Abstract
Thailand’s temporary abstinence campaign aims to persuade drinkers to abstain from alcohol for three months during Buddhist Lent. In recent years, a decline in the popularity of the campaign has been observed. This study aims to determine factors associated with success in complete abstinence during the campaign period and to determine whether the associations change over time to provide insight into the decline in complete abstinence. This study analyzes pooled data of 5898 current drinkers from three waves (2015, 2018, and 2021) of the campaign evaluation survey. The primary outcome is complete abstinence during the campaign. Multivariable analysis indicated that campaign media exposure was associated with complete abstinence (OR, 1.42; 95% CI 1.17–1.72). Similarly, the year 2018, older age, lesser drinking frequency prior to the campaign, and higher level of affordability were positively associated with complete abstinence. There was a statistically significant interaction between year and drinking frequency prior to the campaign (p < 0.001). The decline in complete abstinence was plausibly explained by reduced campaign media exposure, increased drinking frequency among drinkers, and the 2021 period effect (presumably COVID-19). Diversifying campaign media distribution across traditional, community-based, and digital platforms may enhance the campaign’s success by ensuring wider exposure to campaign messages.
Supplementary Information
The online version contains supplementary material available at 10.1038/s41598-025-33405-9.
Keywords: Alcohol, Temporary abstinence, Mass media campaign, Sobriety, Media exposure
Subject terms: Health policy, Public health
Introduction
Alcohol use is a major health risk1–3. In all comparative risk assessments to date, alcohol use had been ranked within the 10 leading causes of burden of disease4. Most recently—for the year 2021—alcohol has been ranked 10th, and has accounted for 0.8% of age-standardized female disability-adjusted life years (DALYs) and 3.9% of age-standardized male DALYs, globally2. Alcohol use is associated with non-communicable diseases (e.g., liver cirrhosis, cancers, and mental health problems), communicable diseases (e.g., tuberculosis, HIV/AIDS, and lower respiratory infections), and injuries (e.g., assault, crime, and road traffic accidents)2,5–9. According to a 2021 systematic review, in middle- and high-income countries, the cost of alcohol use and its consequences to the economy exceeded 2% of gross domestic product10. In Thailand, the setting of this study, alcohol consumption was the 3rd leading contributor of DALYs in 2019, accounting for 1.56 million DALYs (14.0% of total DALYs)11.
Mass media campaigns are employed to modify health-related behaviors to improve health. These campaigns can influence behavior through various pathways, including cognition, emotion, setting public agenda, and establishing social norms12. For alcohol use, campaigns encouraging drinkers to temporarily abstain from alcohol consumption have been conducted in several countries. For instance, the Dry January campaign has been undertaken in the UK to challenge drinkers to pause drinking in January since 201313. In Australia, the Dry July campaign—encouraging alcohol-free in July and raising funds for cancer victims—has been conducted since 200814. There is scant evidence that these temporary abstinence campaigns might be associated with alcohol harm reduction15.
The Buddhist Lent Abstinence campaign, which is among the first of its kind, has been conducted in Thailand since 2003. Taking advantage of Thailand’s predominately Buddhist population (93.5% of Thais were Buddhists according to the 2018 Social and Culture survey16, this campaign seeks to persuade drinkers to abstain from alcohol for three months during Buddhist Lent17. According to Buddhist beliefs, consuming intoxicants, including alcohol, should be avoided during this time. The period is based on the lunar calendar and typically occurs between July and October. The effectiveness of this campaign in reducing population-level alcohol consumption has been demonstrated in previous studies18,19.
The Thai abstinence campaign has been promoted through various media channels—including newspapers, television, radio, billboards, and online media. In some localities, there are also campaign activities at the community level18. This campaign is also included in the National Alcohol Strategy of Thailand20. According to the 2016 campaign evaluation, 32.2% of drinkers completely abstained during the campaign period18, making it the most significant alcohol-related campaign in Thailand. A recent time-series analysis demonstrated that the campaign has been associated with an approximately 10% decrease in population-level alcohol consumption19. Campaign activities at the community level were associated with abstinence lasting for 3 months beyond the end of Lent17.
In recent years, a decline in the popularity of the campaign has been observed. The rate of complete abstinence (i.e., abstaining from alcohol for three months during Buddhist Lent) steadily decreased from 32.3% in 2015 to 12.9% in 202121,22. The explanations for the decline remain unclear. By analyzing data from a repeated cross-sectional survey (2015, 2018, and 2021 waves), this study aims to determine factors associated with complete abstinence during Thailand’s three-month alcohol abstinence campaign and to determine whether the associations change over time. Knowledge about these can provide insight into the decline in complete abstinence. Given the emergence of temporary abstinence campaigns in several countries, this knowledge may improve campaign implementation in Thailand and elsewhere.
Methods
Study design and data source
This study analyzed data from three waves of the Buddhist Lent Abstinence Evaluation Survey, a repeated cross-sectional survey conducted annually since 2014. Data used in this study were obtained from the 2015, 2018, and 2021 waves of the survey. These three waves were selected because they captured the period of an obvious decline in the campaign and because of the comparability of questionnaire items. Data collection was carried out over a two-week period—varying from the final week of the Lent period to the first two weeks after the Lent period ended. The survey’s target population was the general population aged 15 years and older. Eligibility criteria for participants included the age of ≥ 15 years old, fluency in Thai, and being a resident of the sampled subdistricts for at least three months.
A multistage stratified sampling was used in the survey. In the first stage, the provinces of Thailand were stratified into five strata: the Bangkok metropolitan region, the central region, the northern region, the northeastern region, and the southern region. In the 2015 wave, four, five, five, seven, and four provinces were selected from each stratum, respectively. In the 2018 and 2021 waves, three, two, two, three, and two provinces were selected from each stratum, respectively. The probability of a province being selected was proportional to population sizes. In the second and third stages, districts and subdistricts were randomly selected with a probability proportional to their size. In the fourth stage, households were randomly selected from the subdistrict’s population registers. One member of a selected household who met the inclusion criteria was invited to participate in the survey. Those who provided consent were interviewed by trained interviewers21–23. Hence, informed consent was obtained from all survey participants. The total sample sizes were 7099, 3927, and 3916 for the 2015, 2018, and 2021 waves, respectively. The post-stratification weighting was not applied.
Study subjects
The present study analyzed data from current drinkers—i.e., those who consumed alcohol during the past 12 months prior to the beginning of the campaign. Survey participants who were not current drinkers were excluded. This study analyzed data of 2657, 1670, and 1571 participants from the 2015, 2018, and 2021 waves, respectively. In total, data from 5898 participants were included in the analysis. The diagram of participant selection is shown in Fig. 1.
Fig. 1.
Participant selection in the analysis.
Data management
The primary outcome was complete abstinence during Buddhist Lent. The questionnaire item for complete abstinence was, “From the beginning of the Buddhist Lent period (3 months) this year until now, have you abstained from drinking alcohol?” The response options included: abstained for three months, abstained for a certain period, reduced the number of drinks per occasion, and continued drinking as usual. As the survey was conducted at the end of the Lent period or shortly thereafter, this item captured drinking behavior during the three-month Buddhist Lent. A dichotomous complete abstinence variable was created based on these responses. The response, abstained for three months, was classified as ‘Success’; other responses were classified as ‘No success.’ As it is a nationwide campaign, either complete abstinence or reduction in alcohol consumption during the Buddhist Lent period can be considered participation in the campaign. In a sensitivity analysis, an ordinal variable—campaign participation—based on original responses (abstained for three months > abstained for a certain period > reduced the number of drinks per occasion > continued drinking as usual) was used. This sensitivity analysis was conducted to determine whether the classification of the outcome variable affected the regression results.
Based on previous studies18,24,25, the other variables included in the analysis were demographic characteristics, drinking frequency, affordability, and campaign-related variables. The demographic variables included sex (male and female), age (15–19, 20–30, 31–45, 46–60, and 61 years or higher), educational attainment (primary school or lower [Grade 6 or lower], secondary school [Grade 7–12 ], and college or higher), and religion (Buddhism and others).
Drinking frequency prior to the campaign (i.e., an average frequency in the past 12 months) was measured as weekly, monthly, or occasionally (less than once a month). The affordability variable was created by dividing self-reported monthly income (in Thai baht) by an average self-reported drinking expense per occasion (in Thai baht) and then classifying the resulting values into tertiles (the first tertile indicates the lowest level of affordability). This variable was included to account for potential changes in alcohol purchasing power across the three survey periods. The campaign-related variable was exposure to the campaign media (yes and no).
Statistical analysis
Continuous variables were described using means and standard deviations, while categorical variables were summarized using counts and percentages. To determine factors associated with success in complete abstinence during 2015 to 2021, firstly, data from three waves of the survey were pooled. The year of the suvey was included as a variable in the pooled dataset.
A multivariable analysis was performed using a binary logistic regression model. Complete abstinence (dichotomous) was a dependent variable in the regression model. Demographic characteristics, drinking-related variables, campaign-related variables, and year of survey were covariates in the model. Adjusted odds ratios (OR) with 95% confidence intervals (CI) and p-values were estimated to indicate a magnitude of association between each factor and success in complete abstinence. The sensitivity analysis was carried out by replacing the dichotomous complete abstinence variable with the ordinal campaign participation variable in the regression model, and ordinal regression was performed instead of the binary logistic regression. In the ordinal regression model, ORs greater than 1.00 indicate higher odds of shifting toward complete abstinence (the most intensive level of campaign participation). An interaction between year and each factor associated with success in complete abstinence (from the logistic regression model) was tested to determine whether an association between each covariate and complete abstinence varied by time. Of 5898 participants, 543 were excluded from the regression analysis due to missing data on one or more covariates.
The R software version 4.0.3 was used to perform the statistical analysis26. R packages used in data management and data analysis included MASS27, psych28, and epicalc29. The significance level of 0.05 was employed in all analyses.
Results
Characteristics of drinkers
Characteristics of drinkers from three waves of the survey are shown in Table 1. The majority of drinkers were male. The mean age was between 37.9 and 39.9 years old. More than 40.0% of drinkers had secondary education. Nearly all participants were Buddhists. The prevalence of drinkers who drank at least once a week (regular drinkers) increased from 30.5% in 2015 to approximately 50% in 2018 and 2021. The percentage of drinkers exposed to the campaign media was decreasing. Regarding drinking behavior during the campaign period, the percentage of complete abstinence reduced from 32.3% in 2015 to 30.2% in 2018 and to 12.9% in 2021. A more detailed examination of drinking during the Lent period unveiled a considerable reduction in the percentage of drinkers who abstained for a certain period between 2015 and 2018. Concurrently, there was an observed increase in those adhering to their usual drinking behavior during the Lent period from 2015 to 2021.
Table 1.
Characteristics of drinking survey participants across three survey waves (2015, 2018, and 2021).
| Characteristics | 2015 | 2018 | 2021 | |||
|---|---|---|---|---|---|---|
| n | % | n | % | n | % | |
| Total | 2,657 | 100.0 | 1,670 | 100.0 | 1,571 | 100.0 |
| Sex | ||||||
| Male | 1,815 | 68.3 | 1,159 | 69.4 | 1,016 | 64.7 |
| Female | 842 | 31.7 | 511 | 30.6 | 555 | 35.3 |
| Age (years) | ||||||
| Mean (SD) | 37.9 (13.2) | 39.5 (12.4) | 39.9 (13.5) | |||
| 15–19 | 190 | 7.2 | 62 | 3.7 | 69 | 4.4 |
| 20–30 | 700 | 26.4 | 430 | 25.8 | 412 | 26.2 |
| 31–45 | 975 | 36.8 | 573 | 34.5 | 525 | 33.4 |
| 46–60 | 656 | 24.7 | 561 | 33.7 | 463 | 29.5 |
| ≥61 | 132 | 4.9 | 38 | 2.3 | 102 | 6.5 |
| Educational attainment | ||||||
| Primary school or lower | 799 | 30.3 | 219 | 13.2 | 327 | 20.8 |
| Secondary school | 1,062 | 40.3 | 893 | 53.6 | 794 | 50.6 |
| College or higher | 776 | 29.4 | 553 | 33.2 | 448 | 28.6 |
| Religion | ||||||
| Buddhism | 2,589 | 99.0 | 1,653 | 99.0 | 1,527 | 97.7 |
| Others | 26 | 1.0 | 17 | 1.2 | 36 | 2.3 |
| Drinking frequency prior to the campaign | ||||||
| Weekly | 803 | 30.5 | 878 | 52.6 | 752 | 48.0 |
| Monthly | 1,063 | 40.4 | 417 | 25.0 | 465 | 29.6 |
| Occasionally | 767 | 29.1 | 374 | 22.4 | 351 | 22.4 |
| Affordability | ||||||
| First tertile | 1,178 | 52.0 | 605 | 37.1 | 458 | 30.0 |
| Second tertile | 570 | 25.2 | 617 | 37.9 | 479 | 31.3 |
| Third tertile | 516 | 22.8 | 408 | 25.0 | 591 | 38.7 |
| Exposure to campaign media | ||||||
| No | 228 | 8.6 | 343 | 20.5 | 407 | 25.9 |
| Yes | 2,429 | 91.4 | 1,327 | 79.5 | 1,164 | 74.1 |
| Complete abstinence | ||||||
| No | 1,800 | 67.7 | 1,165 | 69.8 | 1,368 | 87.1 |
| Yes | 857 | 32.3 | 505 | 30.2 | 203 | 12.9 |
| Drinking during the Lent period | ||||||
| Continued drinking as usual | 774 | 29.1 | 746 | 44.7 | 1,095 | 69.7 |
| Reduced the number of drinks per occasion | 429 | 16.1 | 255 | 15.3 | 174 | 11.1 |
| Abstained for a certain period | 597 | 22.5 | 164 | 9.8 | 99 | 6.3 |
| Abstained for three months | 857 | 32.3 | 505 | 30.2 | 203 | 12.9 |
SD standard deviation.
Data source: Buddhist Lent Abstinence Evaluation Survey 2015, 2018, and 2021.
Prevalence of success in complete abstinence during the campaign period by characteristics
The prevalence of complete abstinence during the campaign by characteristics is shown in Table 2. Females had higher percentages of abstinence than males across three waves of the survey. Half of the occasional drinkers completely abstained in 2015 and 2018; the figure decreased to only 17.7% in 2021. Drinkers who were exposed to the campaign media had a significantly higher prevalence of complete abstinence than those who did not in 2015 and 2018. Complete abstinence decreased across all categories in 2021.
Table 2.
Percentage of success in complete abstinence during the campaign period by participant characteristics across three survey waves (2015, 2018, and 2021).
| Characteristics | 2015 | 2018 | 2021 | ||||||
|---|---|---|---|---|---|---|---|---|---|
| % | 95%CI | P-value | % | 95%CI | P-value | % | 95%CI | P-value | |
| Overall | 32.3 | 30.5, 34.1 | - | 30.2 | 28.1, 32.5 | - | 12.9 | 11.4, 14. | - |
| Sex | < 0.001 | 0.015 | < 0.001 | ||||||
| Male | 28.0 | 26.0, 30.1 | 28.4 | 25.9, 31.1 | 10.3 | 8.6, 12.4 | |||
| Female | 41.4 | 38.2, 44.8 | 34.4 | 30.5, 38.7 | 17.7 | 14.7, 21.0 | |||
| Age (years) | 0.422 | < 0.001 | 0.256 | ||||||
| 15–19 | 31.6 | 25.4, 38.5 | 24.2 | 15.2, 36.2 | 14.5 | 8.1, 24.7 | |||
| 20–30 | 30.7 | 27.4, 34.2 | 23.5 | 19.7, 27.7 | 9.7 | 7.2, 13.0 | |||
| 31–45 | 31.5 | 28.6, 34.5 | 34.0 | 30.3, 39.0 | 14.3 | 11.6, 17.5 | |||
| 46–60 | 34.0 | 30.5, 37.7 | 30.1 | 26.5, 34.0 | 14.0 | 11.2, 17.5 | |||
| ≥61 | 37.9 | 30.1, 46.4 | 57.9 | 42.2, 72.1 | 12.7 | 7.6, 20.6 | |||
| Educational attainment | 0.160 | 0.486 | 0.710 | ||||||
| Primary school or lower | 29.5 | 26.5, 32.8 | 29.2 | 23.6, 35.6 | 14.1 | 10.7, 18.3 | |||
| Secondary school | 32.9 | 30.1, 35.7 | 29.3 | 26.4, 32.4 | 12.8 | 10.7, 15.4 | |||
| College or higher | 33.8 | 30.5, 37.2 | 32.2 | 28.4, 36.2 | 12.1 | 9.4, 15.4 | |||
| Religion | 0.999 | 0.384 | 0.999 | ||||||
| Others | 30.8 | 16.5, 50.0 | 17.6 | 6.2, 41.0 | 13.9 | 6.1, 28.7 | |||
| Buddhism | 32.4 | 30.6, 34.2 | 30.4 | 28.2, 32.6 | 12.9 | 11.3, 14.7 | |||
| Drinking frequency prior to the campaign | < 0.001 | < 0.001 | < 0.001 | ||||||
| Weekly | 16.9 | 14.5, 19.7 | 18.6 | 16.1, 21.3 | 7.6 | 5.9, 9.7 | |||
| Monthly | 29.7 | 27.1, 32.5 | 31.4 | 27.1, 36.0 | 17.6 | 14.4, 21.4 | |||
| Occasionally | 50.8 | 47.3, 54.4 | 56.2 | 51.1, 61.1 | 17.7 | 14.0, 22.0 | |||
| Affordability | 0.003 | 0.269 | 0.787 | ||||||
| First tertile | 27.8 | 25.3, 30.4 | 27.6 | 24.2, 31.3 | 12.4 | 9.7, 15.8 | |||
| Second tertile | 31.4 | 27.7, 35.3 | 30.6 | 27.1, 34.4 | 12.5 | 9.9, 15.8 | |||
| Third tertile | 35.9 | 31.8, 40.1 | 32.1 | 27.8, 36.8 | 13.7 | 11.2, 16.7 | |||
| Exposure to campaign media | 0.002 | < 0.001 | 0.720 | ||||||
| No | 22.8 | 17.8, 28.7 | 21.9 | 17.8, 26.5 | 12.3 | 9.4, 15.8 | |||
| Yes | 33.1 | 31.3, 35.0 | 32.4 | 29.9, 35.0 | 13.1 | 11.3, 15.2 | |||
Note. Complete abstinence, completely abstained from drinking for three months during the Buddhist Lent period; 95%CI, 95% confidence interval.
Data source: Buddhist Lent Abstinence Evaluation Survey 2015, 2018, and 2021.
Factors associated with success in complete abstinence
The results of the multivariable analysis are shown in Table 3. The year of the survey was associated with complete abstinence. Contrary to the unadjusted analyses, compared to 2015, the year 2018 was associated with 21% higher odds of complete abstinence (OR 1.21; 95% CI 1.04–1.42), and the year 2021 was associated with 64% lower odds of complete abstinence (OR 0.36; 95% CI 0.30–0.44). The higher odds of complete abstinence in 2018, which contradicted the crude prevalence of complete abstinence (Table 2), were largely attributable to adjustment for drinking frequency (see Supplementary Table S1). Females had 14% higher odds of complete abstinence than males; however, the association was borderline and was not statistically significant (OR 1.14; 95% CI 0.99–1.31). The odds of complete abstinence tended to increase with age. Lower frequency of drinking prior to the campaign period was associated with higher odds of complete abstinence. A higher level of affordability was associated with higher odds of complete abstinence. Exposure to the campaign media was associated with 42% higher odds of complete abstinence (OR 1.42; 95% CI 1.17–1.72).
Table 3.
Factors associated with success in complete abstinence during the campaign period (multivariable logistic regression; pooled data from three survey waves).
| Factors | AOR | 95%CI | p-value | |
|---|---|---|---|---|
| Wald’s test | LR-test | |||
| Year | < 0.001 | |||
| 2015 | 1 | |||
| 2018 | 1.21 | 1.04, 1.42 | 0.015 | |
| 2021 | 0.36 | 0.30, 0.44 | < 0.001 | |
| Sex | 0.076 | |||
| Male | 1 | |||
| Female | 1.14 | 0.99, 1.31 | 0.075 | |
| Age (years) | < 0.001 | |||
| 15–19 | 1 | |||
| 20–30 | 0.98 | 0.71, 1.36 | 0.908 | |
| 31–45 | 1.39 | 1.01, 1.91 | 0.041 | |
| 46–60 | 1.36 | 0.98, 1.88 | 0.066 | |
| ≥61 | 2.37 | 1.55, 3.62 | < 0.001 | |
| Educational attainment | 0.555 | |||
| Primary school or lower | 1 | |||
| Secondary school | 1.09 | 0.91, 1.29 | 0.356 | |
| College or higher | 1.02 | 0.84, 1.23 | 0.869 | |
| Religion | 0.308 | |||
| Others | 1 | |||
| Buddhism | 1.39 | 0.73, 2.64 | 0.320 | |
| Drinking frequency prior to the campaign | < 0.001 | |||
| Weekly | 1 | |||
| Monthly | 2.29 | 1.94, 2.71 | < 0.001 | |
| Occasionally | 4.74 | 3.97, 5.65 | < 0.001 | |
| Affordability | 0.003 | |||
| First tertile | 1 | |||
| Second tertile | 1.20 | 1.02, 1.40 | 0.029 | |
| Third tertile | 1.34 | 1.13, 1.58 | < 0.001 | |
| Exposure to campaign media | < 0.001 | |||
| No | 1 | |||
| Yes | 1.42 | 1.17, 1.72 | < 0.001 | |
| McFadden’s pseudo-R2: 0.101 | ||||
Note. AOR, adjusted odds ratio; 95%CI, 95% confidence interval; LR-test, likelihood ratio test.
Dependent variable: complete abstinence (dichotomous)—completely abstained from drinking for three months during the Buddhist Lent period.
Data source: pooled data from the Buddhist Lent Abstinence Evaluation Survey 2015, 2018, and 2021.
Sensitivity analysis: factors associated with campaign participation (ordinal outcome)
The results of the sensitivity analysis are shown in Table 4. The year of the survey was associated with campaign participation. Compared to 2015, the years 2018 and 2021 were associated with lower odds of shifting toward abstinence (OR 0.86; 95% CI 0.76–0.97 for 2018; OR 0.25; 95% CI 0.21–0.28 for 2021). Female sex was associated with higher odds of shifting toward abstinence (OR 1.27; 95% CI 1.13–1.42). Associations of age, drinking prior to the campaign period, and exposure to the campaign media with drinking during the Lent period were consistent with the result presented in Table 3. On the other hand, affordability was not associated with drinking during the Lent period in this analysis.
Table 4.
Factors associated with campaign participation (ordinal regression; pooled data from three survey waves).
| Factors | AOR | 95%CI | p-value | |
|---|---|---|---|---|
| Wald’s test | LR-test | |||
| Year | < 0.001 | |||
| 2015 | ||||
| 2018 | 0.86 | 0.76, 0.97 | 0.002 | |
| 2021 | 0.25 | 0.21, 0.28 | < 0.001 | |
| Sex | < 0.001 | |||
| Male | 1 | |||
| Female | 1.27 | 1.13, 1.42 | < 0.001 | |
| Age (years) | < 0.001 | |||
| 15–19 | 1 | |||
| 20–30 | 1.02 | 0.80, 1.31 | 0.859 | |
| 31–45 | 1.35 | 1.06, 1.71 | 0.016 | |
| 46–60 | 1.31 | 1.02, 1.68 | 0.033 | |
| ≥61 | 2.31 | 1.64, 3.24 | < 0.001 | |
| Educational attainment | 0.064 | |||
| Primary school or lower | 1 | |||
| Secondary school | 1.08 | 0.95, 1.24 | 0.241 | |
| College or higher | 1.14 | 0.98, 1.32 | 0.079 | |
| Religion | 0.597 | |||
| Others | 1 | |||
| Buddhism | 0.88 | 0.56, 1.39 | 0.597 | |
| Drinking frequency prior to the campaign | < 0.001 | |||
| Weekly | 1 | |||
| Monthly | 2.18 | 1.93, 2.47 | < 0.001 | |
| Occasionally | 3.29 | 2.86, 3.79 | < 0.001 | |
| Affordability | 0.402 | |||
| First tertile | 1 | |||
| Second tertile | 0.94 | 0.83, 1.07 | 0.352 | |
| Third tertile | 0.97 | 0.85, 1.11 | 0.645 | |
| Exposure to campaign media | < 0.001 | |||
| No | 1 | |||
| Yes | 1.40 | 1.21, 1.62 | < 0.001 | |
| McFadden’s pseudo-R2: 0.072 | ||||
Note. AOR, adjusted odds ratio; 95% CI, 95% confidence interval; LR-test, likelihood ratio test.
Dependent variable: campaign participation (ordinal)—complete abstinence > abstained from drinking for a certain period > reduced the number of drinks per occasion > continued drinking as usual.
Data source: pooled data from the Buddhist Lent Abstinence Evaluation Survey 2015, 2018, and 2021.
The discrepancy between the results in Tables 3 and 4 was because the two models examine factors associated with different outcomes. The model in Table 3 determined factors associated with success in complete abstinence; drinking behaviors other than complete abstinence (including reduced drinking) were classified as no success. On the other hand, the model in Table 4 determined factors associated with campaign participation (ordinal outcome), in which both reduced drinking and complete abstinence were classified as forms of participation, with reduced drinking regarded as a lower level of participation than complete abstinence.
Interaction between year and factors associated with complete abstinence
The interactions between year and factors associated with complete abstinence are presented in Table 5. There was a significant interaction between year and drinking frequency prior to the campaign. The odds of complete abstinence were lower for occasional drinkers in 2021 compared to 2015 (OR 0.51; 95% CI 0.32–0.82).
Table 5.
Interaction between year and factors associated with complete abstinence (multivariable logistic regression; pooled data from three survey waves).
| Factors | AOR | 95%CI | p-value | |
|---|---|---|---|---|
| Wald’s test | LR-test | |||
| Interaction: Year*Age (years)† | 0.051 | |||
| 2018: 20–30 | 0.78 | 0.36, 1.72 | 0.543 | |
| 2018: 31–45 | 1.30 | 0.61, 2.80 | 0.500 | |
| 2018: 46–60 | 1.15 | 0.53, 2.51 | 0.719 | |
| 2018: ≥61 | 2.85 | 0.95, 8.50 | 0.060 | |
| 2021: 20–30 | 0.59 | 0.24, 1.45 | 0.253 | |
| 2021: 31–45 | 0.96 | 0.41, 2.28 | 0.930 | |
| 2021: 46–60 | 1.03 | 0.43, 2.48 | 0.951 | |
| 2021: ≥61 | 0.91 | 0.30, 2.72 | 0.865 | |
| McFadden’s pseudo-R2: 0.104 | ||||
| Interaction: Year*Drinking frequency prior to the campaign† | < 0.001 | |||
| 2018: Monthly | 0.95 | 0.65, 1.37 | 0.770 | |
| 2018: Occasionally | 1.02 | 0.70, 1.49 | 0.926 | |
| 2021: Monthly | 1.27 | 0.82, 1.99 | 0.289 | |
| 2021: Occasionally | 0.51 | 0.32, 0.82 | 0.005 | |
| McFadden’s pseudo-R2: 0.105 | ||||
| Interaction: Year*Affordability† | 0.310 | |||
| 2018: Second tertile | 1.01 | 0.71, 1.42 | 0.977 | |
| 2018: Third tertile | 0.72 | 0.49, 1.04 | 0.081 | |
| 2021: Second tertile | 0.81 | 0.51, 1.29 | 0.381 | |
| 2021: Third tertile | 0.81 | 0.52, 1.26 | 0.358 | |
| McFadden’s pseudo-R2: 0.102 | ||||
| Interaction: Year*Exposure to campaign media† | 0.120 | |||
| 2018: Yes | 0.98 | 0.61, 1.57 | 0.922 | |
| 2021: Yes | 0.63 | 0.38, 1.05 | 0.073 | |
| McFadden’s pseudo-R2: 0.102 | ||||
Note. AOR, adjusted odds ratio; 95% CI, 95% confidence interval; LR-test, likelihood ratio test.
†Adjusted for year, sex, age, education, religion, drinking frequency prior to the campaign, affordability, and exposure to campaign media.
Dependent variable: complete abstinence (dichotomous)—completely abstained from drinking for three months during the Buddhist Lent period.
Data source: pooled data from the Buddhist Lent Abstinence Evaluation Survey 2015, 2018, and 2021.
Discussion
The analysis of three waves of the Buddhist Lent Abstinence Evaluation Survey revealed that the year 2018, older age, lesser drinking frequency prior to the campaign, higher level of affordability, and campaign media exposure were positively associated with success in complete abstinence. Occasional drinkers in 2021 were less likely to abstain completely compared to occasional drinkers in 2015.
After adjusting for several factors, the year variable still had a significant association with complete abstinence. The year variable in the regression model can represent a combination of factors not accounted for in the model. The most prominent event of 2021 is probably the COVID-19 pandemic. The impact of COVID-19 on alcohol consumption was complex. Globally, consumption decreased, for instance, by 8% from 2019 to 202030. However, even in countries where the level of alcohol use decreased, all systematic reviews showed that some groups increased their consumption30–34, often associated with stress or mental problems35–37. In 2021, the main slogan of the Thai abstinence campaign was “Stop drinking, stay safe from COVID,” and campaign participants were encouraged to express their commitment to abstinence via an online platform38. Hence, the replacement of on-ground campaign activities with the usage of online platforms may contribute to the decrease in complete abstinence in 2021.
A change in level of religiosity during the study period may also contribute to the effect of the year variable. Evidence from the Health, Aging, and Retirement in Thailand survey—a longitudinal study of Thai citizens aged 45 years and older and their spouses—suggests a mixed pattern39. The proportion of respondents who reported frequent prayer increased from 46.9% in 2015 to 53.7% in 2020 and slightly decreased to 51.7% in 2022. The proportion of those engaging in merit-making as part of religious observance increased from 40.8% in 2015 to 53.3% in 2020, then considerably declined to 41.8% in 2022. The proportion of those observing Buddhist teachings showed a significant decline, from 61.9% in 2015 to 29.5% in 2020 and to 22.2% in 2022. As refraining from alcohol is one of the Buddhist precepts (i.e., basic moral guidelines), the decline in the level of observing Buddhist teachings may partly account for the effect of the year variable.
As this campaign can be considered faith-based, one explanation for the association between age and complete abstinence may also be related to religiosity. A recent Gallup survey suggests that older age groups are more likely to identify themselves as being religious or spiritual40. A structural equation modeling study found that age was positively related to religiosity, and religiosity was associated with a healthy lifestyle41. Religious service attendance was associated with less consumption of alcohol and cigarettes42. The Health, Aging, and Retirement in Thailand survey in 2020 and 2022 showed that respondents aged 60–79 years generally had higher rates of participation in religious activities than those aged 45–59 years39,43. Hence, older participants in the present study may also be more religious. Those with stronger religious identities may be more likely to adhere to the conduct suggested in this faith-based campaign and have a tendency toward a healthier lifestyle.
Drinking behavior prior to the campaign period has been consistently shown to be a predictor of campaign completion (completely abstaining from alcohol throughout the campaign period). Two prospective studies of Dry January in the UK found that the frequency of drunkenness in the month preceding Dry January was negatively associated with successful abstinence24,44. A study of the 2016 Thai abstinence campaign found that a lower drinking frequency prior to the campaign was associated with alcohol abstinence during the campaign period18. Baseline drinking behavior has also been identified as a predictor of success in alcoholism treatment44. Hence, drinking behavior prior to the attempt to reduce consumption is a strong predictor of the success of the attempt. It should be noted that while both Dry January and the Thai abstinence campaign aim to temporarily reduce alcohol consumption, they are embedded in distinct cultural and social contexts. Dry January emphasizes individual health improvements and social media support45,46. The Thai campaign is rooted in Buddhist tradition and is more collective in nature, reflecting conformity to community moral norms18.
A considerable amount of evidence demonstrates a consistent relationship between affordability and alcohol consumption and alcohol-related problems47–50. However, the association between the level of affordability and success of an alcohol reduction attempt, as our finding suggests, is still unexplored in other alcohol studies. The observed association in this study may suggest that the level of affordability serves as an indicator for participant socioeconomic status (SES). Evidence from a smoking cessation study showed that smoking cessation rates were consistently greater among those in higher SES categories, as determined by various SES indicators51. Hence, the relationship between affordability along with other SES indicators and attempts to reduce or quit alcohol may be an area for further investigation.
Our findings can shed light on factors plausibly contributing to the decline in campaign success (complete abstinence) from 2015 to 2021. The first factor was a steady decline in exposure to the campaign media over the period—from 91.4% in 2015 to 79.5% in 2018 to 74.1% in 2021. Lesser exposure to traditional media (such as television), where campaign media is usually broadcast, may have contributed to this decline in media exposure—particularly among younger age groups in urban areas52. Meanwhile, there was a higher percentage of drinkers who drank more frequently (who were less likely to abstain). Moreover, occasional drinkers who had high odds of complete abstinence (Tables 3 and 4) became less likely to completely abstain in 2021, as shown by the interaction model (Table 5). Furthermore, the decline in complete abstinence became more prominent in 2021, presumably as a result of the pandemic, as discussed earlier.
Strength and limitations
The primary strength of this study is that it used data from three waves of a repeated nationally representative survey spanning seven years. A limitation of this study is the self-reported nature of drinking behaviors. The assessment of drinking behavior during the campaign is based on self-report, which may be affected by social desirability bias—i.e., survey participants may state they drink less than they actually do, as it is more socially desirable. Additionally, the drinking frequency item and the exposure to the campaign media item were susceptible to recall bias. The drinking frequency item referred to a 12-month period, which is a relatively long timeframe. Furthermore, participants who did not pay attention to the campaign might not have accurately remembered the exposure to the related media.
Another important issue is the classification of the outcome variables. A comparison between the main analysis (which used a dichotomous outcome) and the sensitivity analysis (which used an ordinal outcome) revealed differences in the direction of associations between certain covariates and the outcome. Nevertheless, each classification serves different purposes. When the emphasis is on success in complete abstinence and other categories of alcohol use—including reduced consumption—are regarded as no success, the dichotomous outcome is preferred. On the other hand, if categories of reduced drinking other than complete abstinence are regarded as partial success, the ordinal outcome is more appropriate. Moreover, there are remaining potential confounding factors that were not captured by the surveys. These factors encompass alcohol marketing campaigns and the prices of alcoholic beverages, among others.
Conclusions
Year, age, drinking frequency prior to the campaign, affordability, and exposure to the campaign media were identified as factors associated with success in complete abstinence. During 2015 and 2021, a decrease in exposure to campaign media and an increase in the proportion of drinkers who drank more frequently were plausibly responsible for the decline in complete abstinence. A period effect in 2021, likely linked to COVID-19, may have contributed to the decline as well. To enhance the campaign’s success, campaign media distribution should be diversified across traditional, community-based, and digital platforms to better reach younger, urban populations and ensure wider exposure to campaign messages. To investigate the role of religiosity, the decline in abstinence among occasional drinkers, and the influence of other unmeasured factors, a prospective study guided by psychological frameworks, such as the Theory of Planned Behavior, is recommended for future research.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
The authors would like to thank the Center for Alcohol Studies for providing access to the data source used in this study.
Author contributions
P.S.: Conceptualization, Methodology, Data curation, Validation, Formal analysis, Writing - Original Draft. N.S.: Conceptualization, Formal analysis, Writing - Original Draft. M.S.P.: Conceptualization, Formal analysis, Writing - Original Draft. B.S.: Supervision and Manuscript improvement. J.R.: Supervision and Manuscript improvement. U.S.: Conceptualization, Methodology, Funding acquisition, Supervision, Validation, Writing-Review and Editing, Project administration.
Funding
This work was supported by the Center for Alcohol Studies, Thailand [Grant number 61-02029-0104, 61-02029-0113, and 65-10068-12] and the Thai Health Promotion Foundation [Contract number 67-E1-0081].
Data availability
Authors obtained the data used in this study from the Center for Alcohol Studies under the terms specific to this research. Requests for the data can be made to the corresponding author ( [saengow.udomsak@gmail.com](mailto: saengow.udomsak@gmail.com) ). Approval of access to the data must be granted by the Center for Alcohol Studies.
Declarations
Competing interests
The authors declare no competing interests.
Ethical approval
This study protocol was approved by the Human Research EthicsCommittee of Walailak University, Thailand (WU-EC-MD-3-090-65). Thedata were analyzed anonymously. This study was conducted in accordance with the Declaration of Helsinki.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Brauer, M. et al. Global burden and strength of evidence for 88 risk factors in 204 countries and 811 subnational locations, 1990–2021: a systematic analysis for the global burden of disease study 2021. Lancet403, 2162–2203 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.World Health Organization. Global status report on alcohol and health and treatment of substance use disorders. Geneva: World Health Organization (2024). https://www.who.int/publications/i/item/9789240096745
- 3.Shield, K. et al. National, regional, and global burdens of disease from 2000 to 2016 attributable to alcohol use: a comparative risk assessment study. Lancet Public. Health. 5, e51–e61 (2020). [DOI] [PubMed] [Google Scholar]
- 4.Rehm, J. & Imtiaz, S. A narrative review of alcohol consumption as a risk factor for global burden of disease. Subst. Abuse Treat. Prev. Policy. 11, 37 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Chen, C. M. & Yoon, Y. H. Usual alcohol consumption and risks for nonfatal fall injuries in the united states: results from the 2004–2013 National health interview survey. Subst. Use Misuse. 52, 1120–1132 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rehm, J. et al. Alcohol and noncommunicable disease risk. Curr. Addict. Rep.5, 72–85 (2018). [Google Scholar]
- 7.Rumgay, H. et al. Global burden of cancer in 2020 attributable to alcohol consumption: a population-based study. Lancet Oncol.22, 1071–1080 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Saengow, U., Assanangkornchai, S. & Casswell, S. Alcohol: a probable risk factor of COVID-19 severity. Addiction116, 204–205 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bryazka, D. et al. Population-level risks of alcohol consumption by amount, geography, age, sex, and year: a systematic analysis for the global burden of disease study 2020. Lancet400, 185–235 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Manthey, J. et al. What are the economic costs to society attributable to alcohol use? A systematic review and modelling study. Pharmacoeconomics39, 809–822 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.International Health Policy Program. Burden of Disease Attributable to Risk Factors in Thailand 2019 (2023). https://www.trc.or.th/en/burden-of-disease-attributable-to-risk-factors-in-thailand-2019
- 12.Wakefield, M. A., Loken, B. & Hornik, R. C. Use of mass media campaigns to change health behaviour. Lancet376, 1261–1271 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Dry January. About Dry January. Alcohol Change UK. (2022). https://alcoholchange.org.uk/help-and-support/managing-your-drinking/dry-january
- 14.Dry July Foundation. Impact Reports. Impact Reports (2022). https://www.dryjuly.com/about/impact-reports (2022).
- 15.de Ternay, J. et al. One-month alcohol abstinence National campaigns: a scoping review of the harm reduction benefits. Harm Reduct. J.19, 24 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.National Statistical Office. Statistics on Religion, Art, and Culture (2024). http://statbbi.nso.go.th/staticreport/page/sector/th/04.aspx
- 17.Jirarattanasopha, V., Witvorapong, N. & Hanvoravongchai, P. Impact of Buddhist lent dry campaign on alcohol consumption behaviour: A community level study. Health Soc. Care Community. 27, 863–870 (2019). [DOI] [PubMed] [Google Scholar]
- 18.Saengow, U. Drinking abstinence during a 3-month abstinence campaign in Thailand: weighted analysis of a National representative survey. BMC Public. Health. 19, 1688 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Saengow, U. et al. The effect of an annual temporary abstinence campaign on population-level alcohol consumption in Thailand: A time-series analysis of 23 years. BMJ Global Health9, (2024). [DOI] [PMC free article] [PubMed]
- 20.Saengow, U. Thailand’s National alcohol survey as a policy monitoring tool. Drugs Alcohol Today. 17, 250–257 (2017). [Google Scholar]
- 21.Research Centre for Social and Business Development. The evaluation of Buddhist Lent Abstinence campaign 2015 (2015). https://cas.or.th/content?id=395
- 22.Research Centre for Social and Business Development. The evaluation of Buddhist Lent Abstinence campaign 2021 (2021). https://cas.or.th/content?id=598
- 23.Research Centre for Social and Business Development. The evaluation of Buddhist Lent Abstinence campaign 2018. (2018). https://cas.or.th/content?id=472
- 24.de Visser, R. O., Robinson, E. & Bond, R. Voluntary temporary abstinence from alcohol during ‘Dry january’ and subsequent alcohol use. Health Psychol.35, 281–289 (2016). [DOI] [PubMed] [Google Scholar]
- 25.Beard, E., Brown, J., Kaner, E., West, R. & Michie, S. Predictors of and reasons for attempts to reduce alcohol intake: A population survey of adults in England. PLoS One. 12, e0173458 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Posit team. RStudio: Integrated Development Environment for R (Posit Software, PBC, 2024). http://www.posit.co/
- 27.Venables, W. N. & Ripley, B. D. Modern Applied Statistics with S Fourth Edition. (Springer, 2002). ISBN 0-387-95457-0.
- 28.Revelle, W. & Psych Procedures for Psychological, Psychometric, and Personality Research. Northwestern University, Evanston, Illinois. R package version 2.4.6. (2024). https://CRAN.R-project.org/package=psych
- 29.Chongsuvivatwong, V. Epidemiological Calculator. R package version 3.5.1.7. (2019). http://medipe.psu.ac.th/epicalc/
- 30.Acuff, S. F., Strickland, J. C., Tucker, J. A. & Murphy, J. G. Changes in alcohol use during COVID-19 and associations with contextual and individual difference variables: A systematic review and meta-analysis. Psychol. Addict. Behav.36, 1–19 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kilian, C. et al. Changes in alcohol use during the COVID-19 pandemic in europe: A meta-analysis of observational studies. Drug Alcohol Rev.41, 918–931 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Roberts, A. et al. Alcohol and other substance use during the COVID-19 pandemic: A systematic review. Drug Alcohol Depend.229, 109150 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Schmidt, R. A. et al. The early impact of COVID-19 on the incidence, prevalence, and severity of alcohol use and other drugs: A systematic review. Drug Alcohol Depend.228, 109065 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Sohi, I. et al. Changes in alcohol use during the COVID-19 pandemic and previous pandemics: A systematic review. Alcohol Clin. Exp. Res.46, 498–513 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Srinagesh, A., Forthal, S., Madden, S. P., Stein, L. & Muench, F. Impacts of COVID-19 on alcohol use among help-seeking adults. Adv. Drug Alcohol Res.3, 11159 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Grossman, E. R., Benjamin-Neelon, S. E. & Sonnenschein, S. Alcohol consumption during the COVID-19 pandemic: A Cross-Sectional survey of US adults. Int. J. Environ. Res. Public. Health. 17, 9189 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Koopmann, A. et al. The effects of the lockdown during the COVID-19 pandemic on alcohol and tobacco consumption behavior in Germany. Eur. Addict. Res.27, 242–256 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Ministry of Public Health, Thailand & Hfocus Ministry of Public Health urges Thais to ‘Stop drinking, stay safe from COVID’ on National No Alcohol Day. (2021). https://www.hfocus.org/content/2021/07/22176
- 39.Anantanasuwong, D. et al. Health Aging, and Retirement in Thailand (HART): Wave 4. (2022). https://hart.nida.ac.th/reports-whitepapers
- 40.Jones, J. M. In U.S., 47% Identify as Religious, 33% as Spiritual (2023). https://news.gallup.com/poll/511133/identify-religious-spiritual.aspx
- 41.Abdala, G. A. et al. Religion, Age, Education, Lifestyle, and health: structural equation modeling. J. Relig. Health. 60, 517–528 (2021). [DOI] [PubMed] [Google Scholar]
- 42.Pawlikowski, J., Białowolski, P., Węziak-Białowolska, D. & VanderWeele, T. J. Religious service attendance, health behaviors and well-being-an outcome-wide longitudinal analysis. Eur. J. Public. Health. 29, 1177–1183 (2019). [DOI] [PubMed] [Google Scholar]
- 43.Anantanasuwong, D. et al. Health Aging, and Retirement in Thailand (HART): Wave 3 (2020). https://hart.nida.ac.th/reports-whitepapers
- 44.Staines, G. et al. Predictors of drinking outcomes among alcoholics. Am. J. Drug Alcohol Abuse. 29, 203–218 (2003). [DOI] [PubMed] [Google Scholar]
- 45.Yeomans, H. New Year, new you: a qualitative study of dry January, self-formation and positive regulation. Drug-Educ Prev. Polic.26, 460–468 (2019). [Google Scholar]
- 46.Case, P. et al. Has the increased participation in the National campaign ‘Dry january’ been associated with cutting down alcohol consumption in england? Drug Alcohol Depend.227, 108938 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Jiang, H. & Livingston, M. The dynamic effects of changes in prices and affordability on alcohol consumption: an impulse response analysis. Alcohol Alcohol. 50, 631–638 (2015). [DOI] [PubMed] [Google Scholar]
- 48.Lintonen, T. & Nevalainen, J. Has the role of personal income in alcohol drinking among teenagers changed between 1983 and 2013: a series of nationally representative surveys in Finland. BMJ Open.7, e013994 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Moore, S. C. et al. Alcohol affordability: implications for alcohol price policies. A cross-sectional analysis in middle and older adults from UK biobank. J. Public. Health. 44, e192–e202 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Razvodovsky, Y. E. Affordability of alcohol and alcohol-related mortality in Belarus. Adicciones25, 156–162 (2013). [PubMed] [Google Scholar]
- 51.Hiscock, R., Dobbie, F. & Bauld, L. Smoking cessation and socioeconomic status: An update of existing evidence from a National evaluation of english stop smoking services. Biomed. Res. Int.2015, 274056 (2015). [DOI] [PMC free article] [PubMed]
- 52.McKenzie, J. et al. Digital Divides, generational Gaps, and cultural overlaps: A portrait of media use and perspectives of media in Thailand. Media Psychol.27, 106–134 (2024). [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Authors obtained the data used in this study from the Center for Alcohol Studies under the terms specific to this research. Requests for the data can be made to the corresponding author ( [saengow.udomsak@gmail.com](mailto: saengow.udomsak@gmail.com) ). Approval of access to the data must be granted by the Center for Alcohol Studies.