Pathological findings in mine workers: II. Quality of the PATHAUT data

To assess the feasibility of using the pathology automation system (PATHAUT) for research, the quality of the data was explored by examining who comes to autopsy, the quality of the autopsy material, interobserver variability, and repeatability of diagnoses. The data indicated that autopsy rates in the gold mining industry, especially for whites, are high and that even among blacks, gold miners are represented in proportions exceeding the relative size of the working population. Because of the perception of the autopsy service as a means of obtaining compensation, miners with occupational diseases fully compensated in life are probably underrepresented. The autopsy material submitted for full autopsy is generally better preserved than cardiorespiratory organs that are sent for examination. The gold mining industry has a high proportion of full autopsies as does the Iron and Steel Corporation of South Africa. Full autopsies are more commonly performed on older deceased miners. This was true for both blacks and whites. The allocation of material to pathologists for full autopsies and examinations of the cardiorespiratory organs were clearly not random, and this may affect comparisons among pathologists. Active tuberculosis, silicosis, and emphysema prevalences appeared fairly comparable across pathologists; however, there was wide variability in the prevalence of bronchiolitis as determined by the pathologists. Agreement between the diagnoses on PATHAUT and reclassifications by single pathologists was very good for the severity of emphysema and the histological type of bronchogenic carcinoma.