A comparison of conventional and grid techniques for chest radiography in field surveys

Abstract

A comparison of conventional and grid techniques for chest radiography in field surveys. The effect on the quality of chest radiographs using a reciprocating grid with a moderately high kilovoltage (96 to 105 kV) has been studied. A total of 1 710 mineworkers had two postero-anterior chest radiographs taken at the same visit to a linked pair of mobile x-ray units. One film was taken with conventional exposure factors and the other with moderately high kilovoltage and a reciprocating grid. The grid was exchanged between the two units according to a randomized plan so that the first radiograph was not always taken with the same technique.The 3 420 films so produced were subsequently assessed for quality by five doctors experienced in reading chest films. The films were examined singly in random order and the reader did not know which technique had been used for a given film.Four of the readers recorded improved quality using the grid technique for films from men whose antero-posterior chest measurements exceeded 254 mm (10 in), but they preferred the conventional exposure technique for films from men whose chest measurements were less than 254 mm. Results from all film pairs where a difference in quality was recorded showed no overall advantage for either technique.

Quantifying Radiological Changes in Simple Pneumoconiosis

“…Wise and Oldham (l963a) had noted earlier that there was real and systematic variation in the amounts of progression between groups of men with differing initial degrees of pneumoconiosis. Jacobsen et al (1971) reported a similar effect and this is not attributable to differences between individuals' dust exposures during the periods concerned (Jacobsen, 1973). It appears that, under similar dust conditions over the same time periods, men with no radiological evidence of pneumoconiosis (categories 0/or 0/0) are less likely to show subsequent radiological changes than men with initial X-rays classified higher on the scale.…”

“…Evidently the assumption of what Kostal calls "quasi-stationary environmental conditions" in particular collieries over lO-year periods can be misleading in practice. Even mean concentrations calculated from men's individual exposures are, in fact, random variables rather than fixed dose metameters, since individuals' cumulative exposures over particular periods were calculated by summing the products of sample mean dust concentrations relevant to occupational groups within collieries and the time spent by individuals in those groups (Jacobsen, 1973). Ashford (1959a) considers the problem of analysing dose-response relationships 232 APPLIED STATISTICS when the dose metameter is subject to random errors, and he describes how ordered radiological classifications may be treated as multinomial or semi-quantal responses.…”

“…Various studies have demonstrated that the technical quality of a chest X-ray may affect its classification for abnormality (Fletcher and Oldham, 1949;Liddell, 1961;Wise and Oldham, 1963b;Pearson et al, 1965;Reger et al, 1972) but results have not been consistent with respect to how different technical imperfections affect the 242 APPLIED STATISTICS classifications. Differing definitions of what constitutes "good" film quality obfuscate inter-study comparisons (Pearson et al, 1965) and indeed they influence dramatically anyone reader's assessments of the same film (Washington et al, 1973). Differences between readers in their subjective interpretations of what is nominally the same quality criterion confuses the issue further (Liddell, 1961;Reger and Morgan, 1970;Washington et al, 1973).…”

“…Differing definitions of what constitutes "good" film quality obfuscate inter-study comparisons (Pearson et al, 1965) and indeed they influence dramatically anyone reader's assessments of the same film (Washington et al, 1973). Differences between readers in their subjective interpretations of what is nominally the same quality criterion confuses the issue further (Liddell, 1961;Reger and Morgan, 1970;Washington et al, 1973). Notwithstanding these difficulties Liddell and Lindars (1969) showed that the 4-point ordinal scale could be extended usefully into a 12-point ordinal scale simply by noting whether or not an adjacent main category was considered seriously before the final classification was made.…”