Principal investigator: Omar Barroso Khodr
Author (s): Ciro Avitabile
Original title: Spillovers and Social Interaction Effects in the Demand for
Preventive Healthcare: Evidence from the PROGRESA program.
Location of the Intervention: Mexico (a country in North America)
Sample Size: 506 villages (rural regions of Mexico)
Sector: Health Economics
Primary Variable of Interest: Health screening decision i in a location j at time t.
Type of Intervention: Financial incentives for women to undergo cervical cancer screening.
Methodology: Indirect Treatment Effect (ITE – indirect treatment effect); Average Treatment Effect (ATE – average treatment effectDiff-Diff
Summary
This article analyzes the randomized study design of a large social assistance program in rural areas of Mexico—PROGRESA—aiming to investigate the occurrence of side effects on the propensity for screening for gender-specific and non-gender-specific conditions. The author identified significant evidence of increased demand for cervical cancer screening (Pap smear) among women ineligible for the Program. Conditional Income TransferOn the other hand, similar externalities were not observed in examinations considered gender-neutral, such as blood pressure measurement and blood glucose tests. Thus, the data suggest that the weakening of the social norm reflecting husbands' opposition to screening their wives may be one of the mechanisms driving this indirect effect.
- Policy Problem
The article highlights several policy issues related to the indirect effects (externalities) of programs of Conditional Income Transfer (CCT from English) Conditional Cash Transfers), such as PROGRESA (now called Oportunidades/Prospera), on health screening behavior, especially for cervical cancer. The focus on the Mexican program is due to the main challenges faced by public health policies, which include six different points.
Firstly, to develop effective incentives to reverse the low demand for preventive care in developing countries. The authors observed that, even with the offer of subsidies, adherence to preventive health services remains limited—especially in the case of screening tests for cervical cancer, blood pressure, and diabetes. Furthermore, the study seeks to understand the barriers that hinder this adoption and to investigate how financial incentives, such as CCT programs, can influence people's behavior regarding preventive screenings.
Subsequently, the study investigates possible side effects of CCT programs, such as stimulating the search for health services by ineligible populations. The PROGRESA program requires eligible (low-income) families to undergo health screenings to receive benefits. In this context, the study analyzes whether the program indirectly influences ineligible (non-poor) families to also seek medical examinations—especially those for cervical cancer screening (a gender-specific condition) compared to screenings for non-gender-specific conditions such as diabetes and hypertension.
Thirdly, understanding the role of social norms in healthcare-seeking behavior. The authors investigate whether social norms—such as the stigma associated with male doctors performing examinations related to sexual activity—change as more women undergo examinations as part of the PROGRESA program. The challenge was to test hypotheses about the program's impact in areas with higher participation, seeking more significant effects on gender-specific examinations (such as cervical cancer) than on non-gender-specific examinations. The researchers also examined whether the effects were more intense among women in more rigid social contexts, such as male-headed households.
Next, explore alternative explanations for spillover effects (spilloversThis study examines the impact of new social and informational dynamics on the decision to seek treatment. The social dynamics analyzed include: Information sharing between eligible and ineligible families; Female empowerment, with changes in decision-making within families; Improvements in the provision of health services, driven by PROGRESA; and Economic spillovers, such as increased income and consumption among ineligible families due to the program's local effects.
Fifth point: evaluate the cost-effectiveness of health interventions. Researchers seek to understand whether CCT programs generate positive repercussions that justify their implementation from a cost-benefit perspective.
Finally, the authors analyze the political implications for CCT programs and the design of public health policies. They propose that policymakers consider the influence of social norms when developing interventions. Thus, they investigate whether programs like PROGRESA can generate broader social benefits, going beyond direct beneficiaries. In doing so, they suggest that gender-specific health interventions—such as cervical cancer screening—can benefit from community awareness campaigns aimed at reducing stigma.
In summary, this study contributes to a better understanding of the indirect effects of social welfare programs, going beyond the traditional spillover effects on income and consumption. Furthermore, it offers relevant evidence on how peer influence and social norms shape health-related behavior in low-income contexts. Finally, the results can enrich the debate on strategies to increase adherence to preventive medical care in developing countries.
- Policy Implementation Context
PROGRESA is an income transfer program aimed at combating poverty among low-income Mexican families. According to Angelucci & De Giorgi (2009) and Adato et al. (2000), in November 1999, the average monthly subsidy was 200 Mexican pesos — equivalent to approximately R$ 58 in 2025 values — per family, representing about 23% of these families' food consumption.
Eligibility for the program is determined based on the poverty level, calculated from permanent income data obtained through a census conducted in 1997. Income transfers are targeted to women in beneficiary families. To receive the benefit, children must be enrolled and regularly attending primary and secondary school. In addition, families must participate in periodic health checkups, aimed at pregnant women, young children, and adults. Beneficiaries also need to attend health and nutrition classes, especially geared towards mothers.
The health services offered include vaccination, prenatal care, disease prevention, and medical examinations. In parallel, PROGRESA works to strengthen the provision of these services by investing in medical equipment, medications, and the training of healthcare professionals.
In contrast, people who do not participate in the program rarely attend health classes, even though they are allowed to do so. Ultimately, the central goal of PROGRESA is to reduce poverty by promoting education, access to health, and improving nutrition, using conditional cash transfers. The program also contributes to strengthening health infrastructure, supporting the achievement of these goals.
- Evaluation Details
The study encompassed 506 poor rural villages in seven Mexican states. The treatment group consisted of 320 villages that began receiving benefits from the PROGRESA program starting in May 1998. Conversely, the control group included 186 villages that remained without access to benefits until November 1999.
The period designated as the "pre-program" began in 1997 and extended until March 1998. The "post-program" was divided into three phases of analysis: the first in October 1998, the second in May 1999, and the last in November 1999. The initial study sample included 24.077 households, later reduced to 15.566 households with children, focusing on cervical cancer screening analysis. Participant eligibility was fixed until November 1999. In this context, 54% of households initially classified as non-poor (referred to as "densely populated") were subsequently reclassified as eligible but excluded from the analysis.
Several surveys and data collection efforts were conducted throughout the years of the study. The first, of an administrative nature (1999), involved 338 health centers (SSA/IMSS) in villages participating in PROGRESA, with the objective of evaluating the provision of medical care—including equipment, services, and personnel. Subsequently, the 2007 Rural Assessment Survey (ENCEL) focused on the originally evaluated communities and on new control villages, selected through propensity score matching. In this phase, individual screening data were collected for 4.279 women aged 18 to 50, covering cervical and breast cancer screenings, diabetes, and hypertension.
Researchers looked for screening indicators for various health conditions, such as cervical cancer (Pap smear), diabetes (blood glucose test), and hypertension (blood pressure measurement). The results revealed that families classified as "non-poor" had higher screening rates—for example, 38% for cervical cancer, compared to 28% among poor families. Similar differences were observed for diabetes (30% vs. 25%) and hypertension.
Despite the progress, data collection had limitations. “Huge” households, although later deemed eligible, were excluded from the analysis. Furthermore, there were gaps in the data, such as restricting the pre-program analysis of cervical cancer to women with children. Data on the provision of health services were also based on administrative reports, not on direct observations by participants.
Finally, the conditional cash transfers promoted by PROGRESA appear to have encouraged preventive health checkups, especially in areas that received treatment earlier. Long-term data (2007) suggest sustained effects on women's use of health services.
- Method
The research aimed to evaluate the Indirect Effect of Treatment (Indirect Treatment Effect – ITE) of the PROGRESA program on ineligible (non-poor) households, compared to the Average Treatment Effect (Average Treatment Effect – ATE) in eligible (poor) households. The analysis focused on the rates of health screening, such as cervical cancer screening, blood pressure measurement, and blood glucose testing.
The authors used the Difference-in-Differences (Diff-in-Diff) model to estimate the impacts of the program, comparing treated villages with control villages before and after the implementation of PROGRESA. Controls for pre-existing differences, such as disease prevalence and recall bias, were included. The dependent variable of the study was the screening decision of household i, in village j, at time t. The explanatory variables included: I. Village treatment (1 if j was treated; 0 otherwise); II. Program period (1 for post-program; 0 for pre-program); and, III. Household/locality controls (age of residents, literacy level, and poverty index).
The main findings indicated a 4,9 percentage point (pp) increase in cervical cancer screening (Pap smear tests) among ineligible households, attributable to the ITE. However, no significant changes were observed in blood pressure or blood glucose testing rates. The magnitude of the ITE (at approximately 12% of the pre-program rate) is consistent with previous evidence on indirect effects on food consumption and education. Furthermore, an increase of approximately 20 pp was observed in the performance of all tests among eligible households, a result of the conditionality imposed by the program.
Statistical tests revealed that 'ITEs' varied significantly between screening types (p = 0,029), while 'ATEs' did not show statistically significant differences (p = 0,671). PROGRESA directly promoted increased screening in eligible households due to mandatory compliance with conditionalities. Spillover effects were observed only in gender-specific tests—such as cervical cancer screening—among ineligible households, likely due to improved healthcare service provision.
Additionally, health education classes had a significant impact, resulting in approximately 195 additional Pap smear tests performed on ineligible women, with an estimated 608 lives saved based on cervical cancer detection rates. Finally, the results suggest that the program reached approximately 40% of rural Mexican families, totaling 2,6 million eligible households by 1999.
- Main results
This study examines how health interventions, such as Mexico's PROGRESA program, can create side effects that influence healthcare demand beyond its original target population. Specifically, the study investigates whether PROGRESA's requirement for cervical cancer screening among eligible households affected screening rates among ineligible women in the same communities. The results reveal that while the program generated a positive indirect effect on cervical cancer screening (a gender-specific test), it did not demonstrate a measurable impact on non-gender-specific health screenings, such as blood pressure or diabetes checks.
The authors' analysis explores several potential explanations for this selective side effect. The results suggest that the increase in screening rates was not primarily driven by improvements in healthcare provision or greater financial resources. Instead, the evidence points to changing social norms as a likely factor. In rural Mexican communities, male opposition to female screenings—particularly when performed by male doctors—has traditionally been a cultural barrier. In this way, the program appears to provide social incentives for women-specific health screenings.
These findings have important policy implications for the design of health programs in both developing and developed countries. First, the authors highlight the need to consider potential externalities and indirect effects when designing and evaluating screening programs. Second, they emphasize the importance of directly addressing cultural barriers, particularly gender-related norms, to improve program effectiveness. A third implication suggests expanding health education efforts beyond mothers to include men, as increased male awareness of women's health issues can further improve adherence to screening. The PROGRESA experience demonstrates how conditional cash transfer programs can influence health behaviors not only through direct conditionality but also through indirect changes in social norms within the community.
- Lessons in Public Policy
Epidemiological research identifies women's anxiety and embarrassment as significant barriers to adherence to cervical cancer screening in developing countries. These psychological barriers stem from multiple sources, including men's real or perceived opposition to their wives undergoing screening, concerns about physical privacy during examinations, and the social stigma associated with a potential cervical cancer diagnosis. Studies on the implementation of PROGRESA reveal that many physicians recognized husbands' resistance to Pap smears, especially when performed by male physicians. Although economists have extensively studied social norms in various contexts, including the adoption of health services, the data from the PROGRESA evaluation do not allow for isolating social norms as the sole mechanism behind the increase in screening rates among ineligible women.
This analysis explores whether the weakening of gender-related social norms may contribute to the indirect effects observed in screening behavior. The study employs a diffusion model of social norms, in which individual screening decisions are influenced by community behavioral patterns. As more women in a community undergo screening (particularly in areas with higher PROGRESA eligibility rates), the social acceptability of screening increases, reducing the perceived sanctions for participation. This effect is particularly pronounced among women who initially faced higher social costs for violating traditional norms.
Research shows that PROGRESA increased the bargaining power of women in eligible households, as they controlled income transfers. This change led to measurable shifts in household spending patterns, with increased allocations for food and children's clothing. Similar empowering effects may have extended to ineligible women through social interactions with eligible peers or resource transfers between households. This increased bargaining power could facilitate negotiations with partners regarding participation in screening, although such an effect is not expected in female-headed households.
Although the analysis cannot definitively rule out other potential explanations, the evidence suggests that both changing social norms and increased female bargaining power likely contributed to the observed increase in cervical cancer screening among ineligible women. It is important to emphasize that these effects appear to be specific to gender-related health services rather than general health outcomes, highlighting the complex interaction between program design, social norms, and gender dynamics in access to healthcare.
References
Angelucci, E. & Garlick, R., 2016. Heterogeneity in the efficiency of intrahousehold resource allocation: Empirical evidence and implications for investment in children. Unpublished manuscript, Duke University.
Adato, M., Coady, D. & Ruel, M., 2000. An operations evaluation of Progresa from the perspective of beneficiaries, promoters, school directors and health staff. Washington, DC: International Food Policy Research Institute (IFPRI).
