Effectiveness of dengue training programmes on prevention and control among high school students in the Yangon region, Myanmar

Background Dengue is one of the health problems in Myanmar. Thus, health promotion in schools is considered a key approach for reducing risk-taking behaviours related to dengue. Objectives The study aimed to evaluate a dengue training programme for high school students to measure changes in knowledge, attitude and practices (KAP) towards dengue; evaluate the effectiveness of the programme in improving prevention and control practices among families and determining changes in larval indices in their dwelling places. Methodology The dengue school training programme was conducted for Grades 9 and 10 students in Yangon. In total, 300 students in the intervention school received training and were compared with 300 students as control. KAP was assessed using a self-administered questionnaire, whereas larval and control practice surveys were conducted at the homes of both groups 3 months before and after the programme. Results The KAP scores of the intervention group increased after the programme. Moreover, the programme improved prevention and control practices and decreased the larval indices in the intervention group. Students from the same group with high scores in knowledge and self-reported practices were less likely to exhibit Aedes larval positivity in their residential areas. Conclusion This study demonstrated the impact of the dengue training programme on the KAP of students and short-term family larval control practices, which influenced household larval indices.


Introduction
Dengue affects more than one-third of the world population [1], with 3.9 billion people living in dengue-endemic areas. The global incidence of dengue fever has approximately doubled in the past three decades and is expected to increase in Asia, sub-Saharan Africa

Study design
This study is quasi-experimental intervention study on Grades 9 and 10 students from two high schools in Yangon, Myanmar. The study provided the dengue training programme to students in the intervention group after a pre-test and conducted a post-test after 3 months. In the control group, the students received only the routine health education programme provided by the school and local health sectors. The students in the intervention groups conducted a survey on larval indices at their dwelling places under the close supervision of the investigators and entomologists 3 months before and after the training. However, the control group the larval survey was performed by the researcher teams.
The study was conducted from August 2018 to April 2019. Two townships in the Yangon region (Hmawbi and Kaw Hmu) were selected based on their similarity in larval indices. The distance between these two townships was 80 km. The schools were randomly selected as the intervention or control school. A total of 300 students in Grades 9 and 10 from each high school were recruited. Written informed consent was provided by parents, while assent by students was obtained prior to the study.

Sample size
The sample size was calculated using the two independent proportions formula. The anticipated level of knowledge of the control group was derived from Suwanbamrung et al. [21], i.e. on the basic dengue knowledge questionnaire with regard to dengue: 70.9% (p1). The anticipated proportion of students with knowledge in the study group was 81% (p2). With a test power of 80% and a significance level at 5%, the required minimal sample for this study is 280. Correction for drop-out was set at 10% such that a sample size of approximately 300 per group is considered appropriate.

Interventions
Intervention: Dengue training programme was applied to the intervention group. Tool for training: The main content included the biology and mode of transmission of Aedes, common clinical manifestations and prevention and control methods. Visual aids, such as flipcharts, audio-visual media and video, were used. The students received intensive training by entomologists before performing larval survey and using the survey form under the supervision of entomologists, investigators and school health teachers (i.e. teaching and hands-on training). This training focused on the use of the larval survey methods in the residential areas of students, including the identification of breeding sites and methods for controlling the vector (e.g. using temephos or larvivorous fish or cleaning and covering water containers), identifying of the mosquito life cycle and calculation of the index.
Training method: Actual training was performed by classifying the students into three subgroups with 100 students in each group. The average time for in-class training, including discussion and question-and-answer sessions, was 2-3 h per group. Training materials, such as leaflets and infographics on dengue, were distributed to the students to share with their families (Supplementary Information 2).
The questionnaire was tested for internal consistency among 60 students, whose data were excluded from the final analysis. Cronbach's alpha reached 0.7.
The larval survey method: The larval survey primarily comprised the identification of breeding sites, characteristics of containers, covering methods, frequency and methods for cleaning, use of temaphos and identifying the presentation of Aedes larva using the larval survey form. As per group, the intervention group conducted the larval survey under the supervision of the entomologists and researcher two weeks before and 3 months after the training. However, the larval survey was performed by the research teams ( Fig. 1

Statistical analysis
Data on KAP and the larval survey were coded, cleaned, reviewed and entered into Microsoft Assess (2016). Statistical analysis was conducted using STATA version 14. Additionally, descriptive statistics was performed to obtain the mean, standard deviation, frequencies and percentages. For categorical variables, data were reported as proportions (%) and compared using chi-squared tests. Multiple logistic regression analysis was used to identify the factors associated with each KAP level and Aedes larval positivity at participant homes. P-value <0.05 was considered statistically significant.

Socio-demographic characteristics
Both schools were under the government. The male and female student ratio and the mean age were similar for both groups. More than half of the students came from middle-income families, with Bamar as their main ethnicity. The housing style for both groups was similar. Lastly, the majority of living places had electricity and toilets (Table 1). Table 2 presents the distribution of the correct responses of students in both groups before and after training. In the pre-test, both groups correctly answered that the major vector of dengue was Aedes, which bite primarily during the daytime (p > 0.05). After training, significant improvements were noted in both topics only for the intervention group (p < 0.001). Regarding breeding site, correct responses to common outdoor breeding sites (i.e. discarded tyres, garbage and water tanks) were similar in the pre-test for both groups but significantly increased for the intervention group in the post-test. Prior to intervention, the majority of students misunderstood that unlikely outdoor breeding sites (i.e. wastewater drainage and rice fields) were common breeding sites of Aedes. After the training, however, >95% of the students in the intervention group responded that dirty water in the drainage was not a breeding site, whereas 1/3 of both groups continued to believe that rice fields are breeding sites.

Knowledge on dengue and its prevention and control
For common indoor breeding sites, a similar number of students correctly answered in the pre-test that water containers for daily usage and for toilet flushing and water cup for legs of cupboard (to prevent the food from ants) were breeding sites. After the training, the number of students in the intervention group who provided correct answers significantly increased compared with the control group (p < 0.001). More students in the control group correctly answered that fever was a major symptom of dengue compared with the intervention group in the pre-test (p = 0.003). However, both groups exhibited similar levels of knowledge about skin rash and bleeding as clinical manifestations in the pre-test (p > 0.05). The students of both groups were least concerned about bleeding as a clinical feature of the disease prior to the training. In the post-test, the intervention group displayed a significant increase in correct responses to the three common clinical features of dengue, including bleeding manifestation. Baseline knowledge about prevention and control was similar for both groups. The baseline correct responses to using temephos in water and using larvivorous fish were low for both groups prior to training. Moreover, both groups displayed similar background knowledge about the threat of dengue (e.g. 'Dengue fever is a preventable disease' (p = 0.028) and 'Dengue may cause death' (p = 0.342).
After training, the students in the intervention group exhibited significant improvements in knowledge for all items including biology, transmission, preventive measures and knowledge about the disease. Interestingly, the correct responses of the students in control group also significantly increased, especially in knowledge about preventive and control ('changing stored water on a weekly basis' (p < 0.001) and 'using a mosquito net during the day' (p < 0.001) ( Table 2).

Attitude towards dengue and its prevention and control
More than half of the students in both groups similarly strongly agreed/agreed to the statement that 'dengue can occur all year round in Yangon' (p = 0.274). Regarding repeated infection and death, the perception of the students in the intervention group € The correct answer is "No".
*P-value <0.05 ** P-value <0.001 between the two groups after study intervention. Regarding attitude towards control methods, most students in both groups strongly agreed/agreed to the statements 'A weekly cleaning of mosquito breeding sites around the house is essential' and 'Covering water containers will prevent mosquitoes from laying eggs'. After training, the ratios of agreement to disagreement increased significantly for the intervention group (p < 0.001).
For negative statements, more than one-third of the students in both groups strongly agreed/agreed to the statement 'Using temephos in water could be harmful to health' (p = 0.124), whereas approximately two-third of the students strongly agreed/agreed  with the items 'Regularly monitoring mosquito breeding sites in the home is difficult' (p = 0.207) and 'Wearing protective clothing during daytime is very hot and inconvenient' (p = 0.360) in the pre-test. After training, the number of strongly disagree/disagree responses increased significantly for the intervention group (p < 0.001). For the item 'Sleeping under mosquito net during the day is inconvenient', more students in the intervention group strongly agreed/agreed than those in the control group in the pre-test (p < 0.006). After training, the majority of the students in the intervention group exhibited disagreement towards this negative statement on this preventive measure (p < 0.001), whereas the attitude of the control group remained unchanged (p = 0.207). Prior to training, students from both groups similarly strongly agreed/agreed with the item 'Mosquito coils emit a bad smell and can be harmful to health.' However, they exhibited a higher degree of agreement with the unfavourable smell and the harmful effects of the chemical substances in mosquito coils after the training.
For regular health promotion from the health department, approximately 70% of the students strongly agreed/agreed that health promotion was effective (p = 0.046) before the training. Afterwards, only the students in the intervention group exhibited increased confidence in the existing training programmes conducted by the relevant authorities (p < 0.001; Table 3). Table 4 summarises the self-reported practices of the students. Both groups displayed similar self-reported practices for all items. However, the control group reported a significantly high percentage of practices for 'weekly changing of water in pot trays' (p < 0.001). The number of students who reported the use of temephos in water was low for both groups at baseline (intervention = 173/ 300, 57.6%; control = 165/300, 55.0%; p = 0.510). After training, the percentage of students with self-reported practices significantly increased for the intervention group (p < 0.001). Interestingly, the use of temephos in water remained low after training in the intervention group (pre-training = 173/300, 57.6%; post-training = 203/300, 67.6%; Table 4). Table 5 presents the level of KAP before and after training. KAP scores were categorised into three groups by using the modified Bloom cut-off point [23]. The KAP levels were similar for both groups in the pre-test. Most students exhibited a moderate level of knowledge, neutral attitude and good self-reported practice. After the training, however, the KAP levels for the intervention group significantly increased (p < 0.001; Table 5).

Effectiveness of the school training programme
The study evaluated the effectiveness of the dengue training programme using changes in actual practices, which were collected during the vector survey, changes in larval indices and the association between KAP and larval positivity in the household.

Changes in larval control practices
The larval survey included measurable practices, such as the use of temephos, weekly cleaning of water containers and properly covering water containers. Before the intervention, the number of water containers was 1662 and 2234 for the intervention and control Table 5 Level of knowledge, attitude and self-reported practices of both groups before and after intervention. groups, respectively. The number of water containers slightly decreased to 1612 and 2153 at 3 months after training. The rates of observed practices by family members were relatively lower than those of the self-reported practices of the students. The larval control practices between the two groups were similar before the intervention. In the second survey, significant improvements were noted for 'weekly cleaning of water containers' and 'properly covering water containers' only for the intervention group (p < 0.001), whereas no difference was observed in the use of temephos for both groups in pre-and post-tests (Table 6). Lastly, the larval survey did not observe the use of larvivorous fish for both groups.

Change in household larval indices
The larval indices of the intervention and control groups decreased after training (Table 7).

Association between KAP and larval positivity in the household for both groups
Multivariate analysis employed three levels of each KAP domain to determine the association between KAP levels and larval positivity. The high level of knowledge of the intervention group was associated with less chances to detect Aedes larvae in their homes  Table 8].

Discussion
The study observed changes in KAP for the intervention group. These findings agreed with those of previous school-based studies [36][37][38]. The baseline level for KAP was high for both groups, especially knowledge about the threat of the disease, Aedes breeding sites and preventive and control measures for dengue. This result reflects the efficacy of the existing health educational programme or live messages using other routes, such as social media. Several studies reported the regular exposure of high school students in Myanmar to social media [39,40]. Nowadays, technology has become the main channel for delivering health information [23]. Moreover, electronic health and mobile health are used for healthcare delivery, disease surveillance, health education and health promotion [24,25]. Such online open sources may be related to the health education background in a growing city such as Yangon.
The high baseline scores for knowledge and awareness about dengue as a life-threatening disease (e.g. 'It is a deadly disease' and 'It occurs every season') were detected for both groups. According to the health belief model, awareness of the threat of a disease is a key factor that leads to the active seeking of information, which may lead to other significant changes in practices [36,26]. Nearly all issues related to knowledge and attitude significantly increased for the intervention group after the training. Interestingly, a common serious clinical sign, bleeding, was less recognised, which may be related to the Burmese language. In Myanmar, dengue haemorrhagic fever is called Thawe Lon Tut Kwe, which means flu disease and leaves out the haemorrhage part. The World Health Organization issued best practices for naming new infectious diseases, which were frequently given by people outside the scientific community, thus exerting unintended negative impacts and unnecessarily leading to untoward effects [27].
Furthermore, the students misunderstood that rice fields and dirty water were common outdoor breeding sites prior to the programme; however, this perception changed for the intervention group. These findings were similar to those of a study in Malaysia, where nearly half of the students misunderstood that Aedes mosquitoes prefer to lay eggs in dirty water [36]. Moreover, confusion may have emerged regarding other mosquito-borne diseases, such as Japanese encephalitis, malaria and lymphatic filariasis. Thus, similar training programmes should provide knowledge about other epidemic vector-borne diseases so as to fit with the integrated vector control programmes of the WHO as well as the local context for disease prevention and control measures [28].
Overall attitude exhibited changes after the intervention. However, only the attitude towards the negative perception of harmful mosquito coils was retained by majority of the students after the intervention. The misunderstanding about the 'toxicity of temephos' was high at baseline for both groups but changed to a positive attitude in the intervention group after the training. Regarding the perception of temephos, Legorreta-Soberanis et al. [29] conducted a community-based study in Mexico and reported that 69% of people believed that bathing and drinking water that contain temephos posed health risks. Thus, information provided by health care personnel, community leaders and related personnel in this area should be explored.

Table 6
Changes in larval control practice after dengue school training program.  Although knowledge and awareness of dengue increased after training, the extent to which knowledge can be translated into practices and how the practice will in fact influence the larva population remain unclear. Despite the association between knowledge and preventive behaviours towards dengue in several studies [36,30,31], knowledge alone could not directly translate into practices or predict practices [32][33][34][35]. Koenraadt et al. [41] reported that the gap between knowledge and practice is an important challenge for dengue control and a dynamic target of the reduction of the A. aegypti population. The current study found that self-reported practices significantly increased for the intervention group; however, such practices were observed by less than 50% of households. Similar findings were found for Myanmar wherein low scores for practices were observed among caregivers with high levels of knowledge [42]. Khun and Manderson et al. [18] suggested that a long-term school-based dengue programme was required to ensure the translation of knowledge into practice. Moreover, selective health messages that are relevant to daily practices should be identified to change knowledge into practice for dengue prevention and control.
Despite the increased knowledge and decreased negative attitude towards temephos after training, the self-reported practices of the intervention group remained the same. The survey also found that only 50% of the households actually used temephos after the training. The acceptability of temephos is complicated, which was outside the scope of the study. Thus, we infer that this result is influenced by the misconception of temephos as a harmful chemical or simply the lack of knowledge of how to use it. Thavara et al. [43] reported that the refusal to use temephos may be owing to its unpleasant door, water turbidity and perceived safety risks. In Yangon, the responsibility for regular control practices, including the distribution of temephos, is shouldered by the local health department and community leaders [44] such that the residents assumed that these tasks were the responsibility of the authorities instead of their own. Apart from the issue of temephos utility, this study also demonstrated an increase in the percentage of weekly cleaning and covering of water containers for the intervention group. Nevertheless, less than half of the households actually practiced these measures. Phuanukoonnon et al. [45] found that the size, shape and location of water containers determine the success of these methods in Thailand.
Moreover, the current study did not report on larvivorous fish despite its advantages over other prevention and control methods because of its feasibility, low cost and being environmentally friendly. This concept may be due to culture, because owning a fish tank or pond in the household (which is common in Thailand) is uncommon in Yangon. The larval survey form used in this study was created by researchers based on the Thai culture and lifestyle (Mahidol University). Thus, adapting the tool to the Myanmar culture would have been favourable. A qualitative study should also be conducted prior to the tool creation for the survey among health authorities, community leaders and other stakeholders.
Suwanbamrung et al. [20] conducted a school-based participatory education programme in Thailand, reporting an increase in Table 7 Household larval indices before and after intervention for both groups.  knowledge on dengue prevention and control, which was associated with a reduction in larval indices in the school and the households of the students. The current study obtained similar findings, such as the decreased larval indices for the intervention group after training. Although the level of larval indices apparently decreased, the CI remained slightly higher than the standard criteria set by the WHO [46] as well as the applied larval indices of the MOHS in Myanmar (HI > 10%, BI > 20%) [44]. However, other factors may influence this efficacy because the larval indices also decreased for the control group. The first survey was conducted during the wet season with slightly higher number of water containers compared with the second survey, which was conducted during the dry season. Many studies reported that seasonal fluctuations influence the rate Aedes-positive containers in residential and public areas owing to differences in rainfall, temperature, evaporation and humidity between the two seasons [47,48]. During the dry season, people may have consumed all stored water such that the majority of water containers are empty (during the second survey), which may also influence the larval indices. However, with the disproportionate change in containers (slightly decreased number of containers: marked decreased of larval indices), this hypothesis may not be the only reason for the marked decrease in larval indices during the post intervention survey in both groups. This study also explored the effectiveness of the training by testing the correlation between KAP levels and larval positivity. Only high levels of knowledge and high scores for self-reported practices were significantly related to less larval positivity. Previous studies supported that knowledge gain may lead to successful behaviour modification. For example, Castro et al. [49] proposed that people with better access to information about dengue may assume a better understanding of the disease, which could lead to practices related to dengue. Moreover, many studies supported the direct link between knowledge and the adoption of best practices for the reduction of the breeding sites of A. aegypti after health education activities [31,50]. One of the possible reasons for these strong associations in the current study were the self-experimental training, which enabled the students to actually practice and efficiently translate it to family members compared with given simple reading materials. This result is similar to that of a study conducted on high school students that found that they are capable of communicating well with family members than primary school students in Thailand [51].
The most important goal of school health education programmes should be sustainable behavioural changes [52]. Local health departments and school health teachers should provide health education on dengue at least once a year at the beginning of the monsoon season in Myanmar. However, such programmes, which are provided through conventional lectures and involve printed materials, are not a major part of the curriculum. Thus, identifying a suitable tool for knowledge transfer to students is key for motivating families to increase their awareness and participating in dengue prevention practices. Students are interested in ludic strategies and practical activities, which was demonstrated by the significantly high KAP scores obtained by studies in Brazil [53] and the Philippines [50]. Evidently, these mixed-experimental methods in the training could demonstrate its effectiveness of it. A study provided evidence of the effectiveness of mixed-experiment teaching in health about obesity prevention among school children [54]. Thus, the current study proposes that learning by doing may motivate the students compared with conventional lectures. Nevertheless, mixed-experimental training is unique to each disease and is difficult to implement and sustain because it is time consuming and has limited resources despite the strong evidence of its efficiency at the community level [55]. Conversely, the new life style for sustainable learning through the use of social media and online platforms could be an alternative.
In 2015, UNESCO, Ericsson, the UK Department of International Development (DFID) and Ministry of Education initiated the Information Communication Technology (ICT) for Education Project across rural schools in Myanmar to create innovative teaching approaches and spread knowledge among the teachers and pupils [56]. Evidently, youth from resource-poor communities exhibited greater online engagement and positive changes in self-reported health behaviour for prevention against an arboviral disease in the Dominican Republic [57]. Lwin et al. [58] revealed that the youngest age group reported the highest levels of perceived severity and susceptibility to dengue and were well equipped with media exposure through the Internet and social media. Against this background, schools may need to emphasise a new paradigm for sustained educational training based on online learning programmes and should be adept to access social media and obtain credit for science subjects for lifelong learning.

Conclusion
The student training programme on dengue fever among the high school students in the Yangon region exhibited short-term effects regarding knowledge, attitude and self-reported practices. Despite the significant increase in KAP levels, they could not be translated into family practices and significant changes in the larval indices. Thus, this study recommends the use of the mixed-experimental training for future studies. To achieve the goal of dengue prevention and control, vector control activities may not be adequate despite the significant effort to decrease the dengue burden, and other alternative methods may be required. Vaccine implementation was proposed and the acceptance of the vaccine in this population should be explored [59].

Limitation
The main limitation of this study is that it was conducted only for 8 months. Moreover, assessment of the study outcomes was performed only two times, that is, 3 months before and after the training programme. Thus, a longer period is required to determine the long-term results.

Author contribution statement
Soe Htet Aung: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.