Mapping the expanded often inappropriate use of the Framingham Risk Score in the medical literature

Objectives: To systematically evaluate the use of Framingham Risk Score (FRS) in the medical literature and specifically examine the use of FRS in different populations and settings and for different outcomes than the ones originally developed for.

Study design and setting: We identified all the citations to the article by Wilson et al. (1998), in which FRS was originally described through ISI Web of Science until April 2011. We selected studies that stated in their abstract that they calculated or used the FRS for any reason and extracted information on publication date, population studied, outcome, or disease risk factor with which FRS was associated and study design.

Results: We identified 375 eligible articles corresponding to 471 analyses using the FRS in cohort (n = 141), case-control (n = 16), or cross-sectional (n = 314) settings. Only a minority of the cohort studies had as a primary aim to externally validate the FRS (n = 45). The studied population was different (from general or healthy) in 35 (25%) and 133 (42%) of the cohort and cross-sectional analyses, respectively. All case-control studies examined healthy controls. The studied outcome was different (from coronary heart disease) in 79 (56%) of the cohort analyses and 10 (63%) of the case-control studies. Overall, only 46 (33%) of the 141 cohort analyses examined the same outcome and population as FRS was originally developed for.

Conclusion: A large number of studies use FRS in populations and for outcomes other than the ones it has been developed for and therefore for which its performance is unknown and nonvalidated.