Can soda taxes fight obesity? New research adds to the debate

By Pooja Toshniwal PahariaReviewed by Lily Ramsey, LLMOct 16 2024

Recent studies suggest sugary drink taxes may help reduce obesity rates, but the evidence remains mixed—find out what the latest research reveals.

Study: Associations of the Philadelphia sweetened beverage tax with changes in adult body weight: an interrupted time series analysis. Image Credit: VDZ3 Media/Shutterstock.com

A recent study in The Lancet Regional Health-Americas evaluated the impact of the 2017 Philadelphia sweetened beverage taxation on adult weight.

Background

Consuming sugar-sweetened beverages (SSB) is related to poor metabolic health and chronic diseases like obesity, cardiovascular disease, and insulin-independent diabetes. Sweetened fluids contribute the most to dietary added sugars.

Interventions aimed at reducing SSB intake may consequently benefit population health. An intervention of this type is the taxing of sugary beverages. The Philadelphia beverage tax, which went into effect on January 1, 2017, is a 1.5-cent-per-ounce tax on sugar-sweetened beverages and artificially sweetened drinks.

The Philadelphia beverage tax has dramatically increased beverage prices and decreased SSB sales, which may enhance health results. However, there is limited research on the influence of beverage taxation on adult weight-management results.

Two studies on children indicated that sweetened beverage tariffs in Mexico and Seattle resulted in slight changes in body weight. A birth certificate-based study found slight post-tax decreases in body weight for the week of gestation and the risk of pregnancy-related diabetes. There is limited research on the effects of US beverage taxation on weight parameters among non-pregnant adult individuals.

About the study

In the present study, researchers investigated whether the Philadelphia taxation on sweetened beverages reduced body mass index (BMI) and obesity prevalence among adults in Philadelphia.

The researchers analyzed electronic medical records of adults aged ≤65 years from Philadelphia (study intervention) and regions of New Jersey and Pennsylvania (controls) between 2014 and 2019. All participants provided their height, weight, and residential zip codes.

The primary study outcome was a BMI change. For every quarter, researchers determined patients’ obesity status. A BMI below 25 indicated healthy weight, a BMI above 25 but less than 30 indicated overweight, and a BMI ≥30 indicated obesity.

The secondary outcome was the prevalence of obesity. The panel sample included 175,675 individuals with BMI recorded before (between 2014 and 2016) and after taxation (between 2017 and 2019). The cross-sectional sample included 587,121 individuals with one or more BMI measures recorded between 2014 and 2019.

Controlled interrupted time series models evaluated post-taxation changes in BMI and obesity prevalence rates. Potential confounders included age, sex, race, ethnicity, Yost scores, and pre-tax healthcare consultations. The team used inverse probability of treatment-weighted (IPTW) generalized estimating equations to adjust for these confounders.

The researchers calculated Yost index scores for each residential census to assess their neighborhood-level socioeconomic status. Medicaid status denoted the individual-level socioeconomic status.

For panel individuals, researchers determined the number of visits between 2014 and 2016 to measure their pre-taxation utilization of healthcare services. They stratified the results by gender, race, and weight.

For sensitivity analyses, the researchers aggregated data at patient-month levels to estimate monthly versus quarterly changes, excluding the initial six months after taxation. They also removed outlier values for BMI, i.e., more than three standard deviations above the mean. The analyses included only individuals with BMI measures recorded in each year of the study.

Results

Among Philadelphian individuals with a mean age of 42, 59% were female, 41% were white, 50% were black, 4.0% were Pacific Islander or Asian American, 5.0% belonged to other races, and 4.0% were Hispanic. Most Philadelphia residents resided in low-income areas (65% had Yost index ratings of 1 or 2), and Medicaid covered 20% of their visits.

Before the tax implementation, the mean BMI among Philadelphia panel individuals was 30 kg/m2, and the prevalence of obesity was 45%. After the tax implementation, BMI decreased by 0.030 kg/m2 per quarter compared to controls, indicating a 0.30 kg/m2 decrease after three years.

Cross-sectional sample individuals showed a 0.050 kg/m2 per quarter reduction in BMI compared to controls, indicating a 0.60 kg/m2 decrease at three years. Obesity prevalence findings were similar to those for BMI. Cross-sectional sampling yielded more precise estimates than longitudinal panel sampling.

The stratified analysis showed similar results for both study outcomes across population subgroups. However, the associations were stronger among Blacks than among Whites. Most sensitivity analyses yielded consistent findings. However, restricting the analyses to individuals with annual BMI records attenuated the results. There were no differences between males and females.

The study findings provide limited evidence for reduced body mass index and obesity in Philadelphia three years after implementing the beverage tax. The findings indicated that sweetened beverage taxation may have weight-related health benefits for non-pregnant adults, especially for black individuals. However, further research is required to replicate the results.