Health messaging with Viber application on risk factors of non-communicable diseases among the rural population: community-based quasi-experimental study

International Journal of Community Medicine and Public Health Lwin SMM et al. Int J Community Med Public Health. 2021 Aug;8(8):3823-3829 http://www.ijcmph.com pISSN 2394-6032 | eISSN 2394-6040

DOI: https://dx.doi.org/10.18203/2394-6040.ijcmph20213009

Original Research Article

Health messaging with Viber application on risk factors of

non-communicable diseases among the rural population:

community-based quasi-experimental study

Swe Mar Myint Lwin1*, Khaing Lay Mon1, Aung Soe Htet2

1Department of Health Behavior and Communication, University of Public Health, Myanmar 2International Relations Division, MOHS, Nay Pyi Taw, Myanmar Received: 30 May 2021 Accepted: 29 June 2021 *Correspondence: Dr. Swe Mar Myint Lwin, E-mail: swemarmyintlwin@gmail.com Copyright: © the author(s), publisher and licensee Medip Academy. This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

ABSTRACT Background: Mobile health and mobile phone technology have become increasingly noticed as favorable communication channels for the prevention of non-communicable diseases (NCDs) by reducing the unhealthy behaviors of NCDs. The study was carried out to assess the effect of the health messaging with the Viber application on modifiable risk factors of NCDs. Methods: A community-based quasi-experimental study was conducted among a total of 240 community members who were 30 to 60 years old selected from two villages in Twantay township with a one year intervention. Face-to- face interviews with all participants using the WHO STEP survey questionnaire and measurements of behavioral and metabolic risk factors of NCDs were carried out at baseline and after 12 months follow up. Changes in behavior and metabolic risk factors between the intervention and control group were analyzed using mixed-effects random- intercept linear regression modeling with propensity score adjustment. Because of the small sample in smoking, smokeless tobacco and alcohol users, the Wilcoxon rank-sum (Mann-Whitney) test was used to compare the changes between the two groups. Results: The amount of smoking and smokeless tobacco use among current users in the intervention group was reduced after the intervention (p<0.05). Of metabolic risk factors, participants in the intervention group had a significant net reduction in fasting blood sugar with a mean change relative to controls of -18.7 mg/dl 95% CI (-32.83, -4.56, p=0.010). No significant mean reduction was seen in other behavioral or metabolic risk factors. Conclusions: The study results showed that intervention affected a reduction in some behavioral and metabolic risk factors of NCDs. Keywords: Health messaging, Viber application, Risk factors of NCDs

INTRODUCTION

NCDs are having an increasingly negative impact on health status in populations, with disproportionately higher rates in developing countries in recent years.1 The burden of NCD is projected to increase by 2030, especially in all low and middle-income countries.2 More than 80% of major NCDs can be prevented via eradicating the common risk factors mainly tobacco use,

unhealthy diets, physical inactivity and the harmful use of alcohol.3 Rapid and substantial reduction in mortality can be achieved from the improvement of NCD risk factors in a population.4 Moreover, improving health literacy amongst the community is important in the prevention of NCD risk factors as it focuses on inclusive and equitable access to health information and encouraging health literacy helps in improving knowledge and understanding of health determinants and can motivate people to adopt

International Journal of Community Medicine and Public Health | August 2021 | Vol 8 | Issue 8 Page 3823

Lwin SMM et al. Int J Community Med Public Health. 2021 Aug;8(8):3823-3829

healthy behaviors.5,6 In Myanmar, the health literacy status was low and 31.5% of Myanmar adult people attained a satisfactory health literacy status.7 Besides, Myanmar was high in rank among South-East Asia countries in terms of the prevalence of many NCD risk factors.8 Almost every adult in Myanmar had at least one NCD risk factor and most of the risk factors were high, especially smoking, smokeless tobacco use and current alcohol drinking. Moreover, the prevalence of low consumption of fruit and vegetables, overweight and obesity, hypertension, diabetes and raised total cholesterol was high.8

Nowadays, mobile technology is increasingly used to promote health and prevent disease.9 Mobile smartphone technology is also considered as a promising communication channel that presents the achievable to improve healthcare delivery and promote behavior change amongst vulnerable populations.10 Thus, mHealth has been used effectively to promote the adoption of healthy behaviors and the reduction of unhealthy behaviors.11 The rising burden of chronic diseases calls for innovative and cost-effective chronic disease prevention and management approaches.12 Providing health messages on risk factors for NCDs plays an important role in preventing diseases.13 A predominant advantage of mHealth interventions is the fact that they can be delivered to many people in a low-cost manner and in a quite brief time.13 The majority, 90% of people living in developing countries owned a mobile phone and two-thirds of the global internet users were lived in developing countries.10 Myanmar had fortunately, 22 million internet users as of January 2020 and the number of internet users in Myanmar increased by 1.0 million (4.8%) between 2019 and 2020.14 As Viber application is one of the commonly used and user-friendly applications among Myanmar people. The present study was conducted to assess the effect of regular mobile phone health messaging on risk factors of NCDs using Viber application among the rural community in Twantay township, Yangon region. Changes in participants’ behavioral and metabolic risk factors of NCDs after intervention were evaluated by comparing with a non- intervention control group.

METHODS

Study setting and participants

A community-based quasi-experimental study was conducted among 240 participants within the ages of 30 to 60 years in two selected villages of Twantay township (one situated in the north part of township and the other one situated 25.2 km distance at south part of township) within the duration of one year (July 2019 to July 2020). The two villages were purposively selected to avoid the contamination effect of the intervention. Participants were recruited by a systematic random sampling method from the sampling frame of household, which was obtained from authorities of Twantay township before data

collection. From a chosen household, one eligible person (male or female) was selected using a random sampling among household members to include in the study. The inclusive criteria were those who owned a mobile smartphone (Android, iPhone, tablet, iPad) and no plan to change residential address within one year. If a participant had poor or no skill in using a mobile phone, he/she must have at least one caregiver who knew how to use a mobile phone. We excluded participants who were mentally ill, bed-ridden, pregnant, living only temporarily in the local area and owning only a mobile keypad phone. The intervention comprised sending health messages via the Viber application corresponding to each participant’s risk factor grouping.

Data collection and intervention

Firstly, socio-demographic data and information about behavioral risk factors of NCDs were collected among the participants using the WHO STEP survey questionnaire (version 3.0) with the face-to-face interview method. Physical examination (height, weight, BMI and blood pressure) and biochemical measurement (fasting blood sugar, total cholesterol, triglyceride, LDL and HDL cholesterol) were followed. All measurement equipment was standardized just before taking measurements and the blood samples were collected with standard procedures.15 Basic health education concerning major risk factors of NCDs was given to both the intervention group and the control group before intervention. During the intervention period, however, health messages on the reduction of their risk factors, promotion of healthy lifestyle, the benefits of physical activity and balanced diet and hazard of alcohol drinking and smoking were given only to the intervention group using the Viber application. All health text messages were developed by the standardized message book and IEC material from NCD units and the health literacy promotion unit of the ministry of health and sports, Myanmar. All the messages sent were in the most commonly used Myanmar languages. Each week, a different message was sent to the intervention participants on Friday and each message was presented two times in the year. For prevention of loss to follow up, the participants who missed reading three messages in a month were traced with a phone call. However, participants who could not be contacted and those who missed reading the health messages for two consecutive months were dropped from the study. At the end of the 12 months, information of behavioral risk factors of NCDs, physical and biochemical measurements was recollected using the same procedures as at baseline.

Statistical analysis

Data analysis was conducted using STATA version 14.2. For descriptive analysis, categorical variables were presented as frequency and percentage and continuous variables presented as and mean and standard deviation. Differences in baseline characteristics between intervention and control groups were tested using the Chi

International Journal of Community Medicine and Public Health | August 2021 | Vol 8 | Issue 8 Page 3824

Lwin SMM et al. Int J Community Med Public Health. 2021 Aug;8(8):3823-3829

participants from the control side dropped out due to inability to be contacted. Hence, of 240 participants, 189 participants completed the study and were reassessed at 12 months. No significant difference in socio- demographic characteristics between those who completed and those who were lost to follow up was observed. The baseline socio-demographic status of participants in the intervention group and control group did not differ significantly by age group, marital status or ethnic group. However, differences were evident in the distributions of sex, level of education, occupation and level of income. More female participants were included in the control group (80.8%) than the intervention group (65.8%). Participants in the intervention group were more educated than the control group and nearly half of the participants in the intervention group had received more than primary school level of education (43.3%) compared with only one-fifth (21.7%) in the control group. A higher percentage of working individuals were included in the intervention group (85.8%) than in the control group (56.7%). One-third of participants (31.7%) in the intervention group but only one-fifth (18.3%) in the control group reported a monthly per capita income within the first quartile of less than 45 thousand (MMK) (Table 1). Among current smokers after the 12 months intervention period, the number of smokes per day and the number of uses of smokeless tobacco per day were both reduced significantly in the intervention group compared to the changes in the control group (p=0.012 and p<0.001, respectively). However, no significant differences were observed in the change in the number of drinking occasions in the past 30 days or in the number of standard drinks per drinking occasion between intervention and control groups (p>0.05) (Table 2). For the factors of the number of servings of fruits and vegetables per day and physical activity (MET minute per week), no significant differences were found between the groups after adjustment for the propensity score (p>0.05) (Table 3). Regarding metabolic risk factors, no significant differences between two groups were found in the changes in systolic and diastolic blood pressure, BMI or fasting lipid profile (total cholesterol, triglyceride, LDL, and HDL) (p>0.05). However, at the end of the study, participants in the intervention group had a significant net reduction in fasting blood sugar with a mean change of 18.7 mg/dl (95% CI -32.83, -4.56, p=0.010) relative to the control group (Table 4). Table 1: Characteristics of participants in intervention and control groups (N=240).

square test, Fisher’s exact test, and independent sample t test. A two sample Wilcoxon rank-sum (Mann-Whitney) test was used to evaluate the difference between groups in the change over the 12 months for non-normally distributed variables. Changes in continuous variables were computed by subtraction of final follow up values from the baseline values. To minimize the selection bias arising from the study design, a propensity score, a statistical method for adjusting for baseline differences between study groups was constructed to adjust for potential confounding variables. Mixed-effects random- intercept linear regression modeling was used, after checking the assumptions, to analyze the changes in continuous behavioral, anthropometric and metabolic variables over the 12 months and to compare between the intervention and control groups by including an interaction term between pre/post intervention and group. The interaction term constituted the fixed effect in the model while the individual was considered to be the random element, thereby adjusting the standard errors and confidence intervals for repeated measures on the same subject. The propensity score was used as a covariate in the linear mixed-effects models to control for the potential confounding variables which could affect the estimated relationship between risk factors and outcomes. Because of the small sample size for the outcome variables of smoking, smokeless tobacco and alcohol drinking, the Wilcoxon rank-sum (Mann-Whitney) test was used for comparing mean changes between the intervention and control group. All statistical tests were evaluated using two-tailed 95% confidence intervals (CI). All analyses were carried out at a 0.05 significant level.

RESULTS

The process of the study including the number of participants at baseline and follow up is shown in Figure 1. A total of 120 participants who met the inclusion criteria were selected from each of the two selected villages. The intervention of sending health messages to each group was conducted weekly and each message was repeated two times in the year. At the end of the 12 month intervention, 32 participants from the intervention side dropped out due to limitation of their free time, loss of their phone, change of their phone number or inability to be contacted due to switched off mobile phones, while 19

Continued.

International Journal of Community Medicine and Public Health | August 2021 | Vol 8 | Issue 8 Page 3825

Lwin SMM et al. Int J Community Med Public Health. 2021 Aug;8(8):3823-3829

*Chi squared test; †Fisher’s exact test.

Table 2: Mean changes in smoking, smokeless tobacco and alcohol in both intervention and control group after the

intervention.

*Wilcoxon rank-sum (Mann-Whitney) test.

Table 3: Linear mixed-effects modeling on change in fruits and vegetable serving and physical activity before and

after intervention with propensity score adjustment.

*The p value is for the null hypothesis of no difference in the change from before to after between the 2 groups. Table 4: Linear mixed-effects modeling of anthropometric measurement and fasting lipid profile before and after

intervention in intervention and control groups, with propensity score adjustment.

*The p value is for the null hypothesis of no difference in the change from before to after between the 2 groups.

International Journal of Community Medicine and Public Health | August 2021 | Vol 8 | Issue 8 Page 3826

Lwin SMM et al. Int J Community Med Public Health. 2021 Aug;8(8):3823-3829

Figure 1: Flow chart of the study.

DISCUSSION

Using mobile health technologies for improving population-level health outcomes around the world has emerged in the last decade.16 The results of the present study show that mobile health messaging intervention can be effective in reducing the number of smoking and smokeless tobacco among current users and a reduction in fasting blood sugar levels among the study population. This concurs with the findings of a previous study from Myanmar, which reported that sending health messages via Viber application resulted in a reduction in smoking frequency among current users.17 In contrast, findings from Lebanon showed that health messaging intervention did not affect smoking.18 Besides, no significant changes were observed in reducing the frequency of alcohol drinking among the current users after the intervention in the present study, although another study in Myanmar has reported that health message intervention reduced alcohol

drinking frequency.17 It has been stated that alcohol intake was regarded as a difficult issue to be tackled by text messaging or phone conversation only, without face- to-face encounters.19

Although some studies stated positive effects of the mHealth intervention on diet behavior and physical activity in developing countries, no significant changes were observed in the present study population.10 Several other studies have also pointed out that sending health messages did not affect physical activity and this is in agreement with a study that showed that lifestyle behavior such as physical activity cannot be modified through self-care education intervention, whether via SMS or pamphlets or by face-to-face meetings.19-22 However, different findings were reported in the study from Delhi, India, where a significant reduction in behavioral risk factors (unhealthy diet and insufficient physical activity) was seen in the intervention group

International Journal of Community Medicine and Public Health | August 2021 | Vol 8 | Issue 8 Page 3827

Lwin SMM et al. Int J Community Med Public Health. 2021 Aug;8(8):3823-3829

compared with the control group.13 The possible obstacle for improvement in physical activity between the intervention and the control group in the present study might be related to the reduction in the frequency of going outside because of the concurrent outbreak of COVID-19 at the time of the second data collection. However, such an explanation requires confirmation by conducting further studies.

The study finding of a considerable reduction in fasting blood sugar level in the intervention group is in agreement with other studies.13,20 However, the intervention did not result in reducing the blood pressure among the study population. The findings were consistent with a study in Latin America, which showed that the intervention did not reduce the blood pressure but reduced systolic blood pressure in a study of Myanmar.17,19 In contrast, a study in Lebanon mentioned that sending weekly short health messages resulted in reducing systolic blood pressure.18

In the present study, the intervention did not affect body mass index (BMI). Consistent findings were reported in one study from India, whereas another study in India reported a contrasting finding of a significant reduction of BMI between the intervention group and the control group after the intervention.13,20 In addition, the intervention in the present study showed no effect on reducing total cholesterol, triglyceride or LDL levels. The reason is not clear and further studies need to be conducted for a detailed explanation.

In conclusion, the study has provided evidence that the intervention can reduce the amount of smoking and using smokeless tobacco among current users and reduce fasting blood sugar. These facts favor considering giving health messages via Viber application to the community who have risk factors of NCDs as a national policy because the incidence of risk factors of NCDs is rising around the world and particularly in developing countries including Myanmar. From the feasibility point of view, mobile phone usage and the number of internet users in Myanmar have increased in recent years. The ministry of health and sports is strongly promoting e-health development among the community and mobile phone tablets have already been distributed to basic health staff in most areas of Myanmar. Sending readymade health messages at regular intervals, in addition to in-person contact via the Viber group to many members of the community having at least one NCD risk factor, by basic health staff in their respective local areas can be an effective way to reduce the burden on providing health education in person especially during the pandemic period. The findings of the study may apply to other similar contexts to the study area.

The strengths of the present study include the geographic separation between the control and intervention group (two villages were 25.2 km apart) which would have prevented contamination of health messages. Moreover,

propensity score adjustment was used to reduce confounding by covariates that were not equally distributed at baseline. Utilization of WHO standard questionnaire and standardization of instruments minimized potential information bias. Another point to mention is that the participants in the present study can be considered to represent most of the rural population of the country because the most rural settings in Myanmar are more or less similar. As the limitations, participant’s response to behavioral risk factors was self-reported based on recall, which was prone to recall and self- desirability bias. Another limitation was the occurrence of the COVID-19 pandemic, which started between the first and second time of data collection. However, the impact would have been more or less similar in the intervention and control groups. The pandemic induced people from both groups to adopt a more sedentary lifestyle.

CONCLUSION

The study results showed that intervention affected a reduction in some behavioral and metabolic risk factors of NCDs.

ACKNOWLEDGEMENTS

We gratefully acknowledge all my supervisors and senior colleagues from the university of public health, Yangon, and Ajarn Alan Geater, lecturer, epidemiology unit, Prince Songkla university, Thailand for his constructive suggestions on this work.