The health care community is recognizing the importance of addressing cancer not just as a biological disease but as a multifaceted issue influenced by various social, environmental and economic factors.

By incorporating social determinants of health (SDOH) into risk stratification models and utilizing trustworthy AI to eliminate bias, health care providers can more effectively identify and support populations at higher risk of cancer. This approach is crucial for improving health outcomes, reducing disparities and ensuring that everyone has access to appropriate care.

The role of social determinants of health in cancer risk

In a recent Health Pulse Podcast, Dr. Otis Brawley, Bloomberg Distinguished Professor of Oncology and Epidemiology at Johns Hopkins University, discussed how social determinants of health – such as income, education, neighborhood environment and access to health care – play a significant role in shaping an individual's cancer risk. These factors influence behaviors, exposure to carcinogens and access to preventive services, all of which contribute to the likelihood of developing cancer.

For instance, individuals from lower socioeconomic backgrounds often face higher exposure to cancer risk factors like tobacco use, poor diet and occupational hazards while also experiencing barriers to early detection and treatment.

As an example, Brawley shared some relevant breast cancer findings. “The death rate for black women in Massachusetts is somewhere in the neighborhood of 18 or 19 deaths for every 100,000. But if you go to Louisiana or Mississippi, the death rate is 32 or 33 per 100,000. It's not race as much as it is the socioeconomics and the environment associated with the socioeconomics.”

By incorporating SDOH into cancer risk assessments, health care providers can gain a more holistic understanding of patient risk profiles, allowing for more targeted interventions. This approach, which is being researched at organizations like Renown Health, can lead to earlier detection, tailored prevention strategies and ultimately better outcomes for at-risk populations.

Risk stratification: Identifying populations at risk

“Technology is already helping us figure out people who are at higher risk for certain diseases, so we might be able to intervene and prevent those diseases,” notes Brawley.

Risk stratification involves categorizing patients based on their likelihood of developing cancer or experiencing poor outcomes. Traditionally, this has been done using clinical data, such as age, family history and genetic markers. However, by integrating SDOH, health care systems can refine these models to better capture the full spectrum of risk.

For example, a patient living in an area with limited access to healthy foods and health care services may have a higher risk of colorectal cancer than their clinical data alone would suggest. By recognizing this, health care providers can prioritize resources for screening and prevention in these communities, ultimately reducing the incidence of late-stage diagnoses.

Moreover, incorporating SDOH into risk stratification helps identify disparities in cancer care. By analyzing how social factors impact cancer outcomes, health systems can develop targeted programs to address these gaps, ensuring that vulnerable populations receive the necessary support and resources.

The promise of trustworthy AI in eliminating bias

As artificial intelligence (AI) becomes increasingly integrated into health care, its potential to revolutionize cancer care is immense. AI can analyze vast amounts of data, identify patterns and make predictions that inform clinical decisions. However, AI systems are only as good as the data on which they are trained. If these systems are fed biased or incomplete data, they can perpetuate existing disparities in cancer care.

Trustworthy AI aims to eliminate bias by ensuring that AI models are trained on diverse, representative datasets that include SDOH. This approach helps prevent the reinforcement of inequities in health care, particularly for marginalized communities. For instance, an AI model used for predicting cancer risk might incorporate data on income, education and access to health care alongside clinical information. By doing so, the model can provide more accurate and equitable risk assessments.

Additionally, trustworthy AI involves continuously monitoring and evaluating AI systems to identify and correct biases. This includes engaging with diverse stakeholders in the development process, such as clinicians, patients and community leaders, to ensure that AI tools meet the needs of all populations.

New capabilities, such as model cards, offered as a component of SAS’ trustworthy AI solution, provide data scientists, researchers and clinicians with peace of mind that the models they have developed to identify individuals at risk eradicate bias and that care is provided equitably.

A holistic approach to cancer prevention and care

By integrating SDOH into risk stratification and employing trustworthy AI, health care systems can move toward a more equitable approach to cancer care. This holistic strategy improves the accuracy of risk assessments and ensures that interventions are tailored to the specific needs of at-risk populations.

For example, a health system might use AI to identify neighborhoods with high lung cancer rates and limited access to health care. By combining this data with information on SDOH, the system can deploy targeted outreach programs, offering smoking cessation support, mobile screening units, and education on lung cancer prevention. This proactive approach helps address the root causes of cancer disparities and ensures that resources are allocated where needed most.

The fight against cancer requires more than advancements in medical treatments – it necessitates a comprehensive understanding of the social factors contributing to health. By incorporating SDOH into risk stratification models and utilizing trustworthy AI to eliminate bias, we can create a more effective and just health care system. This approach holds the promise of reducing cancer disparities and improving outcomes for all patients, regardless of their background or circumstances.

Check out the video below to view the entire episode or listen via your podcast streaming platform of choice.

Learn more about how to improve health care outcomes and efficiencies. Get the new e-book Data-driven health care.

 

WANT MORE GREAT INSIGHTS MONTHLY? | SUBSCRIBE TO THE SAS INSIGHTS NEWSLETTER