In the framework of a quality analysis project for the improvement of digital subtraction angiography (DSA) equipment, an inventory was made of the image quality and radiation dose of DSA equipments in six hospitals in the Netherlands. The image quality was investigated with a contrast detail (CD) phantom. The entrance dose of the radiation on this phantom and the skin dose at the level of the eye lenses and the thyroid gland were measured in these hospitals using a human phantom during a standardised simulated DSA examination of the aortic arch and brachiocephalic arteries, by means of thermo-luminescence dosimeters (TLD). To establish the relation of these measurements on the human phantom and real patient examinations, the same measurements were carried out in our own hospital on 16 patients during a comparable DSA examination. To find the difference from the dose in conventional angiography (CA) the same measurements were carried out in our hospital on 11 patients during a comparable examination. These dose measurements were also carried out on the human phantom with the use of the same CA equipment. We found large differences in image quality in the various hospitals. Within one hospital, monitor images were better than hard copy images. These differences were strongly related to the amount of radiation used, to the technique of storing the images (digital or analogue) and to the quality of the equipment used to make hard copies (the imager). Recommendations are made for improvement and quality control.