Obesity has been rising across many countries, and the reasons behind that rise are still debated. Scroll to explore why what Americans believe about obesity might not match the data.
Obesity has increased almost everywhere over the past few decades. Estimates from the World Health Organization indicate that global adult obesity has nearly tripled since 1975, reaching about 1 in 8 adults worldwide by 2022 [4].
The interactive visualization on the left uses standardized data from the NCD Risk Factor Collaboration tracking obesity prevalence from 1990 to 2022 [5]. Try pressing the "play" button to watch how obesity has become more common over time. Notice how different regions change color (from lighter to darker shades) as obesity rates increase.
While the timing and speed of rising obesity differ across countries, obesity is becoming more common. That raises two questions: why is it increasing at all, and what explains the differences between countries?
Now take a closer look at the patterns. Click on different countries to see how their obesity rates have changed over time. The visualization will show trends for adults, women, men, and, when available, children.
Notice how the timing, slope, and scale vary dramatically. Some countries like the United States show an early, steady rise. Adult obesity roughly tripled over several decades, reaching around 40% today [6] [7]. Others like Japan show a much slower increase from a lower starting point, remaining among the lowest rates in the OECD [8] [9].
Some Gulf countries like Saudi Arabia, Qatar, and Kuwait experienced a rapid late surge; obesity was relatively uncommon decades ago, then shot up to 30-40% of adults as income and food environments changed dramatically [10] [11].
As you explore, ask yourself: What patterns do you notice in which regions have higher obesity? What might be driving this sudden rise across so many countries? Is it reduced physical activity? Changes in diet? Economic development? Something else entirely?
At its core, obesity occurs when energy intake is consistently higher than energy expenditure. The harder question is which side of that balance has shifted in the last few decades. Public health organizations typically point to two possibilities: people may be burning less energy through daily activity, or they may be consuming more energy as diets shift [1].
A 2006 Pew survey captured Americans’ beliefs about what was causing obesity. Most respondents pointed to personal behavior as the main driver, and they overwhelmingly pointed to one explanation in particular: not getting enough exercise. In fact, they rated lack of activity as the most important reason people are overweight, ahead of diet, marketing, or genetics [2]. Although the study is almost a decade old, it captures a then dominant view among Americans, shared with health organizations, that declining physical activity could be a major contributor to rising obesity [1].
A cross-cultural study in PNAS directly tested the assumption that rising obesity is driven by declines in physical activity. The researchers measured total daily energy expenditure using the doubly labeled water method, drawing from the International Atomic Energy Agency’s global database and adding three additional populations: the Aymara in Bolivia, the Tuvan in Siberia, and the Daasanach in northern Kenya [1].
The final study sample included 4,213 adults across 34 population groups. The original team had access to individual level measurements, but access to that data requires a formal application. This analysis uses only the published population means from their study. Also, each population represents a community sample, not a national estimate.
We included hunter gatherers, pastoralists, horticulturalists, small scale farmers, and industrialized groups because they differ in lifestyle, and market integration. These differences allow better comparisons across variables that may predict obesity levels. To measure obesity we did not use body mass index but body fat percentage to avoid penalizing individuals with higher muscle mass.
The map shows where the 34 sampled communities are located. Populations are categorized using the Human Development Index, which ranks groups by life expectancy, education, and income, along with other economy type data in the dataset. You can hover over each economy type to see what it means.
High HDI groups are heavily represented, especially in Europe, the US, and East Asia. These larger samples produce more stable averages. In contrast, subsistence groups hunter gatherers, horticulturalists, and agropastoralists appear in only a few regions and often with small sample sizes. They are essential for comparing lifestyles, but the map visualization helps demonstrate that there is far less data for them than for high HDI groups.
This plot compares body fat percentage with Physical Activity Level (PAL), which is total energy expenditure divided by basal energy expenditure. PAL gives a size-adjusted measure of daily activity.
For men, the slope is strongly negative and statistically significant. The p value is 0.01 and R² is about 0.21, so PAL explains around 21% of the variation in body fat which is a moderate amount, but not large. For women, the slope is not significant. The p value is 0.25 and R² is near zero, which means PAL has no detectable linear relationship with body fat in this sample.
The main takeaway is that higher PAL may predict lower body fat for men, but the effect is still only moderate. For women, PAL did not meaningfully predict body fat.
This plot compares body fat percentage with the share of ultra-processed foods (UPF) in the diet. Unlike the activity plots, the points line up more clearly here. For both men and women, the relationship between UPF intake and body fat is statistically significant. The p values are 0.01 for men and 0.005 for women. The slopes are positive and similar in size, which means higher UPF intake predicts higher body fat across sexes.
For men, PRE is about 0.31. For women, it is about 0.44. These are much higher than what you saw with PAL. They are still moderate, not huge, but they indicate that UPF explains a meaningful share of the variation in body fat, especially for women. There is still variation. Some groups eat relatively little UPF and differ in body fat, and some high-UPF groups sit below the trend line, but the overall slope is stronger than what we saw for PAL.
One additional pattern also starts to emerge. Populations with higher UPF intake tend to have higher HDI scores. As countries industrialize, they usually shift toward more processed and packaged foods, and that transition aligns with the higher-UPF, higher-body-fat clusters in the plot.
It’s also worth noting that UPF data were available for only 25 of the 34 populations, so the coverage is not complete. Even with that limitation, the relationship appears more consistent than the activity based measures. This doesn’t prove that UPF is the cause, but it suggests that dietary composition may align more closely with differences in body fat across populations than physical activity does.
You can change both the x and y axes to test how any two variables relate. When you switch axes, check three things: whether the relationship is statistically significant (p values is less than 0.05), whether the line goes up or down, and how spread out the points are around the line.
Try pairing different x and y axes and ask yourself what each comparison reveals. • HDI Rank vs UPF As HDI improves, does UPF intake rise? Does this match the idea that industrialization shifts diets more than it changes activity? • TEE vs PAL Does burning more total energy actually mean being more physically active? • Economy Type vs Any Metric Where do you see the most or least variance within a Economy Type?
Across populations, physical activity did not show a clear pattern that explains large differences in body fat. In the subset of groups with dietary data, higher intake of ultra processed foods shows a more consistent link with higher body fat. This is correlational, not causal proof, but it suggests diet composition may explain population level differences better than activity.
Public discussions often blame inactivity or willpower. Some of that framing has been shaped by industry funded messaging that emphasized exercise over diet [3]. If UPFs track more closely with body fat than activity does, strategies focused mainly on individual behavior will have limited effect. Policies that target food environments, marketing, and UPF availability likely offer a stronger lever.
Dietary data cover only 25 of the 34 populations, so any dietary pattern is based on a partial sample. The analysis relies on population means, which prevents us from seeing how much variation exists within each group or how precise the averages are. The data are cross sectional and correlational, meaning they describe differences between populations at one point in time and cannot establish causation or track changes over time. Energy expenditure measurements come from different studies that used slightly different methods, adding measurement noise. These limits narrow what we can infer, but the broad patterns still point toward the need for better dietary data, long term studies, and policy approaches that address UPFs directly.
These population patterns suggest that diet may play a far more important role in obesity than many public health organizations initially recognized. This growing recognition of diet’s influence also frames how we interpret the recent decline in U.S. obesity rates, which further challenges the old notion that physical activity was the main driver of rising obesity.
Since 2022, U.S. obesity rates have been declining! Gallup's 2025 National Health and Well-Being Index reports a drop from 39.9 percent in 2022 to 37.0 percent in 2025, an estimated 7.6 million fewer adults with obesity [14].
The decline aligns with the rapid uptake of GLP-1 drugs, which help people feel full sooner and eat less. National spending on these medications increased from $13.7 billion in 2018 to $71.7 billion in 2023 [12]. Survey data indicate that roughly 40% of prescriptions are for weight management, and adult use has risen from 5.8% in early 2024 to 12.4% in 2025 [15] [14].
The steepest reductions in obesity are occurring in the very age groups with the highest GLP-1 usage (ages 40–64),suggesting these medications may be contributing to the first sustained population-level decline in U.S [14]. obesity rates.
As shown earlier, the populations with the highest intake of ultra processed foods also tend to have the highest body fat percentages. The effectiveness of GLP-1 drugs helps explain why. UPFs interfere with the leptin–melanocortin pathway, the brain circuit that regulates hunger, fullness, and metabolism [16]. GLP-1 drugs work by boosting signals in the same pathway [17], which makes it clear why disrupting that system with UPFs can drive weight gain in the first place.
UPFs are energy-dense, rapidly absorbable, and engineered to be hyper-palatable. When consumed regularly, they drive repeated spikes in glucose, insulin, inflammatory signals, and leptin. Over time, this barrage leads the hypothalamus to become progressively less responsive to leptin and insulin, a state called leptin and insulin resistance [18]. In this state, the brain perceives the body as having less stored energy than it actually does, triggering the same biological response seen in genetic obesity: hunger increases, satiety decreases, and metabolic rate slows.
Here's why the obesity decline matters, GLP-1 medications act on this same energy-regulation circuitry. They stimulate satiety-promoting neurons and dampen hunger-promoting neurons, essentially overriding the dysfunctional signals that arise from leptin resistance. By counteracting the dysfunction of this pathway caused by UPFs, GLP1s help change eating behavior. Specifically, they restore satiety signals and in doing so reduce the appeal and consumption of hyper-palatable ultra-processed foods [19].
This dual mechanism, restoring satiety and reducing UPF consumption, is likely why we’re seeing the recent sustained population level obesity decline in the U.S. It’s not about willpower or exercise. It’s about correcting a biological dysfunction caused by UPFs that drives overconsumption.
Understanding that GLP-1 medications work by targeting the same biological pathway disrupted by UPF has important policy implications. Since UPF consumption drives obesity through pathway dysfunction, and GLP-1s restore that pathway, then ensuring access to these medications in regions with the highest obesity rates becomes a public health priority.
The interactive state-by-state visualization below shows an encouraging pattern: states with higher obesity rates tend to have higher GLP-1 prescription counts. This suggests that GLP-1 medications are reaching the locations where they are needed most. However, it's important to note that this data reflects only prescription patterns and does not account for off-label usage or non-prescription access, which may be substantial in some regions.
Despite this limitation, the alignment between obesity rates and prescription patterns is a positive sign. It indicates that clinical need is driving adoption. However, access remains unequal. Monthly out-of-pocket costs often range from $1,000–$1,300 without insurance [20], making these medications unaffordable for many people in the regions with highest obesity rates (Midwest and Southern) as respective states are mostly also the ones with much lower average incomes (feel free to check this out by clicking the 'Size by Income' button in the visualization). As a result, the true distribution of GLP-1 usage is almost certainly more uneven than the prescription patterns alone suggest.
The visualization below shows GLP-1 prescription patterns across all 50 U.S. states and how they relate to state-level obesity rates. Use the interactive controls to explore regional patterns and identify which states have the highest adoption of these medications.
This emphasis a crucial role for government intervention: expanding access to GLP-1 medications in high-obesity regions should be a policy priority. If we understand obesity as a biological dysfunction caused by environmental factors (namely, UPF-dominated food environments), then treating it requires systemic solutions that include making effective medications accessible to those who need them most, regardless of income or geography.
While this encouraging decline in obesity rates is significant, it does not address the underlying drivers of the epidemic. GLP-1s can counteract a dysfunctional leptin–melanocortin pathway and restore satiety signals, but relying solely on medication to manage weight ignores that obesity is only one manifestation of poor dietary quality.
The goal should not simply be to eat less, it should be to eat better. UPF don't just drive weight gain; they are independently associated with cardiovascular disease, Type 2 diabetes, metabolic dysfunction, chronic inflammation, and many cancers [21]. A person on GLP-1 medication who continues consuming a diet dominated by UPFs may lose weight, but they remain exposed to the broader health consequences that obesity, and body fat, metrics alone don't capture.
This emphasis the importance of policy action when it comes to our food system. While expanding access to GLP-1 medications is a crucial public health intervention, it must be paired with efforts to address the root cause: the pervasive availability and aggressive marketing of ultra-processed foods that makes unhealthy eating the default choice. Policy solutions should include improving food environments, especially in underserved communities, supporting nutrition education, regulating misleading health claims on processed foods, and incentivizing the production and accessibility of whole, minimally processed foods.
Across this report, we challenged the familiar assumption that obesity is a failure to exercise and showed that, when we compare physical activity across populations, activity levels do not strongly predict differences in body fat. Instead, we saw how UPF align far more closely with body fat patterns and how the recent U.S. decline in obesity lines up with the spread of GLP-1 drugs that target the brain circuits UPFs disrupt. By exploring these relationships, we see that understanding obesity requires looking beyond individual effort and toward the biological and environmental forces that shape our behaviors and bodies.
[1] Pontzer, Herman, et al. "Energy Expenditure, Dietary Energy Intake, and Body Composition Across Populations." Proceedings of the National Academy of Sciences, vol. 122, no. 3, 2025, e2420902122. PNAS, https://doi.org/10.1073/pnas.2420902122.
[2] "Americans See Weight Problems Everywhere But In the Mirror." Pew Research Center, 11 Apr. 2006, https://www.pewresearch.org/social-trends/2006/04/11/americans-see-weight-problems-everywhere-but-in-the-mirror/.
[3] Nixon, Laura, et al. "We're Part of the Solution: Evolution of the Food and Beverage Industry's Framing of Obesity Concerns Between 2000 and 2012." American Journal of Public Health, vol. 105, no. 11, Nov. 2015, pp. 2228–2236. PMC, https://doi.org/10.2105/AJPH.2015.302819.
[4] "Obesity and Overweight." World Health Organization, https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight.
[5] "NCD Risk Factor Collaboration (NCD-RisC)." NCD-RisC, https://ncdrisc.org/.
[6] "Prevalence of Obesity and Severe Obesity Among Adults: United States, August 2021–August 2023." Centers for Disease Control and Prevention, Data Brief No. 508, Mar. 2024, https://www.cdc.gov/nchs/products/databriefs/db508.htm.
[7] "Obesity Rate Nearly Triples in the United States Over Last 50 Years." USAFacts, https://usafacts.org/articles/obesity-rate-nearly-triples-united-states-over-last-50-years/.
[8] Roser, Max, and Hannah Ritchie. "Obesity." Our World in Data, https://ourworldindata.org/obesity.
[9] Ikeda, Nayu, et al. "Adult Mortality Attributable to Preventable Risk Factors for Non-Communicable Diseases and Injuries in Japan: A Comparative Risk Assessment." PLOS Medicine, vol. 9, no. 1, Jan. 2012, e1001160. PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC9163402/.
[10] "World Obesity Atlas 2023." World Obesity Federation, https://data.worldobesity.org/rankings/.
[11] Al-Nohair, Sulaiman. "Obesity in Gulf Countries." International Journal of Health Sciences, vol. 8, no. 1, Jan. 2014, pp. V–VI. PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC4039587/.
[12] "Spending on GLP-1s has grown dramatically. Here are the details." American Medical Association, https://www.ama-assn.org/public-health/behavioral-health/spending-glp-1s-has-grown-dramatically-here-are-details.
[13] "United States of Ozempic: Where anti-obesity drugs are taking off." Axios, 18 Jan. 2024, https://www.axios.com/2024/01/18/ozempic-wegovy-weight-loss-drugs-states-map.
[14] "Obesity Rate Declining in U.S." Gallup, https://news.gallup.com/poll/696599/obesity-rate-declining.aspx.
[15] Montero, Alex, et al. "KFF Health Tracking Poll May 2024: The Public's Use and Views of GLP-1 Drugs." KFF, 10 May 2024, https://www.kff.org/health-costs/kff-health-tracking-poll-may-2024-the-publics-use-and-views-of-glp-1-drugs/.
[16] Hall, Kevin D., et al. "Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake." Cell Metabolism, vol. 30, no. 1, July 2019, pp. 67-77.e3. PMC, https://doi.org/10.1016/j.cmet.2019.05.008.
[17] Müller, Timo D., et al. "Glucagon-like Peptide 1 (GLP-1)." Molecular Metabolism, vol. 30, Dec. 2019, pp. 72-130. PMC, https://doi.org/10.1016/j.molmet.2019.09.010.
[18] Farooqi, I. Sadaf, and Julia M. O'Rahilly. "Leptin: A Pivotal Regulator of Human Energy Homeostasis." The American Journal of Clinical Nutrition, vol. 89, no. 3, Mar. 2009, pp. 980S-984S. PMC, https://doi.org/10.3945/ajcn.2008.26788C.
[19] Semaglutide Effects on Cardiovascular Outcomes in People With Overweight or Obesity (SELECT) Trial. New England Journal of Medicine, vol. 389, Nov. 2023, pp. 2221-2232. https://doi.org/10.1056/NEJMoa2307563.
[20] "How Much Do GLP-1 Injections Cost? Average Prices, Savings Tips, and More." CallOnDoc, 2025, https://www.callondoc.com/en/blogs/how-much-do-glp-1-injections-cost-average-prices-savings-tips-and-more.
[21] Pagliai, Giuditta, et al. "Consumption of Ultra-Processed Foods and Health Status: A Systematic Review and Meta-Analysis." British Journal of Nutrition, vol. 125, no. 3, Feb. 2021, pp. 308-318. https://doi.org/10.1017/S0007114520002688.