1. Basic epidemiological terms and calculations

Epidemiology is traditionally defined as the study of the distributionand determinants of disease within a human population (Woodward,2005). To accomplish this, epidemiological studies attempt to identifyfactors which may explain or contribute to disease outbreak. Onceidentified, these factors not only inform future clinical etiologicalstudies, but also, and perhaps more importantly, they inform diseaseprevention and control programmes (Mausner and Kramer, 1985).The success of epidemiologists in reducing the occurrence of humandisease over the last century is undeniable. The identification offactors that contribute to the occurrence of diseases such as lungcancer (smoking), sexually transmitted diseases (unprotected sex),and cardiovascular disease (high blood pressure) have permittedtargeted community health initiatives aimed at preventing orcontrolling risk factor exposure. These initiatives, in turn, have helpedreduce the rate of disease in targeted populations (Mausner andKramer, 1985; Koepsell and Weiss, 2003; Woodward, 2005).

Considering the success of human epidemiology, it is notsurprising that epidemiological methods have been adopted by thosewishing to understand and reduce disease outbreak in non-humananimals (epizootiology) (Nutter, 1999). The term epidemiology is nowwidely adopted by those studying disease and disease determinants innon-human organisms, including honey bees, and will be the termused in this paper. Nutter (1999) argued that the application ofepidemiological methods for understanding disease occurrence inplant, human, and animal populations involves the implementation ofsix common steps which include defining disease in quantitative termsand quantifying state and rate variables of the disease system. Analternative way to look at this process is to consider the "virtuousepidemiological cycle" (Fig. 1) which outlines the various stepsinvolved in quantifying disease in a population, determining riskfactors contributing to disease occurrence, determining methods toreduce disease occurrence and then evaluating the effectiveness ofthese methods (Toma  et al ., 1991).

A comprehensive review of all of these steps is well beyond thescope of this paper. Similarly, much of the data used byepidemiologists are derived from surveillance efforts, a discussion ofwhich is also beyond the scope of this paper, but has receivedattention in other recent work (Hendrikx  et al ., 2009, vanEngelsdorpet al. , 2013). Instead, we focus on presenting and defining thevocabulary needed to implement epidemiological studies, and thenoutline study design, analysis and interpretation. It is also the intentof this paper to present a framework for understanding and initiatingongoing and future studies of honey bee health. Unless otherwisenoted, the following terms and concepts have been adapted fromKoepsell and Weiss (2003).

Fig. 1. The virtuous circle of epidemiology: Step 1. describe healthcharacteristics of the population in space and time (descriptiveepidemiology); Step 2. analyse data and mechanisms of developmentof the disease to understand behaviour (analytical epidemiology);Step 3. produce, select and apply control or preventive measures(operational epidemiology); Step 4. give necessary information thatpermits the follow up of measures (evaluative epidemiology); Inaddition, changing epidemiological methods should be supported bytheoretical epidemiology (modelling).

Figure 1

1.3. Measures of disease in a population