Investment case for the prevention and reduction of childhood and adolescent overweight and obesity in Mexico

Despite efforts to curb the rise in Mexico's child and adolescent overweight and obesity rates, prevalence in Mexico has grown by 120% since 1990 to 43.3% in 2022. This investment case identifies policies that will produce the largest returns for Mexico. The investment case model builds beyond a cost‐of‐illness analysis by predicting the health and societal economic impact of implementing child and adolescent overweight and obesity interventions in a cohort aged 0–19 from 2025 to 2090. The Markov model's impacts include healthcare expenditures, years of life lost, and reduced wages and productivity. We projected and compared costs in a status quo scenario to an intervention scenario to estimate cost savings and calculate return‐on‐investment (ROI). Total lifetime health and economic costs amount to USD 1.8 trillion—USD 30 billion on average per year. Implementing five interventions can reduce lifetime costs by approximately 7%. Each intervention has a low cost per disability‐adjusted life year averted over 30‐year, 50‐year, and lifetime horizons. The findings demonstrate that a package of interventions mitigating child and adolescent overweight and obesity offers a strong ROI. The novel investment case methods should be applied to other countries, particularly low‐ and middle‐income countries.

Besides the many health impacts of living with overweight and obesity, there are also numerous economic impacts. Children and adolescents affected by overweight and obesity on average tend to miss more days of school and, while they are at school, have an increased risk of exhibiting behavioral problems. 10,[13][14][15][16][17] In adulthood, those affected by overweight and obesity miss more days of work (absenteeism) and are, on average, less able to work at full capacity while they are at work due to obesity-attributable diseases (presenteeism). 18 Adults affected by overweight and obesity are more likely to be unemployed and have lower wages. 19,20 Research from Mexico shows that women affected by overweight and obesity face hiring discrimination. 21 Finally, deaths and disabilities resulting from obesityrelated diseases imply a loss of potential future contributions and result in significant impacts on the economy. 22,23 Successive Mexican governments have demonstrated awareness and a high level of policy responsiveness to child health and welfare, as seen, for example, in the past with Prospera (formerly Opportunidades and Progresa), which was rigorously evaluated and used iterative program design to improve implementation and impact. 24 Nonetheless, this policy previously focused on all aspects of malnutritionparticularly undernutrition-and evidence of the increasingly obesogenic environment subsequently led Mexican policymakers to enact measures to specifically address the country's progressively obesogenic environment with measures such as taxation of sugarsweetened beverages and junk food, restrictions on marketing unhealthy foods to children, front-of-pack nutrition labels, banning unhealthy food in schools, breastfeeding promotion, and encouraging physical activity at school. [25][26][27][28][29] However, some of these interventions were not adequately implemented due to industry interference. 30 Moreover, the prevalence of overweight and obesity among children and adolescents has continued to grow to its current level of 43.3% in 2022-a 120% increase since 1990. 23,31 This stark statistic illustrates the need to address critical gaps and adjust and reinforce policies to address the systemic risk factors for overweight and obesity so that they can be more effective. 32 While there have been cost-effectiveness analyses of various child and adolescent overweight and obesity interventions in low-and middle-income countries (LMIC), 33,34 our review of the literature found that there are no existing investment cases (or cost-benefit analyses) that compare cross-sectoral interventions for child and adolescent overweight and obesity in LMIC contexts. Investment cases have been demonstrated to provide policymakers with evidence needed to prioritize scarce health resources. [35][36][37] While a costof-illness (COI) analysis defines the disease burden's cost to society in economic terms if action is not taken, an investment case builds on the COI analysis by examining the potential economic and health impacts of implementing or scaling health interventions. 38 Hence, this investment case aims to begin to fill that gap by quantifying the health and economic impacts of child and adolescent overweight and obesity in Mexico, an upper-middle income country, and the potential gains from prevention and reduction. It identifies which policies and interventions will produce the largest health and economic returns for Mexico (the return on investment, or ROI) to support resource allocation and prioritization to respond efficiently to this growing global health challenge.

| METHODS
The investment case assesses the health and economic impact of child and adolescent overweight and obesity prevention and reduction interventions using methods from two previous economic models that have been used only in high-income countries to date. The Assessing Cost-Effectiveness in Obesity (ACE-Obesity) model examines the impacts of child obesity on future mortality, and the Early Prevention of Obesity in Childhood (EPOCH) model, which also includes healthcare costs and productivity loss, assessed the cost-effectiveness of child and adolescent overweight and obesity prevention interventions in Australia and other countries, including the United States. [39][40][41][42] We developed a deterministic Markov cohort model to assess the health impacts of overweight and obesity in children and adolescents in Mexico. By taking a societal cost perspective, we estimated the reductions in mortality and morbidity from implementing evidencebased interventions and the resulting economic impact in terms of healthcare cost savings and impacts on wages and productivity during adulthood, including from education. Applying a broad societal cost perspective includes all costs and health effects regardless of who pays or who is affected, offering insight into the interventions' impact on the wider population if implemented. 43,44 Because of the societal approach of the analysis, tax revenue impact should not be considered because they are transfers of resources across society, not net gains or losses. 44 The model cohort includes children aged 0-19 in 2025 and estimates impacts from 2025 to 2090. In 2090, the mean age of the cohort will be 75 years, which is the current life expectancy in Mexico. The year 2025 was selected as the base year under the assumption that it would require additional time to prepare and implement intervention programs and policies.
The analysis defines overweight for adults as a body mass index (BMI, calculated as weight in kilograms divided by the square of height in meters) of 25 to <30, and obesity is defined as a BMI of 30 and above, while for children, overweight is defined as a BMI-for-age above one standard deviation of the World Health Organization (WHO) Growth Reference median for children of the same age and sex and obesity as a BMI-for-age above two standard deviations of the median. 45 The analysis was performed with Microsoft Excel and used localized evidence where available. It was also conducted according to the Consolidated Health Economic Evaluation Reporting Standards' practice guidelines (Table S6). 46

| Baseline scenario
The baseline scenario estimates the health outcomes for children and adolescents, absent any new policy, throughout their lifetime. We assume that current trends in mortality, morbidity, and risk factors in Mexico remain unchanged and project future mean BMI and overweight and obesity prevalence, disaggregated by age and sex. For children aged 0-19 in 2025 in Mexico (the model cohort), we project future BMI for every year from age 5 to 19 and then in 5-year increments from age 20 to 84 (Tables S3a,b and S4a,b). 31,39 We use singleyear age groups for children and adolescents because the relationship between BMI and overweight and obesity is more variable during this period and becomes more stable in adulthood, where we use 5-year age groups. To do this, we used cohort and age effects on the BMI of children, adolescents, and adults in Mexico using data from the NCD-Risk Factor Collaboration, estimated using multiple linear regression. 31,39 We then convert mean BMI into prevalence of overweight and obesity. The relationship was modelled separately by sex and singleyear age group for children aged 5-19 and by sex and five-year age group for adults aged 20 and above since the relationship between BMI and overweight and obesity changes through the life cycle. We used historical mean BMI, overweight prevalence, and obesity prevalence, disaggregated by sex and 5-year age groups, from the NCD Risk Factor Collaboration. 31 Because the relationship between BMI and overweight and obesity prevalence is nonlinear for children and adolescents aged 5-19, we modelled the relationship between BMI and overweight prevalence and obesity prevalence using a cubic spline of mean BMI, with knots at every 5 years, as undertaken by Stevens et al. 47 For adults aged 20 years and over, we used linear regression to project future overweight and obesity prevalence of the model cohort. 48,49 We use projected overweight and obesity prevalence to calculate future health and economic impacts as described below. First, we estimated the years of life lost (YLLs), years lived with disability (YLDs), and disability-adjusted life-years (DALYs) among children and adolescents affected by overweight and obesity during their lifetime. We discounted the effects at 3% annually, the standard discount rate for global health economics research. 44 where q x,a is the mortality rate in year x for the 5-year age group within which a falls, obtained from World Population Prospects (WPP) 53 ; pop a is the population in age group a in year x also obtained from WPP; WHO standard life expectancy is 90 years for both males and females, 54 and r is the discount rate of 3%. 50,51 Overweight and obesity-attributable YLLs were estimated using the proportion of allcause YLLs that are due to overweight and obesity-attributable diseases or conditions 55,56 obtained from projected all-cause and causespecific mortality rates from 2026 to 2090 using an autoregressive integrated moving average (ARIMA) model. 57 The analysis uses gross domestic product (GDP) per capita as a proxy for the economic value of a life year to capture economic losses from premature mortality. 50 Using GDP per capita as a proxy offers a more equitable approach to valuing each individual's economic contribution across the life course regardless of their employment status. 23

| Baseline YLDs
To calculate the YLDs due to each overweight and obesityattributable condition for each 5-year age group, we followed the procedure of Menon et al. 58 This approach entails obtaining Mexico's age-specific YLDs and YLLs from the 2019 Global Burden of Disease study and calculating the YLD to YLL ratio by 5-year age groups for each overweight and obesity-attributable condition included in the analysis. 2 The YLD to YLL ratio was applied to the overweight and obesity-attributable YLLs we previously calculated.

| Baseline DALYs
DALYs due to each overweight and obesity-attributable condition were calculated as the sum of the YLLs and YLDs.
We estimated the economic impact of children and adolescents affected by overweight and obesity at baseline as a combination of obesity-attributable healthcare costs during childhood and adulthood and the impact obesity has on their future labor productivity. consumer price index (CPI) and discounted at 3% annually. 50,51,61 Thus, the obesity-attributable healthcare cost for age group a in year y was calculated using the formula: Obesity attributable healthcare cost a,y ¼ where N is the number of individuals affected by obesity, C is the additional annual obesity-attributable healthcare cost per person, r is the discount rate, and y is the future year.

| Impact on labor productivity
We modelled the lifetime loss in wages due to lower educational attainment and work absenteeism and presenteeism among children and adolescents affected by overweight and obesity. To estimate the impact on educational attainment, we rely on a Swedish study which found that having overweight and obesity during childhood and adolescence is significantly associated with a 27.6% lower likelihood of attaining 12 or more years of education (after controlling for gender, ethnicity, and attention deficit hyperactivity disorder). 62 Using this finding, the number of individuals affected by overweight and obesity at age 18 and the proportion of the Mexican population aged [25][26][27][28][29][30][31][32][33][34] with tertiary education, we estimated the number of individuals who would not attain tertiary education due to overweight and obesity.
The lifetime loss in wages is then calculated as Lifetime loss in wages a ¼ N a w tert u tert À u sec where N a is the number of individuals in sub-cohort a who did not attain tertiary education due to obesity; u tert is the expected wages with tertiary education; u sec is the expected wages with secondary education; w sec is the proportion of population age 20-24 that attained secondary education who are currently working and w tert is the proportion of population age 20-24 with tertiary education who are currently working. Wages and the proportion of the working population, by education attainment and sex, are obtained from Mexico's National Institute of Statistics and Geography (INEGI). 63 We accounted for the four additional years during which individuals who did not attend tertiary school may have worked (since college education lasts for 4 years in Mexico) and assumed that every individual who attained tertiary education would enter the workforce at age 22 and exit at age 65. The lifetime wages lost by each subcohort a were discounted at 3% annually. 50,51 The study took a human capital approach to estimate the productivity loss due to obesity-related absenteeism and presenteeism from work. Prior research found that overall work impairment (combining absenteeism and presenteeism) of 4.4% among a nationally representative sample of individuals in Mexico with obesity and no significant difference for individuals affected by overweight compared to healthy weight individuals. 64 To calculate the productivity loss due to obesityattributable absenteeism and presenteeism, we multiplied the 4.4% work reduction attributable to obesity by the average annual salary in Mexico 65 and the number of individuals with obesity in the workforce till the age of 65, when Mexican citizens can retire with a pension. 66,67 We chose to apply the productivity losses associated with an individual's obesity status rather than use productivity loss values attributable to obesity-related diseases (for example, the productivity loss value associated with a general incident case of diabetes, regardless of the underlying cause) as the productivity losses experienced among people affected by obesity may not be disease-specific or could be the result of comorbidities.

| Intervention scenario
To identify and select priority overweight and obesity interventions for children and adolescents in Mexico, we conducted literature reviews of the costs and cost-effectiveness of interventions to prevent and reduce child and adolescent overweight and obesity in Mexico and globally, with attention to the policy environment in Mexico (Table 1). We also reviewed the WHO's recommendations and guidelines for ending child and adolescent obesity and consulted with UNICEF nutrition specialists. 68,69 We used purposive sampling to identify and conduct semi-structured interviews of six key informants from a range of stakeholder groups in Mexico specializing in overweight and obesity, nutrition, and public policy to gather their views on the interventions deemed to be of the greatest value and suitable T A B L E 1 Intervention selection criteria. for the context. 70 A summary of their feedback is provided in Appendix S1 (Table S9). Based on the evidence collected, the following five priority interventions were selected.
Strengthen the restrictions on marketing unhealthy foods to children: Extend the existing regulation that restricts TV and some movie advertising of certain unhealthy foods and beverages to children and have not yet been shown to be as cost-effective as the included interventions. 76 Additionally, the analysis does not model primary health care interventions as Mexico already has plans to expand primary health care nutrition services to address all forms of malnutrition. 79 Despite their higher cost, systems to support weight management in the primary healthcare system among children and adolescents who are already affected by overweight and obesity are important.
WHO suggests this is undertaken using a multicomponent, familybased approach delivered by a multidisciplinary team. 68 One Mexican study found that a physical activity and nutrition lifestyle intervention for children with overweight and obesity in Mexico delivered by primary healthcare professionals lowered BMI by 1.8 over 1 year. 80 This study highlights that tertiary prevention is effective, albeit much more costly than primary prevention, underscoring the importance of strategies to prevent child and adolescent overweight and obesity before they manifest. Primary healthcare interventions were not included in this study because the selected interventions were aligned with immediate priorities for the Mexican government, with a window of policy opportunity. While there is limited literature on the specific synergistic effect of multisectoral interventions, the authors recognize that for interventions to be successful at reversing the overweight and obesity epidemic in Mexico and globally, they will need to be multisectoral and integrated within countries' food systems, retail, transport, urban planning, media, education, health, and recreation sectors. 81 We gathered evidence from the literature on the effect size in terms of BMI reduction or reduction in overweight or obesity prevalence and the costs of each included intervention. Note: This table details the interventions selected for the analysis, including information on their baseline coverage in Mexico, their individual effect sizes, their unit costs per child (2020 USD), and the corresponding data sources.
(0-19). Social marketing in schools and school-based interventions were modelled for ages 6-17 and breastfeeding promotion for 0-12 months. All costs were discounted at 3% annually. See Table S2 for a comprehensive summary of the baseline intervention coverage in Mexico and the target coverage of implementation we modelled.
We assumed that the effects of the interventions are realized 1 year after their implementation. 25 We then estimated the impact of the interventions on YLLs, YLDs, and DALYs saved, healthcare costs averted, productivity gained, and lifetime wages gained due to increased educational attainment.
We calculated the effect of each intervention on baseline BMI of the model cohort for each age and sex group(s). We assumed that the change in BMI was maintained into adulthood and estimated the resulting reduction in projected overweight and obesity prevalence by sex and age. The analysis also estimates the effect a multilevel intervention strategy has on BMI by combining the effect sizes from each intervention, Eff i , as detailed by Watkins and colleagues 91 : As the literature suggests that multilevel interventions have a stronger impact on BMI among children and adolescents than interventions implemented individually, the ROI associated with the combined package of interventions from the main analysis may be a conservative estimation. 20,21 We estimated the impact of the selected interventions on reductions in mortality and morbidity using a potential impact fraction (PIF).
PIF is the proportional change in mortality and morbidity attributable to a change in exposure to a risk factor due to implementing the interventions in a population. For each obesity-attributable condition, the PIF is calculated as where P j is population distribution of each BMI category j (healthy, overweight, and obese) in the baseline scenario, c P^is the population distribution of BMI category j in the intervention scenario, and RR j are the relative risks of mortality and/or morbidity due to each overweight and obesity-attributable condition for each BMI category  Table S2 for more details on the baseline and intervention scenarios).

| Indicators
The investment case includes two indicators of the efficiency of the interventions. The first is the cost per DALY averted through the interventions. This offers a means of comparing the cost of an intervention to gains in key outcomes, such as morbidity and mortality. An intervention is considered low cost if its cost per DALY averted is less than three times the GDP per capita. 93 Interventions are considered very low cost if their cost to avert a DALY is less than one times GDP per capita.
The second indicator of efficiency is a return on investment (ROI) analysis. The ROI (or benefit-cost ratio) is calculated by dividing the total economic value gained from the interventions by the cost to implement the interventions. The ROI analysis compares the cost of implementation to all economic benefits-averted mortality (YLLs), healthcare costs averted, and wages and productivity gained.
The analysis reviews these indicators of efficiency over 30 years, 50 years, and lifetime horizons to capture all salient costs and benefits of the intervention. 94

| Sensitivity analysis
We the economic value of premature mortality. 23,96 Additionally, as there is a growing discussion around using a higher discount rate in low and middle-income countries to illustrate a preference for receiving benefits earlier, we have also conducted an additional sensitivity analysis that applies a higher discount rate of 5%. 97 Finally, while the analysis estimates the combined impact of implementing a package of interventions on the assumption that the impact of the interventions is additive and independent, the sensitivity analysis examines the ROI results of a less-than-additive impact and a more-than-additive impact by estimating the ROI of the combined package of interventions if each intervention's impact was reduced by 50% and if each intervention's impact was increased by 10%.

| RESULTS
3.1 | Current health and economic impacts of child and adolescent overweight and obesity Figure S1 shows the different components that make up the economic impact of child and adolescent overweight and obesity that are included in the investment case.  (Table 4).
In addition to healthcare costs, the economic impact of the indirect costs of reduced productivity, lost wages due to lower educational attainment, and the cost of lost human life are also substantial.
The total costs of child and adolescent overweight and obesity are presented in Table 4.  Note: This table provides the health outcomes for the model cohort if Mexico were to continue under a business-as-usual scenario. Only diseases and conditions causally associated with high BMI were included (see Table S1).

| Health and economic benefits of interventions
T A B L E 4 Direct and indirect costs attributable to childhood and adolescent overweight and obesity, 2026-2090 total. Note: This table provides the direct and indirect costs attributable to childhood and adolescent overweight and obesity if Mexico were to continue under a business-as-usual scenario. Costs are expressed in currency of 2020. The value of a YLL is GDP per capita for each year from 2026 to 2060 using the GDP projection for Mexico obtained from the OCED and held constant after 2060 due to unavailability of projection data. Cost only includes direct healthcare cost due to childhood obesity and does not include cost due to childhood overweight, as evidence and data on the cost of childhood overweight (nonobese) is limited.  (Table 7). Table 8      Over a lifetime, the estimated savings from averting child and adolescent overweight and obesity-attributable premature mortality is approximately 2.5 times Mexico's total annual healthcare expenditure in 2019. 98 The average annual savings from averting premature mortality is the equivalent of 2% of Mexico's annual expenditure in 2019. 98 Savings from lifetime improvements in productivity and the reduction of healthcare expenditures equal 30% of Mexico's annual health expenditure in 2019. 98   Mexico. Fifth, our estimate of the additional healthcare cost due to child and adolescent overweight and obesity included only the additional cost associated with obesity and did not include overweight due to limited data on the additional healthcare cost of overweight.
Additionally, while healthcare costs may rise at a different rate than other goods and services, as data to extrapolate the cost of healthcare over time is unavailable, we applied the CPI to address the possibility of inflation. Sixth, as age-and sex-specific rates of productivity reduction attributable to obesity in Mexico is not currently available, the model applies fixed parameters for obesity-related absenteeism and presenteeism rates from a nationally representative study of Mexico's population. 64 Seventh, while the selection of interventions was partly based on the recommendations and perceptions of the feasibility of a range of key informants, the analysis did not systematically assess their interview responses to compare implementation feasibility. Further country-specific systematic assessments of child and adolescent overweight and obesity interventions will allow stakeholders to create tailored solutions to address the growing burden of overweight and obesity among youth. Eighth, the study assumes that the intervention effect sizes are additive and independent, as data describing the positive or negative coincident effect of multiple interventions is limited.
However, the literature suggests that multilevel interventions have a stronger impact on BMI among children and adolescents than interventions implemented individually. 101,102 Further research should investigate the effects of a combination of interventions compared to individual strategies. 103,104 Ninth, the analysis does not examine the potential equity implications across socioeconomic groups due to an absence of data disaggregated by socioeconomic groups. Moreover, the study assumes that each intervention is applied equally across sexes, limiting the analysis' capacity to compare the implications by sex. Future analyses should further investigate the potential equity implications of implementing interventions to address child and adolescent overweight and obesity. Finally, since the focus of our model is child and adolescent overweight and obesity, it did not account for the additional spillover benefits resulting from the interventions, such as improved cognition and immunity among children and protective effects against cancers in mothers resulting from breastfeeding, as well as the potential for the population level interventions to impact upon adult overweight and obesity.

| CONCLUSION
This study's findings provide Mexican policymakers with evidence for a robust policy response to the epidemic of child and adolescent over-

ACKNOWLEDGMENTS
We acknowledge UNICEF, supported by a grant from Novo Nordisk.