How can we improve communication of statistical health information? For people working with health information systems, the process of communicating statistical health information has sometimes proven to be problematic. This challenge has been targeted by psychological research, and different explanations on why communication problems arise and how to solve them have been proposed. Psychological research has especially focused on how to improve risk related health information. The key findings from this work shows how communicators of statistical health risk often presents risk statements in ways that makes it hard for people to understand what the information really means. The psychological research also shows how this problematic situation is made worse by people’s low ability to deal with, and understand numbers. One of the most important findings in the psychological research has been how specifying reference class in risk statements can improve peoples understanding. The efficiency of this strategy is supported by findings from two experiments performed as a part of this thesis. The experiment show that health risk statements related to health information systems are better understood when reference class is specified.