This wide-ranging article reviews the challenge and promise of implementing predictive analytics on large healthcare datasets. The border between causal inference and mathematical modeling is porous and often difficult to define. While modeling has the potential to produce tangible benefits for practitioners and provides a basis for research, it is difficult to design studies demonstrating causal links on the basis of register data alone. The preponderance of observational studies in the field of pre-hospital care decision making as previously noted, can be seen as the result of an increased availability of ePCR datasets in the absence of experimental research. As pressure increases on EMS agencies to adopt progressive treatment and transport protocols, high-quality research is required to ensure that such processes can be implemented in an evidence-based manner.
The requirements placed on data for modelling and causal research are intertwined, and EMS must keep pace with hospital-based medicine in improving data quality and embracing data standards to improve interoperability. It’s not just billing and accountability – With clean data in a standardized format, the possibilities for business intelligence, predictive modelling and quality research are immense.