3.1 The epidemiological triad
The epidemiological triad is best represented diagrammatically (see Figure 19). This represents the interaction between an agent, host or persons and environment or place within a specific time dimension. The epidemiological triad can be applied to non-infectious diseases where the agent could be ‘unhealthy behaviours, unsafe practices, or unintended exposures to hazardous substances’ (Miller, 2002, p. 64).
Within the epidemiological triad the agent is known as a ‘necessary’ factor. It has to be present for morbidity, although it may not inevitably lead to disease. For the disease to occur it needs the combination of what have been called ‘sufficient’ factors. These would include a host, which might be an individual or group of individuals who are susceptible to the agent. Susceptibility might be on the basis of age, sex, ethnic group or occupation. Environmental factors can also be sufficient factors that combine with the agent.
With reference to the case of chlamydia as the agent in Figure 3, there is some very limited information on the hosts and their environment within a particular timescale (i.e. 2002). The map in Figure 3 shows the geographical distribution of male and female cases of chlamydia reported in genito-urinary medicine clinics. The source of that data – the statistical returns made from GUM clinics (the KC60 form) – records age as well as sex. This indicates that the highest rates of diagnoses were among women aged between aged 16 and 19 and men aged between 20 and 24, and that the rate for women begins to fall dramatically after the age of 24 (Health Protection Agency, 2003). In order to contribute to causal understanding, Bhopal suggests that three main questions need to be asked:
- How does the pattern of disease vary over time in this population?
- How does the place in which the population lives affect the disease pattern?
- How do the personal characteristics of the people in the population affect the disease pattern?
Thinking point: Which of the three questions do you think would be the most fruitful line of investigation in the case of a sexually transmitted disease?
The third question about personal characteristics would seem most useful because the risk factor is behavioural in that the practice of safe sex seems to be important.
However, identifying causal variables is far from straightforward. Aggleton (1990) cites three conditions identified by Armstrong (1983) which must be met before two variables can be said to be causally related. These are outlined below:
- The variable must be in the correct temporal sequence. The variable which is thought to be the cause must precede the one that it is predicted to affect.
- There must be a correlation between the variables that are believed to be related. As one varies, so should the other. Correlations can be positive; that is, when one variable increases, so does the other. They can also be negative; that is, as one increases, the other decreases.
- There must be no hidden or confounding variable (i.e. a variable that could also be a causative factor) causing both of the variables to change.