failakawi [43] had a large sample size (n = 128) with participants assigned equally among the control and intervention groups, however, the study is limited in terms of blinding. Four interventions were pre-post studies [38, 40, 44, 45], and one used a post-test design [42], which threatens internal validity in terms of selection bias. Furthermore, only two studies [41, 43] included in this review reported confounders. In addition to these methodological concerns, the lack of detailed reporting on confounders and other potential statistical errors (such as improper use of p-values or effect sizes) further limits our ability to accurately gauge the interventions’ true effects, in line with the concerns raised by Brown et al. [50]. These methodological limitations warrant more rigorous study designs for future school-based interventions. Habib-Mourad et al. [51] pointed out that weight-related interventions involving children would require large sample sizes and sufficient follow-up periods to observe significant changes in the outcomes. The studies included in this review had sample sizes between 28 and 2890 and follow-up times between 1 and 12 months. The finding regarding the methodological limitations of the included studies is in line with another systematic review on obesity interventions for adults in the GCC states, which found that most evaluations of interventions did not have control groups [32].

Various intervention’s components were used in the studies included in this review, typically in multi-component weight-related interventions in school settings [52]. However, the development of the intervention components was not explicitly discussed in some of the included journal articles. Previous studies that reported successful outcomes, such as weight-related measures, physical activity, and nutrition behaviours, adopted one or more theoretical frameworks [52, 53]. None of the journal articles included in our systematic review reported any theoretical underpinning except the study conducted by Kutbi and colleagues [41], which was based on the social cognitive theory. However, one of the doctoral theses was based on the social-ecological model [44], and the other two were based on social cognitive theory [43, 45]. Particularly in school settings, where multiple stakeholders can work together to achieve a common goal in health promotion interventions, it would be crucial to consider theoretical concepts in the design of the studies [24]. Theory-driven interventions are also important for translating evidence into practice and in making relevant decisions for applying the intervention components in practice [54]. Our review found that a variety of theoretical frameworks underpinned the interventions, emphasising the importance of a multifaceted approach to obesity intervention. These theories highlight the significance of multiple and combined factors in shaping health outcomes. However, the inconsistent application of these theories across studies suggests a need for more robust theoretical grounding in future research.

Despite the limitations of the included interventions, this review suggests that though there is some potential, the overall evidence supporting the effectiveness of school-based interventions in addressing obesity is mixed and warrants cautious interpretation. This is consistent with other reviews of similar interventions in high-income countries [55, 56] and low to middle-income countries [57, 58].

Studies have revealed that the success of interventions could potentially be attributable to the school’s environmental support in terms of finance, incentives, and applying weight management-related policies [56‐59], highlighting the importance of multiple concomitant approaches to counteract obesity or its linked factors among students at schools. Environmental components were only included in three of the school-based interventions [38, 40, 41], indicating more research is required in this area. Considering the social context of the school environment, facilitators and barriers are crucial to determine the success or failure of weight-related interventions [33, 60].

There was considerable variation in the outcomes of the included studies, with some showing improvements in students’ anthropometrics in terms of reducing the prevalence of overweight or obesity, and decreases in BMI or waist circumference, while others did not, highlighting the need for a cautious interpretation of these results. Students’ weight outcomes were improved in some of the included studies [35, 37, 39, 40, 42]. Some systematic reviews in other parts of the world have suggested a range of outcomes, from mild to significant effectiveness of school-based interventions for addressing childhood obesity, particularly if they were long enough, used multiple components, and had parental involvement in delivering the intervention [58, 59]. However, it is important to note that results vary widely, and some studies, including a notable Cochrane Review, report limited effectiveness [61], two meta-analyses conducted by Kanekar and Sharma [62], covering studies from the USA and UK and the other by Harris et al. [63], including twelve studies from the USA, three from Canada and one each from Australia, Chile, and Sweden. Accordingly, there is various evidence that school-based interventions can improve students’ weight status.

The studies in this review demonstrated significant improvements in dietary habits, such as decreased energy and soft drink consumption and increased intake of fruits and vegetables, water, fish, and dairy. These findings align with other similar systematic reviews about school-based weight-related interventions from the USA and Europe [64, 65]. Overall, this suggests that school-based interventions in the GCC countries may improve students’ dietary habits and eating behaviours.

The intervention outcomes of the included studies suggest that physical activity and sedentary behaviour can potentially be improved with school-based interventions in the Gulf countries, which is consistent with studies on school students from other parts of the world [66, 67]. The included studies used various measures for physical activity and sedentary behaviour, which did not allow for direct comparisons across the studies. This calls for standardised, valid, and reliable measurements to improve evidence-based health promotion [33].

A key strength of this study is that to the best of our knowledge, it is the first comprehensive systematic review of school-based weight-related interventions in the GCC countries, a region with one of the highest obesity rates in the world. This review fills a significant gap by providing information on the participants, characteristics, components, outcomes and strengths and limitations of the interventions. A further strength of this systematic review is that we utilised a rigorous and comprehensive study design, including PROSPERO registration, following the PRISMA statement, and using the EPHPP tool for quality appraisal of the interventions. However, including only studies published in English is a limitation of our systematic review, potentially introducing language bias, although scientific studies and policy documents from the GCC countries are usually published in English rather than Arabic. An additional limitation was the preclusion of a meta-analysis due to the heterogeneity of the included studies in terms of study design and intervention components.