Given the advantages, in terms of efficiency, both users of public services as the analysts seem to discover the theoretical decentralized supply of some goods and, at a time like this, where the process of transferring skills health to the autonomous communities has been completed, one wonders if indeed the health decentralization has contributed to increasing the productive efficiency of the centers.

With the intention of bringing some element of discussion in this regard, this article proposes an analysis that assesses and compares the production efficiency of the health centers in La Rioja, Alava and Navarre, whose management centers, in the year under review (1997), was in the hands of INSALUD (Rioja) and centers in autonomous communities that, at that time, had transferred all the powers on health (Navarre and Alava).

In a context such as the production of primary care, where technological differences and there are hardly any room for maneuver in the management of the centers is low, a marked differential in favor of behavioral health centers could be transferred to the indicative Decentralization has enabled some factor other than the purely technological (such as the introduction of incentives or adequate control mechanisms, better structural adjustment or the use of the advantage conferred by proximity to the particular needs of users, etc.) with positive effects on productive efficiency of the health centers.

In any case, since the direct allocation to the decentralization of the differences in production efficiency of health facilities is difficult to prove through statistical models, this article will cause decentralization as the nature of working hypotheses.

Summary
Objective: The purpose of this study was to evaluate whether the productive behavior of health centers in autonomous communities with competence in health is more efficient than that among centers belonging to Spanish public health system (INSALUD).
Methods: The technical efficiency of 66 health centers in Alava, Navarre and Rioja was analyzed. Centers in autonomous communities that in 1997 had been granted complete authority from the central government to manage their healthcare services were compared with centers whose administration, in the same year, was still in the hands of INSALUD. The method used to measure and quantify the efficiency of these centers was Data Envelopment Analysis.
Results: Nonparametric contrast of the health centers’ mean efficiency rates revealed no significant differences in the (in) efficiency of centers from Rioja, Navarre and Alava
Conclusions: The results obtained from the model of efficiency measurement used did not indicate that decentralization improves the productive efficiency of primary care centers.
Key words: Decentralization. Efficiency. Primary care. Data Envelopment Analysis.

Of all the health facilities surveyed, half of slack variables are zero and an efficiency index unit requirements for the EDA requires that a school is considered efficient. For its part, the index assigned to other schools is indicative of the degree of inefficiency because, according to the underlying problem, such centers could increase its production to the extent that this rate exceeds the unit without the need to alter the level of resources available. Accordingly, the centers are inefficient, ineffective half rate of 40%, although this rate varies between 94.69% of CSN10 heart of Navarre and Alava CSA10 of 4%.

The comparison of the rates of efficiency and inefficiency of health facilities according to their geographic origin shows that the largest proportion of centers is efficient in Rioja (63.6%), besides being inefficient facilities where minors riojanos have potential for improvement (a 31.86% against 32.64% in Alava and 45.30% of the inefficient Navarran) (Table 3). Results of which might be inferred that the conduct of efficient and inefficient health care facilities may be affected by the different organizations and / or production that may exist between sites belonging to different health administrations. However, the contrast non-parametric (Kruskal-Wallis contrast) of the average efficiency rates of the three groups of centers, depending on the health administration to which they belong, are outlined in the sample under study shows that despite significant differences in the variables that determine the efficiency index, they are not in the productive performance of the various health centers analyzed (Table 4). For example, although riojanos centers offer consultations on average more nursing and general medicine that Navarre and Alava, as it is also used for much of nurses and doctors, the differences between the indices of efficiency of some and not others are statistically significant.