Framed by the literature on research-policy transfer, this paper explores a "real world" task of translating adolescent risk-taking data into "whole-of-system" services development. It aims to explore challenges and opportunities in using large-N quantitative data analyses of such complex constructs to inform holistic policy-making. It offers a translational research-into-policy model developed using analyses of a dataset of 5122 Tasmanian students in Years 8 and 10. This model provides three levels of translation of the data analyses aimed at meeting the needs of holistic policy-making: broad directions for how services could be linked and/or be separate; multi-service directions targeting particular risk-taking behaviours; and constellations of interventions for specific risk-taking areas. The translational model is described with reference to specific policy decision-making challenges that are about re-imagining what services should stand alone, and what could be brought together, in what ways, and to what end. The model simplifies a complex process and is incomplete; however, it offers a basis for exploring why diagnostic models of research practice often used to consider complex challenges like adolescent risk-taking may not do enough to meet the needs of policy-makers. In so doing it raises deeper questions about research practice for the twenty-first century. (Contains 1 figure.)