*Geek Box: Hazard Ratios
Hazard ratios, or ‘HRs’, are used to express the chance of an event occurring in a reference or treatment group compared to the control group, over a period of time. They are calculated by looking at the rate at which events occur, for example, how many deaths occur per 1000 person-years (‘person-years’ being a calculation based on the number of years a participant spent in a study). They are used commonly in prospective cohort studies because they reflect the time survived during the period of observation until an event (for example, a heart attack – if that is an outcome of interest) occurred. Because the risk associated with diet-disease interactions isn’t constant – someone doesn’t have a heart attack after eating one cheeseburger, but may have one over 10-years of poor diet – HR’s are used because they reflect risk over time. HR’s are useful because they provide a quantitative summary of the results that is more easily communicated. It is important to note that expressing a HR as a percentage, for example, “a 30% increase in risk for heart disease” is always the risk relative to the reference or control group. So it is always a comparison, and this is important for interpreting nutrition research, because no nutrient exists in a vacuum: if one nutrient is high, often another has been replaced by it. So ‘high’ or ‘low’ can be relative to whatever has been replaced in the diet. Lesson: always look a little closer at the data to see what the actual levels of a nutrient were, the type of nutrient looked at, and how that might relate to the differences in risk observed between two groups.