AI exposes deep inequality in breast cancer screening across India

The study finds that breast cancer screening uptake in India is extraordinarily low, with fewer than one percent of women in the target age group reporting that they have ever been screened. This figure highlights a stark disconnect between policy intent and real-world outcomes. While national guidelines promote early detection as a cornerstone of cancer control, implementation has failed to reach most women, particularly those in rural and economically disadvantaged settings.

Breast cancer is now one of the leading causes of cancer-related deaths among women in India, yet preventive screening remains almost entirely out of reach for large segments of the population. Despite national screening programs and expanding healthcare infrastructure, most Indian women have never been screened for breast cancer, leaving early detection efforts critically underused. New research published in Frontiers in Artificial Intelligence shows that this failure is not random but driven by deep and measurable social, economic, and structural inequalities.

The study “Predicting and Identifying Correlates of Inequalities in Breast Cancer Screening Uptake Using National Level Data From India,” applies machine learning to nationally representative data to identify who receives breast cancer screening in India and why so many women are excluded from preventive care.

Screening rates remain critically low nationwide

The study finds that breast cancer screening uptake in India is extraordinarily low, with fewer than one percent of women in the target age group reporting that they have ever been screened. This figure highlights a stark disconnect between policy intent and real-world outcomes. While national guidelines promote early detection as a cornerstone of cancer control, implementation has failed to reach most women, particularly those in rural and economically disadvantaged settings.

The research shows that screening is heavily concentrated among a narrow segment of the population. Women living in urban areas, those with higher levels of education, and those from wealthier households are far more likely to have undergone screening. In contrast, women who are poor, less educated, rural, or socially marginalized face overlapping barriers that effectively exclude them from preventive care.

Geography plays a significant role. Rural women are much less likely to be screened than their urban counterparts, reflecting disparities in healthcare access, transportation, and availability of trained providers. Distance to health facilities and lack of affordable transport emerge as persistent obstacles, particularly for women who cannot travel alone or must prioritize household responsibilities over personal health.

The study also highlights the role of social norms and gender dynamics. Many women lack autonomy in healthcare decision-making, meaning they cannot seek screening without permission or accompaniment. This constraint sharply reduces screening uptake, even in areas where services are nominally available.

Overall, the findings suggest that low screening rates are not the result of individual choice or awareness alone. Instead, they reflect systemic inequalities embedded in India’s social and healthcare systems.

Machine learning reveals the drivers of inequality

The study leverages machine learning to move beyond descriptive statistics and identify the strongest predictors of screening inequality. The researchers apply models such as Decision Trees, Random Forest, and XGBoost to predict screening uptake and rank the importance of various social, economic, and access-related factors.

The results show that education is one of the most powerful predictors. Women with higher levels of schooling are significantly more likely to be screened, not only because they may have better health literacy but also because education often correlates with greater autonomy and access to information. Household wealth follows closely, reinforcing the role of financial capacity in accessing preventive healthcare.

Contact with community health workers emerges as another critical factor. Women who have interacted with frontline health workers are more likely to be screened, underscoring the importance of local outreach and trust-building in preventive care. These workers often serve as the primary link between the healthcare system and underserved communities, particularly in rural areas.

Women’s autonomy in healthcare decision-making is also a major determinant. The ability to seek medical care independently, without needing permission or accompaniment, substantially increases the likelihood of screening. This finding highlights how gender inequality directly translates into health inequality.

The study further uses inequality decomposition methods to show how these factors interact. Economic and educational disparities do not operate in isolation but combine with access barriers and social constraints to reinforce exclusion. For example, a poor rural woman with limited education and low autonomy faces compounded disadvantages that make screening highly unlikely.

Targeted interventions needed to close the gap

The authors argue that universal screening strategies, while well intentioned, are insufficient without targeted approaches that address the specific barriers faced by disadvantaged women.

Improving screening uptake will require more than expanding facilities or issuing guidelines. The research points to several areas where intervention could reduce inequality. Strengthening the role of community health workers is one such area. By expanding outreach, education, and follow-up at the local level, health systems can reach women who are otherwise disconnected from formal care.

Improving women’s access to information is another priority. Health communication strategies must be tailored to low-literacy and rural populations, using culturally appropriate messaging and trusted channels. Digital health tools may play a role, but only if they are designed with inclusivity in mind and supported by human intermediaries.

Addressing gender norms and autonomy is a more complex challenge but a necessary one. Policies that support women’s ability to seek healthcare independently, including safe transport and women-friendly services, could have a measurable impact on screening rates.

The study also highlights the value of data-driven policymaking. By applying machine learning to large-scale health data, policymakers can identify high-risk groups and allocate resources more effectively. Rather than relying on broad averages, interventions can be designed to reach those most likely to be left behind.