The authors have declared that no competing interests exist.
Obesity is a common and preventable Non-Communicable Disease that is of great importance. Population-based interventions are recognised to have a profound effect on improving health outcomes. One of these approaches includes the adoption of the Sugar-Sweetened Beverage (SSB) Tax.
There were three objectives, 1) explore the associations between gender, age, nationality, and change in SSB consumption, 2) explore SSB consumption during Covid-19 lockdown, and 3) inform policy decision making.
A cross-sectional survey in the United Arab Emirates. We performed descriptive analysis and chi-square for independence to test the difference between the expected and the observed frequencies in one or more categories.
Since the introduction of SSB tax, change in SSB consumption by gender, age or nationality was not statistically significant. Further analysis of the proportion of sugar intake per day was statistically significant (P-value <0.001) by nationality. There was no statistically significant change in SSB consumption by age, gender, or nationality during the Covid-19 lockdown. Further analysis within the group that reported change in SSB consumption suggests a majority (80.5%) reported a reduction in SSB consumption.
Change in SSB consumption by gender, age or nationality was not statistically significant since the introduction of SSB tax, or during Covid-19 lock-down; thus, we accept the Null Hypothesis. Imposing a levy on frequently consumed SSB or revisiting levy by the gram, volume, or type of added sugar (or in combination) may prove to be more effective in reducing SSB consumption. Further research is needed to determine the extent other demographic factors influence SSB consumption as well as the enablers and barriers associated with SSB consumption.
Obesity is a common and preventable Non-Communicable Disease (NCD) of great clinical importance and public health concern.
A cross-sectional, nationally representative survey was developed following a review of the evidence and research group discussions.
Data were collected in two languages (Arabic and English) through the Qualtrics survey system. Responses in Arabic were translated into English through the Qualtrics system and checked by a native Arabic speaker. All survey data was downloaded onto Microsoft Excel. Data were cleaned and coded on Microsoft Excel as per the survey questions, and responses were transferred into Statistical Package for Social Sciences (SPSS) Subscription version 2021 for analysis.
Analysis of responses was done for age, gender, and nationality. Age was analysed by the seven categories from the survey (18-25, 26-30, 31-35, 36-40,41-45,46-59 and 60 plus), gender as male and female, and nationality by Emirati and non-Emirati.
7,500 participants were recruited into the study, and we received 1 290 surveys (participation rate = 17.2%). The distribution of gender was 472 (36.7%) male and 817 (63.3%) female. For Nationality, we report 548 (42.5% Emirati) and 742 (57.5%) Non-Emirati. For age group, we report 276 respondents aged 31-35 years, 274 aged 46-59years, 270 respondents aged 41-45 years, 254 respondents aged 36-40 years, 157 respondents aged 26-30 years, 31 respondents aged 8-25 years, and 26 respondents aged over 60 years. Two missing responses were noted for age, 179 for change in SSB since tax, 179 for the percentage of sugar intake, and 179 for change in SSB consumption since Covid-19 lockdown. Overall, the demographic profile for gender, age, and nationality was consistent with national population data distribution.
Change in SSB consumption since the introduction of SSB tax by gender, age, and nationality is presented in
|
|
|
|
|
|
|
|
Male | 1290 (100%) | 413 | 143 | 80 | 181 | 8 | 1 |
Female | 698 | 193 | 138 | 341 | 19 | 7 | |
Total | 1111 (86.1%) | 336 | 218 | 522 | 27 | 8 | |
|
|
|
|
||||
Pearsons Chi Square | 8.334a | 4 | 0.080 | ||||
Likelihood Ratio | 8.670 | 4 | 0.070 | ||||
Linear-by-linear Association | 7.423 | 1 | 0.006 | ||||
No. of Valid Cases | 1111 | ||||||
a. 1 cell (10%) have expected count less than 5. The minimum expected count is 2.97. | |||||||
|
|
|
|
|
|
|
|
Age (18-25) | 1290 (100%) | 27 | 4 | 3 | 18 | 2 | 0 |
Age (26-30) | 129 | 33 | 29 | 64 | 3 | 0 | |
Age (31-35) | 244 | 72 | 45 | 121 | 5 | 1 | |
Age (36-40) | 214 | 66 | 46 | 94 | 7 | 1 | |
Age (41-45) | 235 | 63 | 46 | 115 | 8 | 3 | |
Age (46-59) | 236 | 90 | 42 | 99 | 2 | 3 | |
Over 60 | 24 | 8 | 6 | 10 | 0 | 0 | |
Total | 1109 (86%) | 336 | 217 | 521 | 27 | 8 | |
|
|
|
|
||||
Pearsons Chi Square | 27.869a | 24 | 0.266 | ||||
Likelihood Ratio | 29.313 | 24 | 0.209 | ||||
Linear-by-linear Association | 5.766 | 1 | 0.016 | ||||
No. of Valid Cases | 1109 | ||||||
a. 11 cells (31.4%) have expected count less than 5. The minimum expected count is 0.17. | |||||||
|
|
|
|
|
|
|
|
Emirati | 1290 (100%) | 458 | 139 | 82 | 221 | 13 | 3 |
Non-Emirati | 653 | 197 | 136 | 301 | 14 | 5 | |
Total | 1111 (86.1%) | 336 | 218 | 522 | 27 | 8 | |
|
|
|
|
||||
Pearsons Chi Square | 2.022a | 4 | 0.732 | ||||
Likelihood Ratio | 2.026 | 4 | 0.731 | ||||
Linear-by-linear Association | .282 | 1 | 0.595 | ||||
No. of Valid Cases | 1111 | ||||||
2 cells (20.0%) have expected count less than 5. The minimum expected count is 3.30. |
The percentage of daily sugar intake from SSB is presented in
|
|
|
|
|
|
|
|
|
|
Male | 1290 (100%) | 411 | 74 | 124 | 56 | 40 | 40 | 8 | 69 |
Female | 700 | 107 | 194 | 113 | 95 | 76 | 9 | 106 | |
Total | 1111 (86.1%) | 181 | 318 | 169 | 135 | 116 | 17 | 175 | |
|
|
|
|
||||||
Pearsons Chi Square | 7.439a | 6 | 0.282 | ||||||
Likelihood Ratio | 7.518 | 6 | 0.276 | ||||||
Linear-by-linear Association | 0.191 | 1 | 0.662 | ||||||
No. of Valid Cases | 1111 | ||||||||
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 6.29. | |||||||||
|
|
|
|
|
|
|
|
|
|
---|---|---|---|---|---|---|---|---|---|
Age (18-25) | 1290 (100%) | 27 | 3 | 13 | 2 | 2 | 4 | 0 | 3 |
Age (26-30) | 129 | 14 | 38 | 29 | 15 | 16 | 0 | 17 | |
Age (31-35) | 243 | 38 | 63 | 42 | 37 | 26 | 6 | 31 | |
Age (36-40) | 220 | 37 | 59 | 33 | 29 | 23 | 3 | 36 | |
Age (41-45) | 237 | 38 | 66 | 31 | 28 | 28 | 5 | 41 | |
Age (46-59) | 230 | 44 | 71 | 29 | 23 | 19 | 2 | 42 | |
Over 60 | 23 | 7 | 8 | 2 | 1 | 0 | 1 | 4 | |
Total | 1109 (86%) | 181 | 318 | 168 | 135 | 116 | 17 | 174 | |
|
|
|
|
||||||
Pearsons Chi Square | 40.124a | 36 | 0.292 | ||||||
Likelihood Ratio | 43.797 | 36 | 0.174 | ||||||
Linear-by-linear Association | 0.002 | 1 | 0.967 | ||||||
No. of Valid Cases | 1109 | ||||||||
a. 17 cells (34.7%) have expected count less than 5. The minimum expected count is .35. | |||||||||
|
|
|
|
|
|
|
|
|
|
Emirati | 1290 (100%) | 457 | 59 | 115 | 67 | 60 | 62 | 9 | 85 |
Non-Emirati | 654 | 122 | 203 | 102 | 75 | 54 | 8 | 90 | |
Total | 1111 (86.1%) | 181 | 318 | 169 | 135 | 116 | 17 | 175 | |
|
|
|
|
||||||
Pearsons Chi Square | 21.700a | 6 | 0.001 | ||||||
Likelihood Ratio | 21.675 | 6 | 0.001 | ||||||
Linear-by-linear Association | 17.729 | 1 | <0.001 | ||||||
No. of Valid Cases | 1111 | ||||||||
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 6.99. |
|
|
|
|
|
|
Male | 1290 (100%) | 413 | 210 | 203 | |
Female | 698 | 364 | 334 | ||
Total | 1111 (86.1%) | 574 | 537 | ||
|
|
|
|
|
|
Pearsons Chi Square | .176a | 1 | .675 | ||
Likelihood Ratio | .128 | 1 | .721 | ||
Fishers Exact Test | .176 | 1 | .675 | ||
Linear-by-linear Association | .709 | .360 | |||
No. of Valid Cases | .176 | 1 | .675 | ||
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 199.62.b. Computed only for a 2x2 table | |||||
|
|
|
|
|
|
---|---|---|---|---|---|
Age (18-25) | 1290 (100%) | 27 | 15 | 12 | |
Age (26-30) | 129 | 73 | 56 | ||
Age (31-35) | 244 | 133 | 111 | ||
Age (36-40) | 214 | 114 | 100 | ||
Age (41-45) | 235 | 106 | 129 | ||
Age (46-59) | 236 | 123 | 113 | ||
Over 60 | 24 | 8 | 16 | ||
Total | 1109 (86%) | 572 | 537 | ||
|
|
|
|
||
Pearsons Chi Square | 9.720a | 6 | .137 | ||
Likelihood Ratio | 9.776 | 6 | .134 | ||
Linear-by-linear Association | 4.080 | 1 | .043 | ||
No. of Valid Cases | 1109 | ||||
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 11.62. | |||||
|
|
|
|
|
|
Emirati | 1290 (100%) | 458 | 230 | 228 | |
Non-Emirati | 653 | 344 | 309 | ||
Total | 1111 (86.1%) | 574 | 537 | ||
|
|
|
|
|
|
Pearsons Chi Square | .653a | 1 | .419 | ||
Likelihood Ratio | .558 | 1 | .455 | ||
Fishers Exact Test | .653 | 1 | .419 | ||
Linear-by-linear Association | .428 | .227 | |||
No. of Valid Cases | .653 | 1 | .419 | ||
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 221.37.b. Computed only for a 2x2 table |
Our study provides new insight on the association between key demographic factors (gender, age, and nationality) and SSB consumption in the UAE. The data analysis suggests that changes in SSB consumption were not statistically significant since the introduction of the SSB tax. The percentage of sugar consumed per day from SSB indicates a statistical significance by nationality (P-value <0.001). Several factors may explain this finding. Approximately 88.53% of the UAE population are non-Emirati derived from over 200 nationalities, and tourism has been steadily rising; for example, Dubai was ranked sixth globally on the tourism index with 16.73 million visitors in 2019 and with this, an increasing need to cater for a variety of dietary needs.
Some studies have explored the impact of taxation on SSB. Colchero et al. estimated changes in sugar-sweetened beverages (SSB) sales and plain water after a 1 peso per litre excise SSB tax were implemented in Mexico in January 2014.
The study explored SSB consumption during Covid-19 lockdown. The number of respondents who reported consuming less SSB during the Covid-19 lockdown was slightly higher than those that reported no change (574 vs 537, respectively). It is plausible that the lockdown may have restricted the temptation for SSB, led to the prioritisation of beverages, or created a false impression of price elasticity among respondents. While research on SSB during Covid-19 lockdown is limited, there is evidence to suggest restricting access and availability of SSB in school and the workplace may lead to substitution or reduction in consumption.
The study does not explore the nature of SSB taxation or its impact on obesity and health outcomes, however; the logical pathway remains relevant as a framework for understanding SSB taxation. There are challenges in implementing the framework due to a lack of consensus on the policy approach and a lack of standardised tools to quantify consumption. This means consumption may vary considerably among individuals or, indeed, population groups. Consumption may be inappropriately estimated due to insufficient knowledge or recall bias, and respondents may find it difficult to quantify their sugar consumption between food and drink. Thus the need to develop validated research tools to quantify consumption more accurately. This would provide greater insight on the effects of SSB tax within population groups and at scale. Additionally, such tools would make it possible to conduct comparative analysis across different regulatory jurisdictions and maximise opportunities to determine superior SSB policy directives.
The study has limitations. We may have excluded specific groups within the population through our sampling approach, such as disease-specific groups or groups with limited access to the internet. Despite achieving the minimum study sample for our population, the study would benefit from a higher response rate for greater generalisability. They may have been alternative explanations to our results due to several factors that were not accounted for, such as self-reporting and interpretation of questions and responder ability to segregate sugar consumption between ' food and drink. Also, adopting a quantitative approach to measure SSB consumption may limit our understanding of the findings' underlying issues.
There are several policy implications from the study. Demographic factors should be considered when developing fiscal policy for SSB and different fiscal policy approaches should be considered. For example, imposing a levy on frequently consumed SSB or revisiting levy by the gram, volume, or type of added sugar or in combination may prove to be more effective in reducing SSB consumption. Performance measures should be considered at different policy implementation stages to determine the impact and provide the opportunity to mitigate the unintended consequences.
Change in SSB consumption by gender, age or nationality was not statistically significant since the introduction of SSB tax, or during Covid-19 lock-down; thus, we accept the Null Hypothesis. Imposing a levy on frequently consumed SSB or revisiting levy by the gram, volume, or type of added sugar (or in combination) may prove to be more effective in reducing SSB consumption. Further research is needed to determine the extent other demographic factors influence SSB consumption as well as the enablers and barriers associated with SSB consumption.
Ethical approval for the study and survey was granted from the Mohammed Bin Rashid School of Government Ethics Committee (REC-80-2020).
The authors would like to acknowledge Mohammed Bin Rashid School of Government, Dubai, UAE, and the Alliance for Health Policy and Systems Research at the World Health Organization for financial support as part of the Knowledge to Policy (K2P) Center Mentorship Program (BIRD Project).
Figure S1.
Figure S2.