[Evaluation in the health sector: concepts and methods]

The practice of evaluation has existed in one form or another for as long as one can remember and is central to all processes of learning. Today, evaluation is a popular concept grouping together multiple and diverse realities. This article aims to propose a conceptual framework for evaluation that is broad and universal enough to allow all those concerned with evaluation of health services (regardless of their disciplines and interests) to better understand each other, to perform better evaluations, and to use them in a more pertinent manner. We will begin by defining evaluation as the process which consists of making a judgement on the value of an intervention by implementing a system which can provide scientifically valid and socially legitimate information on regarding this particular intervention (or any of its components) to the different stakeholders concerned, such that they can form an opinion from their perspective on the intervention and reach a judgement which can translate into action. We define "intervention" as any organized system of action (a structure, actors and their practices, processes of action, one or many finalities and an environment) aiming to, in a given environment, during a given time period, modify the foreseeable course of a phenomenon to correct a problematic situation. An intervention can be a technique, a medication, a treatment, an organisation, a program, a policy or even a complex system like the health care system. Various interventions, regardless of their nature, can be the object of two types of evaluation. Normative evaluation is based on appreciation of each component of the intervention according to criteria and standards. This type of evaluation is defined as an activity which consists of making a judgement regarding an intervention by comparing the resources utilized and their organisation (structure); services and goods produced (process) and results obtained to criteria and standards (in other words, summaries of what is good and right). Did the intervention correspond to what should have been done according to the standards utilized? Evaluative research aims to employ valid scientific methods to analyze relationships between different components of an intervention. More specifically, evaluation research can be classified into six types of analysis, which employ different research strategies. Strategic analysis allows appreciation of the pertinence of an intervention; logical analysis, the soundness of the theoretical and operational bases of the intervention; productivity analysis, the technical efficiency with which resources are mobilized to produce goods or services; analysis of effects, effectiveness of goods and services in producing results; efficiency analysis, relations between the costs of the resources (or the services) used and the results; implementation analysis, appreciation of interactions between the process of the intervention and the context of implementation in the production of effects. The official finalities of all evaluation processes are of four types: (1)strategic, to aid the planning and development of an intervention, (2) formative, to supply information to improve an intervention in progress, (3) summative, to determine the effects of an intervention (to decide if it should be maintained, transformed or suspended), (4) fundamental, to contribute to the advancement of empirical and theoretical knowledge regarding the intervention. In addition, experience acquired in the field of evaluation suggests that evaluation is also productive in that it allows actors, in an organized setting, to reconsider the links between the objectives given, practices developed and their context of action. This task of achieving coherence is continuous and is one of the intrinsic conditions of action in an organized setting. In this perspective, evaluation can have a key role, given that it is not employed to legitimize new forms of control but rather to favor debate and th