Background
Non-communicable diseases (NCDs) are the leading global cause of death and are responsible for 70% of deaths worldwide [
1]. In Sri Lanka, NCDs account for 75% of total deaths, and 40% of these NCDs are cardiovascular diseases (CVD) [
2]. Regarding the risk factors of NCDs in Sri Lanka, the prevalence of overweight and obesity is 18.9 and 3.5% in men and 32.9 and 10.0% in women [
3], respectively, and the prevalence of high blood pressure is 30.5% in men and 26.2% in women [
2]. Unhealthy behaviours such as lack of physical activity, unhealthy diet, tobacco use, and the harmful use of alcohol are common, which in turn may lead to obesity, raise blood pressure, and ultimately develop NCDs [
1].
Middle-income countries in the east Asia and Pacific region will face workforce shortages of health professionals (i.e. physicians and all other health workers) because their demand will exceed supply due to economic growth, rapid population growth, and aging [
4]. Several intervention studies have suggested an important role of non-professional healthcare workers, including community health workers, to combat NCDs in low- and middle-income countries [
5‐
7]. A few intervention studies provided training programs for school-aged children to act as change agents for the improvement of their parents’ lifestyle [
8‐
10]. In Sri Lanka, the promotion of healthy habits initiated by school children led to the reduced body weight and increased physical activity of their mothers [
9]. A study in Brazil reported a reduction of CVD risk factors among parents of children who underwent an educational program for health promotion [
8]. One study in China indicated that an education programme for primary school children was effective in lowering the salt intake of their families [
10]. These approaches are also beneficial for youths themselves in improving their self-efficacy by experiencing community empowerment [
11], as well as their health in later life [
12]. Interestingly, the above-mentioned study in Sri Lanka [
9] presented a case report in which some children involved their neighbourhood homes in the health promotion (e.g. playing games and exercising with their parents). However, no study has examined whether youths, including children, who received health education can improve the health of community adults who are not their parents or relatives. One study [
13] suggested that certain segments of society may be more receptive to youth-led engagement than others, which could be applied to non-professional healthcare workers of youths. It is important to ascertain whether youths can take on the role of non-professional youth healthcare workers (or change agents, as termed in the present study), because this approach has the potential to improve the health status of many more adults than could previous approaches targeting children’s parents.
We developed an intervention program enabling youths, including school children who were members of a youth club, to act as change agents in promoting healthy lifestyles for community adults to reduce CVD risks. Under these hypotheses, the present community-based intervention study in Sri Lanka aimed to investigate the effect of youth change agents on body weight and blood pressure as primary outcomes and health-related lifestyle (e.g., physical activity and dietary intake) as secondary outcomes among community adults over a 12-month intervention.
Results
Figure
1 shows a study flow diagram with the number of clusters and participants at each phase of the trial. From each of the 12 GN divisions in the intervention and the control group, 47 households (24 in intervention and 23 in control) were invited to the baseline survey. Five households refused to participate in the survey (2 in intervention and 3 in control). Eligible adults were invited among households (292 adults in intrevention and 277 adults in control). Of the 512 adults (262 in intervention and 250 in control) who agreed to participate and completed the baseline survey, 29 adults (17 intervention and 12 control) did not attend the follow-up survey at 12 months, leaving 483 adults (245 intervention and 238 control) for completer analyses.
Table
1 shows baseline characteristics of the study participants. The two groups were similar in terms of ethnicity, region, socioeconomic status, body weight (or BMI), blood pressure, and lifestyle variables. Mean (SD) body weight and BMI were 63.3 kg (13.7) and 25.0 kg/m
2 (4.9) in the intervention group and 62.7 kg (12.7) and 24.9 kg/m
2 (4.6) in the control group, respectively. The proportion of overweight (BMI 25.0–29.9 kg/m
2) and obese (BMI ≥ 30 kg/m
2) were nearly 35–36 and 12% in both groups. Mean (SD) systolic and diastolic blood pressure were 127.3 mmHg (18.8) and 83.4 mmHg (10.9) in the intervention group and 126.7 mmHg (20.7) and 83.6 mmHg (12.7) in the control groups.
Table 1Baseline characteristics of study participants
No. of participants | 245 | 238 |
Age (years), mean ± SD | 46.1 ± 8.1 | 44.8 ± 8.2 |
Women | 130 (53.1) | 124 (52.1) |
Ethnicity |
Sinhalese | 234 (95.5) | 226 (95.0) |
Tamil | 10 (4.1) | 11 (4.6) |
Muslim or others | 1 (0.4) | 1 (0.4) |
Religion |
Buddhism | 218 (89.0) | 218 (91.6) |
Hindu | 7 (2.9) | 10 (4.2) |
Roman Catholic/Christian | 19 (7.8) | 10 (4.2) |
Islam | 1 (0.4) | 0 (0) |
Education attainment |
Primary level (grade 1–5) | 34 (13.9) | 29 (12.2) |
Junior high school | 145 (59.2) | 150 (63.0) |
High school or higher | 66 (26.9) | 59 (24.8) |
Household income (Rp/month)a |
≤40,000 | 200 (81.6) | 200 (84.0) |
40,001–60,000 | 32 (13.1) | 25 (10.5) |
≥60,001 | 13 (5.3) | 13 (5.5) |
Current workers | 123 (50.2) | 137 (57.6) |
Current diseases |
Type 2 diabetes | 42 (17.1) | 39 (16.4) |
Dyslipidemia | 25 (10.2) | 33 (13.9) |
Hypertension | 55 (22.5) | 47 (19.8) |
BMI categories |
<18.5 kg/m2 | 21 (8.6) | 20 (8.4) |
18.5–24.9 kg/m2 | 110 (44.9) | 103 (43.3) |
25.0–29.9 kg/m2 | 85 (34.7) | 86 (36.1) |
≥30 kg/m2 | 29 (11.8) | 29 (12.2) |
Primary outcomes |
Body weight (kg), mean ± SD | 63.3 ± 13.7 | 62.7 ± 12.7 |
BMI (kg/m2), mean ± SD | 25.0 ± 4.9 | 24.9 ± 4.6 |
Systolic blood pressure (mmHg), mean ± SD | 127.3 ± 18.8 | 126.7 ± 20.7 |
Diastolic blood pressure (mmHg), mean ± SD | 83.4 ± 10.9 | 83.6 ± 12.7 |
Secondary outcomes |
Physical activity |
Leisure-time physical activities, recommended levelb | 47 (19.2) | 43 (18.1) |
Sedentary times (minutes/day), median (IQR) | 180 (120, 300) | 180 (120, 300) |
Dietary intake |
Vegetables, ≥ two servings/day | 83 (33.9) | 92 (38.7) |
Fruits, ≥ one serving/day | 84 (34.3) | 86 (36.1) |
Snacks, ≥ twice/day | 90 (36.7) | 91 (38.2) |
Sugar-sweetened beverages, ≥ once/day | 20 (8.2) | 15 (6.3) |
Current smoking | 38 (15.5) | 36 (15.1) |
Alcohol intake, > low risk of drinking levelc | 11 (16.9) | 18 (26.9) |
Table
2 shows the differences in the outcome variables between the intervention and the control groups at the 12-month follow-up survey. With respect to the primary outcomes at the follow-up, the intervention group had a significantly lower mean (SD) body weight (61.8 kg [12.7]) and BMI (24.4 kg/m
2 [4.4]) than did the control group (body weight 64.0 kg [12.8] and BMI 25.5 kg/m
2 [4.7]); the mean difference (95% CI) upon comparing the intervention with the control group was − 2.83 kg (− 3.31 to − 2.35) for body weight and − 1.12 kg/m
2 (− 1.32 to − 0.94) for BMI (difference between groups,
p < 0.001). In a subgroup analysis among adults who were overweight (BMI ≥ 25 kg/m
2) at baseline, we observed a more pronounced difference; the mean (95% CI) difference between the intervention and control groups was − 3.69 kg (− 4.48 to − 2.90) for body weight and − 1.50 kg/m
2 (− 1.80 to − 1.20) for BMI (difference between groups,
p < 0.001) (Appendix Table 1).
Table 2Effect of intervention on primary and secondary outcomes at 12th-month follow-up
Primary outcomes |
Body weight (kg) | 61.8 ± 12.7 | −1.51 ± 3.18 | 64.0 ± 12.8 | 1.36 ± 2.30 | − 2.83 (− 3.31, − 2.35)b | < 0.001 |
BMI (kg/m2) | 24.4 ± 4.4 | − 0.59 ± 1.27 | 25.5 ± 4.7 | 0.54 ± 0.95 | − 1.12 (− 1.32, − 0.94)b | < 0.001 |
Systolic blood pressure (mmHg) | 127.1 ± 18.9 | 0.44 ± 15.4 | 128.4 ± 18.1 | 1.11 ± 13.4 | −0.88 (− 3.18, 1.42)b | 0.45 |
Diastolic blood pressure (mmHg) | 84.0 ± 10.7 | 0.44 ± 10.6 | 85.4 ± 10.1 | 1.01 ± 9.20 | −0.94 (− 2.64, 0.77)b | 0.28 |
Secondary outcomes |
Physical activity |
Leisure-time physical activities, recommended leveld | 31 (12.7) | | 20 (8.4) | | 1.58 (0.84, 2.96) | 0.15 |
Sedentary times (minutes/day) | 180 (120, 300)+ | 3.33 ± 180 | 180 (120, 300)+ | 5.27 ± 140 | 1.16 (− 28.0, 30.3)b | 0.94 |
Dietary intake |
Vegetables, ≥ two servings/day | 106 (43.3) | | 92 (38.7) | | 1.24 (0.79, 1.94) | 0.35 |
Fruits, ≥ one serving/day | 112 (45.7) | | 82 (34.5) | | 1.71 (1.10, 2.65) | 0.02 |
Snacks, ≥ twice/day | 52 (21.2) | | 107 (45.0) | | 0.32 (0.21, 0.48) | < 0.001 |
Sugar-sweetened beverages, ≥ once/day | 14 (5.7) | | 15 (6.3) | | 0.86 (0.35, 2.09) | 0.74 |
Current smoking | 29 (11.8) | | 29 (12.2) | | 0.86 (0.32, 2.31) | 0.77 |
Alcohol intake, > low risk of drinking levele | 9 (11.8) | | 17 (22.1) | | 0.50 (0.15, 1.71) | 0.27 |
Blood pressure, which was one of the primary outcomes, did not significantly change from the baseline to the follow-up in either group. There was no significant intervention effect on blood pressure; the mean differences (95% CI) between the intervention and control groups in systolic and diastolic blood pressure were − 0.88 (− 3.18, 1.42) and − 0.94 (− 2.64, 0.77), respectively. Among those who had hypertension at baseline, there was also no intervention effect (Additional file
2: Table S2).
With regard to health-related lifestyle in secondary outcomes, there were no statistically significant differences in the odds of being engaged in recommended levels of leisure-time physical activities between the two groups at the follow up. The intervention group showed significantly higher odds of consuming ≥ one serving/day of fruits (OR 1.71, 95% CI 1.10–2.65) and significantly lower odds of consuming ≥ twice/day snacks (OR 0.32, 95% CI 0.21–0.48) than the control group. There was no significant difference in the consumption of vegetables or sugar-sweetened beverages between the two groups.
Discussion
In this randomised controlled trial that enlisted youths as change agents to lower CVD risks among community adults in Sri Lanka, the intervention group showed a statistically significantly greater reduction in body weight than the control group. As regards diet, the proportion of adults who consumed the recommended levels of fruits was significantly higher and the proportion of adults who consumed snacks frequently was significantly lower in the intervention group than in the control group. There was no significant difference in blood pressure between the two groups. To our knowledge, this is the first study to report the effectiveness of the intervention by youths for the health of community adults.
The present findings regarding the effect of the intervention on body weight are in line with our previous study in Sri Lanka [
9], showing a larger weight reduction in mothers who received support from their school-aged children to identify and improve determinants of health in their family; the mean difference (95% CI) of body weight between the intervention with the control group was −2.49 kg (−3.38, −1.60). In contrast, there was no intervention effect on body weight observed in a Brazilian study [
8], where school children received an educational program for CVD prevention and asked parents to modify the lifestyle of the whole family. Such different results for body weight reduction may be attributed to the difference in outcome settings and strategy of the intervention. In the present and the previous study in Sri Lanka [
9], body weight was one of the primary outcomes and the intervention was designed to modify the lifestyles of adults with an emphasis on weight control, as we will discuss in the next paragraph in greater detail. In the Brazilian study [
8], the education for the children covered nutrition, exercise, and smoking to reduce the overall CVD risk profile of parents, but did not specifically address weight reduction. Moreover, the current and previous study in Sri Lanka [
9] actively intervened in the lifestyle of adults through children or youths, whereas in the Brazilian study [
8] children were the target of education and were asked not to feel responsible for lifestyle modifications of their family.
The weight reduction observed in the present study may be primarily attributed to the improvements in diet. The intervention group consumed snacks less frequently than the control group after the intervention. According to the case reports reported by youths, some adults tried to cut their spending on unhealthy foods and reduce their excess dietary intake by recording their dietary habits. The monthly plotting of body weight in the chart might facilitate adults in avoiding energy dense foods and eating too much. We also observed a significant increase in fruits intake in the intervention group. There is evidence suggesting a protective role of fruit fibre against obesity [
26]. The replacement of snacks and other energy dense foods with fruits might be effective in weight management [
27].
Turning to leisure-time physical activity, the youths encouraged adults to do some exercise, such as outside activities or games with other adults and/or their children together. According to the reports from the youth, adults living in about 60% of the intervention GNs did physical exercise enthusiastically with the support of the youths. In spite of these promotions, the prevalence of adults who engaged in the recommended level of leisure-time physical activity in the intervention group was not significantly higher than that in the control group after the intervention. One reason for this finding could be the insufficient ability of the questionnaire we used to detect moderate changes of the physical activity.
We observed no intervention effect on blood pressure. In contrast, a significant reduction in blood pressure was reported in a Chinese study [
10] in which students were educated to deliver salt reduction message to their parents. Unlike the Chinese study [
10], the present study did not specifically address salt reduction, whose effect on lowering blood pressure is convincing [
28]. Future studies for the prevention of hypertension should incorporate salt reduction as a major component of the intervention, especially in countries with high salt consumption.
It remains to be determined whether the present findings are applicable to other settings and how the present approach of community health promotion by youth club could be scaled up in a wider context. Besides the presence of active youth club or similar association in the community, it is necessary to develop a facilitator-support system from which youths can receive comprehensive guidance. Youths should be guided to promote their awareness, understanding, and tolerance of other people, cultures, and societies so that they can provide standardized and high-quality intervention universally as change agents.
The current study is strengthened by its cluster randomised controlled design, preventing the influence of measured and unmeasured confounders. The study limitations also warrant mention. First, lifestyle factors such as physical activity and dietary habits were self-reported, and misclassification due to inaccurate responses might thus have diluted the effect of the intervention. Second, lifestyles were assessed by interviews, raising a concern of interviewer bias. However, the interviewers were blinded to the intervention allocation. Therefore, such a possibility would be low. Third, there were variations across the intervened communities in the type of activities and frequency of community visits of youths according to youths’ skills and motivations. The effect of intervention may be attenuated due to the inclusion of communities with low levels of health promotion activities. Fourth, we cannot clarify the long-term effect of this program beyond 12 months. Fifth, although this study applied an intention-to-treat approach [
29], we excluded refusals and non-attendance at the follow-up survey from all randomized subjects (303 subjects in the intervention group and 288 subjects in the control group). Sixth, although most of the adults received support from the youths to improve their health behaviours, we did not record them. Therefore, there was a possibility that there were adults who did not receive a direct approach from youths, whereas we expected that adults might receive not only a direct approach from change agents but also a spillover effect [
30] of the change agents on health behaviours among adults in the community. Finally, it remains unclear whether this type of program is generalisable for the following reasons: 1) the present study was performed in a semi-urban area of Colombo, 2) we selected one DS division based on logistical feasibility before the randomization setting, and 3) there was a possibility of some influence of the health survey, which was conducted before the trial, on health behaviours for both the intervention group and the control group, although we could not assess whether this health survey attenuated or promoted the effect of change agents of youth.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.