Mobility
Earliest studies of the impact of mobility on tuberculosis rates112 113 indicated that individuals moving from an area of low prevalence of tuberculosis to one of high prevalence experienced a very dramatic course with rapid infection, disease and death.
Presently, the marked difference in notification rates from one country to another provides the opportunity classic migrant studies. The assodation between notification rates in the country of birth with the rates in those born in the same country but who had moved to a county with a lower rate has been reported114 Further study of immigrants, comparing notification rates in immigrants to Canada from Scandinavia and from Finland95 with those in their countries of birth, showed clearly that rates were determined by the location of birth rather than location of residence. These immigrants had lived in Canada for forty years and had moved at age twenty, suggesting that the probability of having been infected with Mycobacterium tuberculosis early in life was the determining factor.
3. Time
Time comprises various elements: calendar period (used to measure trend), age, cohort and period.
Trend
Tuberculosis is seasonal44 with a peak of occurrence in late winter/early spring. This peak has also been shown to occur in children115 but a second peak in mid summer has been shown. The former may be explained by reduced ventilation and prolonged exposure associated with inclement weather but the latter remains unexplained.
The trend of tuberculosis in the absence of intervention has been described44, showing a steady decline in tuberculosis in Europe at least since the middle of the eighteenth century. Subsequent studies15 have demonstrated a steady decline in tuberculous infection in many low income countries where tuberculosis case management has been notably deficient, indicating a strong probability of decline in tuberculosis in a natural state, likely associated with the low probability of the key transitions which maintain the disease in a community.
Cohort
The unchanging picture of tuberculosis, in the face of a clearly changing level of the disease, has been shown in a series of analyses79 in which the lifetime experience of succeeding birth cohorts were compared, revealing the underlying characteristics of tuberculosis as a disease which kills small children and which preferentially affects young adults. The similar pattem (at variable levels) implies an inescapable biological characteristic which determines subsequent experience. This unchanging pattern suggests predictors of lifetime experience unrelated to environmental variables (or at least only to those associated with aging) and clearly shows the error of considering advancing age as a risk factor for the development of tuberculosis.
The cohort effect is reflected by a varying levei of this disease pattern which implies a common experience (the likelihood of becoming infected) in the early life of the cohort under study with a continuing effect throughout life but which changes from one cohort to another.
Period
The change of tuberculosis over time is often presented as a comparison from one period to another, in which comparison, the differences are presumed to be due to varying environmental conditions. Factors causing a rise in tuberculosis are likely to be related to changing environmental conditions and are "period" effects. When social disruption occurs due to war, tuberculosis mortality116 and tuberculous infection117 increase. In recent years, tuberculosis rates have been rising in other circumstances: in those countries affected by the HIV epidemic118, for obvious reasons, and in other countries, most notably eastern Europe119 for less obvious reasons.
An evaluation, using various models, of the trend in tuberculosis and the impact of medical measures was made using information from several locations120. The impact of medical measures could be more much more clearly demonstrated by a cohort approach, as compared with a period approach, in some locations but not in others, where environmental conditions were thought to explain the differences.
E. Determinants of tuberculosis
The most powerful determinants to explain the key transitions in the dynamics of tuberculosis are exposure, time since infection, host immunity and bacterial load. It is likely that, if each of these is known precisely, the dynamics of tuberculosis transmission could be predicted with a high degree of accuracy.
1. Exposure
It goes without saying that, without exposure to Mycobacterium tuberculosis, there would be no death, disease or infection with tuberculosis. This is the ultimate target of interventions to prevent tuberculosis.
The main components of exposure are concentration and duration. Exposure is the key predictor of the probability of becoming infected. The only method of direct measurement of exposure ever reported used caged guinea pigs and calculated mortality among them121 to measure contagiousness of patients during treatment. Indirect measures of exposure include: intensity of exposure (household or non household23), concentration of bacteria (estimated by degree of positivity of sputum smears122); and duration of exposure (estimated by studies of delay in onset of treatment123).
2. Time since infection
The probability of developing disease is exponentially related to the time which has passed since becoming infected27 28. This is such a powerful predictor of the transition from infection to disease that it is impossible to evaluate other predictors without taking it into account. The greatest problem in doing so, however, is the striking difficulty in determining the time of initial infection; where
