Scientific Tasks in Biomedical and Oncological Research: Describing, Predicting, and Explaining

The traditional classification of studies as descriptive and analytical has proven insufficient to capture the complexity of modern biomedical research, including oncology. This article proposes classification based on scientific tasks that distinguish three main categories: descriptive, predictive, and explanatory. The descriptive scientific task seeks to characterize patterns, distributions, and trends in health, serving as a foundation for highlighting disparities and inequities. The predictive scientific task focuses on anticipating future outcomes or identifying conditions, distinguishing between diagnostic (current) and prognostic (future) predictions, and employing multivariable models beyond traditional metrics like sensitivity and specificity. The explanatory scientific task aims to establish causal relationships, whether in etiological studies or treatment effect studies, which can be exploration or confirmatory, depending on the maturity of the causal hypothesis. Differentiating these scientific tasks is crucial because it determines the appropriate analysis and result interpretation methods. While research with descriptive scientific tasks should avoid unnecessary adjustments that may mask disparities, research with predictive scientific tasks requires rigorous validation and calibration, and study with explanatory scientific tasks must explicitly address causal assumptions. Each scientific task uniquely contributes to knowledge generation: descriptive scientific tasks inform health planning, predictive scientific tasks guide clinical decisions, and explanatory scientific tasks underpin interventions. This classification provides a coherent framework for aligning research objectives with suitable methods, enhancing the quality and utility of biomedical research.