Estilo de liderança do enfermeiro: construçãoe validação de instrumento de avaliação.

Abstract

The leadership in the hospital context is subject of great interest for being a tool capable of promoting changes in the work practice and improvements in the quality of care. Given the unavailability of a specific instrument of leadership that encompassed care activities performed in any inpatient unit, the aim of this study was to design an instrument to measure the leadership styles of the nurse in the perspective of the Situational Theory, as well as to perform apparent validation and of content, and to evaluate parameter properties of the instrument. It is a methodological study, developed according to the steps: the elaboration of the instrument with 26 care activities, divided into 7 domains (primary care, medication, invasive procedures, dressings, guidelines, notes, general); apparent validation and of content by judgment of expert judges in the area; pilot study; validity of criteria from professional variables of the leader and analysis of reliability through test - re-test during an interval of 15 days. It was carried out in a school hospital in the countryside of Minas Gerais, with the participation of 46 nurses paired to 46 technicians/assistants. As to the validity, all care domains presented discriminative power for a desired leadership however there was no statistically significant difference. Considering management as the most important criteria for invasive procedures, the analysis of a posteriori power revealed that the statistical difference for a significance level of 0.05 would be reached with a sample size three times bigger. In the analysis of reliability, the primary care domain (k=0.41; p=0.01), invasive procedures (k=0.54; p<0.001); notes (k=0.42; p<0.001) and general (k=0.51; p=0.03) presented moderate agreement and were statistically significant. The medication domain (k=0.37; p=0.01) presented statistically significant p value and regular reliability. The domains dressing and guidelines did not present statistically significant differences and concordances were insignificant, K=0.01 and K=0.16 respectively. It is interesting to remark that the study indicated large proportion of discordant answers regarding the opinion of leaders and followers which might have affected the validity and reliability values. The results described indicate that the instrument Leadership of the Nurse Evaluation of the Leader and the Follower can be a valid and reliable instrument which can be applied to hospitals, but it is necessary that this instrument to be applied to larger and heterogeneous populations in order to better evaluate its metric properties.