Machine Learning Identifies Sociodemographic Cancer Patterns
A systematic review in J Med Internet Res (2026) examined how machine learning is used to detect associations between sociodemographic factors and cancer outcomes. The authors screened 328 records and included 19 studies, finding predominance of supervised methods—particularly random forest and XGBoost—common use of age, sex, education, income, and geographic variables, and reliance on cross-validation with infrequent external validation. The review urges integrating contextual social determinants, improving transparency, and expanding external validation for greater equity and generalizability.
Key Points
- 1Identify 19 of 328 studies applying ML to sociodemographic-cancer associations, mostly supervised methods.
- 2Highlight dominance of random forest and XGBoost, with cross-validation prevalent but external validation scarce.
- 3Recommend integrating contextual social determinants, enhancing transparency, and conducting external validation for generalizability.
Scoring Rationale
Comprehensive systematic review across oncology ML provides actionable recommendations, but limited study count restricts generalizability.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
