A Norwegian system for defining and classifying DRPs is proposed. The system builds on cross disciplinary agreement between physicians and pharmacists from various clinical and scientific positions. The classification is a tool to handle challenges in relation to drug treatment and the system could contribute to improved documentation of various problem areas.
The panel’s professional and geographical heterogeneity contributes to the classification system’s relevance for various aspects of the drug treatment (prescription, monitoring, use, documentation), for various aims and for different parts of the health services. Although it was a goal to include all Norwegian experts in the field, and it should be simple enough to get an overview of the professional environment in the country, it is possible that not all have been included.
Consensus-based procedures are suitable for integration of research-based and experience-based knowledge. The modified Delphi technique is an established method for development of clinical guidelines and quality indicators (17). This method ensures that various meanings are promoted, independent of the participants’ relations, position and status. Communication by e-mail enables participation of experts who are geographically far apart. On the other hand, the method is demanding and lack of discussion may prevent identification of good ideas and elimination of bad ones.
Van Mil and collaborators have assessed 14 published classification systems of DRPs (18). The group points out that classification systems should be validated and also that the results of this procedure should be published. However, only a few of the classifications have been validated. We have gone through a case-based validation procedure among a heterogeneous review panel to assess the content of the classification and to reveal validity (face).
The classification system has an open hierarchical structure that can be adapted and expanded with several categories according to need, setting and access to clinical information. The intention was to construct a general model that comprises many different problem areas and at the same time prioritize simplicity and flexibility rather than in depth detailed descriptions. Previously published classifications have been considered to ensure comparability with international models.
We have chosen to include both actual and potential problems in the definition of DRPs (15). This choice is founded in our understanding of the importance of identifying problems before they have become manifest and thereby prevent a possible negative outcome, as for example lack of effect or increased morbidity. Both potential and actual problems can be identified by conducting regular systematic reviews of patients’ total drug use.
The participants in the hearing group agreed that undertreatment («need for additional drug») would be part of the classification system. This problem is not strictly associated with one or more specific drugs, but rather to a presumption of effective treatment or to adherence to guidelines to prevent disease; e.g. anticoagulation after a heart attack. Our view coincides with that of Van Mil and collaborators. They criticize the lack of undertreatment as a category in several published classification systems and point at evaluation of treatment effectiveness of a certain condition as a crucial part of medication reviews (18). However, this presumes access to relevant clinical information such as symptoms and laboratory tests, which in some settings will be inadequate, for example in pharmacies.
Validation of the classification system showed an average of 70 % agreement on choice of category. Limit values had not been predefined as there was not sufficient published material to base such definitions on. Our findings are however in agreement with an Australian validation procedure for classification of DRPs; they found an agreement of 69.9 % (16). For some cases there was a larger variation in the classification. Some were relatively heterogeneous and it was challenging to classify these as one single problem in one single category. It is known that such validations render partition between processes difficult, i.e. problem perception and classification in itself (18). Belonging to a professional group did not affect the choice of category and this is in favour of the categories’ lack of ambiguity and the system’s robustness.
We considered the agreement to be sufficient to use the classification in research projects and clinical practice, for example in communication between physicians, pharmacists working in clinical settings or pharmacies and with patients. An evaluation and possible revision of the classification should be done after it has been used for a while.