The correlation of the medical and environmental data (i.e. the derivation of the dosage-response relationship) in a study such as the National Coal Board's Pneumoconiosis Field Research (P.F.R.) is subject to many complicating factors compared with the more conventional types of biological assay. Several methods have been developed within the Research to overcome these difficulties, and the new procedures are described. Each is concerned with the estimation of the direct relation between the radiological abnormality associated with simple pneumoconiosis and some single measure of the past hazard, but the basic techniques are sufficiently general to be applicable in other fields of study.
The first development involves the definition of an underlying continuous scale of radiological abnormality. This prepares the way for the derivation of the “quantitative” relation between exposure and response, to replace the “semi-quantal” relation which is inherent in the use of a small number of discrete categories of radiological abnormality. The effect of errors of observation of dosage and response on the corresponding quantitative and quantal relationships is then determined.
The second development concerns the use of a “multi-dimensional” representation of past hazard. Most of the men under observation had worked in a number of different mining occupations before their first chest radiograph was taken, but this exposure cannot be assessed in terms of dust concentrations, for which reliable data are not available. Nevertheless, it is shown that past hazard can usefully be represented by three “dimensions” corresponding to the periods spent in three main types of environment—(a) the coal-face (coal-getting shift), (b) the coal-face (preparation shift) and (c) elsewhere underground. Each man's past exposure up to the time of his first chest radiograph can be expressed in terms of these three dimensions and the effect of each environment separately can be determined.
The third development extends the multi-dimensional approach to cover not only the working history before the first medical examination, but also the recorded exposure (in terms of measured dust concentrations), to which each man has been subject between the first and subsequent “follow-up” surveys. This measured exposure is regarded as one dimension of the man's total exposure up to the time of his second (or later) examination, and it is possible in this way to determine the direct relation between radiological abnormality and measured exposure, even when this component represents only part of the total hazard to which the man has been subject.
The application of the methods is illustrated by the analysis of some of the data which have been obtained in the Pneumoconiosis Field Research.