Most cardiovascular disease (CVD) risk prediction algorithms are designed for use in individual patients and cannot be applied population-wide as they incorporate variables that are inaccessible in routine health datasets. This research will: 1) develop 'low-information' risk prediction algorithms through anonymous linkage of routinely-collected New Zealand and Auckland health data for people without CVD; and 2) compare the performance of these algorithms with locally-developed high-information risk prediction algorithms. Deriving low-information risk prediction algorithms that can be applied population-wide would complement CVD risk prediction at the individual level. The proposed research offers a powerful opportunity to enhance targeting of CVD primary prevention treatment and improve CVD outcomes in New Zealand and internationally. This work will also enable the applicant to further develop her epidemiological and biostatistical skills in a supervisory environment of the highest calibre, in order to establish independence as a researcher.