Torbjörn Andersson

INTRODUCTION This chapter deals with the ex officio application before Swedish courts of the Unfair Terms Directive. The Directive contains two related, but different, means of providing consumer protection before domestic courts and other public enforcement mechanisms (in the wide sense). One of the means concerns consumer protection in individual cases and the other aims at consumer protection as a public interest. The former draws on the traditional tools and objectives of civil procedure (individual compensation with a retrospective focus on the circumstances in the individual case), while the latter is designed to allow for public interference with sellers and suppliers for the prospective objective of correcting individual abusers, but primarily to clarify the law and deter future abuses. The following will be limited not only to one aspect of the general subject but also to one part of that aspect: the obligation on Swedish courts to act of their own motion to safeguard consumer interests in individual cases where consumers are defendants. First, the public enforcement of consumer protection will be dealt with only insofar as it concerns individual disputes where unfair terms in consumer relations are of relevance. Second, situations where consumer protection emerges in cases where the consumer acts as claimant will be left out. This delimitation may be surprising since recently the EU legislature has been particularly active in relation to public enforcement and devising mechanisms for conflict resolution where consumers are initiators. Still, I would say that the most frequent way in which an unfair term under a consumer contract comes to be tried before a court or a quasi-court (or to be settled without the involvement of the judiciary) is where the seller/supplier directs a claim for payment against the consumer. Hence, I have considered it relevant to use that angle in a discussion on the practical implementation in Sweden of the requirements of consumer protection before the courts. The focus will be on the requirements of the Directive, as revealed through the case law of the European Court of Justice (ECJ), in relation to the obligation on national courts to protect consumer interests ex officio.

In diagnostic radiology increasing attention has been focused on dose reduction while maintaining a clinically good image quality. With the use of digital detectors balancing dose vs image quality is done differently than in film-screen radiography, since dose and image brightness are uncoupled in digital imaging. In this study a new direct digital detector (flat-panel detector) was used in a dose-image optimisation of a simulated pelvic examination. X-ray images were taken with a direct digital detector (DDD), of the pelvic of a phantom using varying tube current (varying stochastic noise). The entrance surface dose was measured for each image. These images were scored by two radiologists according to EU guidelines. A dose comparison was made with an older PCR system (storage phosphor plates). With decreasing tube current the noise in the images increased and the image with the lowest dose and still acceptable image quality was identified. The results showed that the entrance surface dose using the DDD decreased from 1.4 mGy (PCR value) to 0.48 mGy (DDD standard settings). Through the optimisation the dose could be further decreased to 0.24 mGy while still maintaining an acceptable image quality. A substantial dose reduction was obtained with this new direct digital detector. This simple but efficient optimisation approach is easily applicable to other examinations and both DDD and storage phosphor plate detectors.

In identification from a deterministic point of view an algorithm is said to be robustly convergent if the true system is regained when the noise level tends to zero. In this paper we introduce a concept close to this performance measure: robust global identifiability. A model structure, i.e. a smoothly parametrized set of models, is said to be robustly globally identifiable if there exist an identification algorithm such that the true parameters are regained when the noise level tends to zero. We show that global identifiability implies robust global identifiability when the model structure in consideration is a characteristic set of differential polynomials.<>