Background: Here we describe the Systematically Tabulated Outcomes Research Matrix (STORM) method to generate research questions from pre-existing databases with the aim of improving patient outcomes.
Materials and methods: STORM can be applied to a database by tabulating its variables into a matrix of independent variables (y-axis) and dependent variables (x-axis) and then applying each unique pairing of an independent and dependent variable to a patient population to generate potentially meaningful research questions.
Results: To demonstrate this methodology and establish proof-of-principle, STORM was applied on a small scale to the National Trauma Data Bank and generated at least seven clinically meaningful research questions.
Conclusion: When coupled with rigorous clinical judgment, the STORM approach complements the traditional method of hypothesis formation and can be generalized to outcomes research using registry databases across different medical specialties.