Nursing leadership in intensive care units and its relationship to the work environment 1

AIM: To establish whether there is any relationship between the work environment and nursing leadership at intensive care units (ICUs). METHOD: Correlational study conducted at four ICUs in southern São Paulo (SP), Brazil. The study population was comprised of 66 pairs (nurses and nursing technicians) established by lottery. The nurses responded to three instruments: 1) characterization; 2) a validated Portuguese version of the Nursing Work Index Revised (B-NWI-R); and 3) Grid & Leadership in Nursing: ideal behavior. The nursing technicians responded to 1) characterization and to 2) Grid and Leadership in Nursing: actual behavior, relative to the corresponding randomly-assigned nurse. The data were analyzed by means of analysis of variance (ANOVA) at p ≤ 0.05. RESULTS: The work environment was not associated with actual nursing leadership (p = 0.852). The public or private nature of the institutions where the investigated ICUs were located had no significant effect on leadership (p = 0.437). Only the nurse-physician relationship domain stood out (p = 0.001). CONCLUSION: The choice of leadership styles by nurses should match the ICU characteristics. Leadership skills could be developed, and the work environment did not exert any influence on the investigated population.


Introduction
The labor market has been increasingly demanding active leadership from nurses. To meet this demand, understanding the relationship between this variable and other variables is critical for the formulation of action plans that promote the development of this skill.
Concern with nursing leadership is crucial in intensive care units (ICUs) due to their dynamic nature, which results from the severe condition of the admitted patients and which requires nurses to provide highly complex care. Within this setting, nurses must lead their staff to achieve the best outcomes, for which leadership skills are essential. Thus, the relationship between that dynamic, interactive environment and nursing leadership represents an appropriate subject of study.
The correlation between leadership and the work environment has been studied by several authors at the institutional level but not specifically in the ICU setting.
A total of 24 studies included in a systematic review (1) showed that leadership styles focusing on people and relationships contributed to improving the outcomes of the nursing workforce, the work environment, and the productivity and effectiveness of health organizations.
By contrast, the staff satisfaction was lower in 10 studies in which leadership focused on tasks.
Transformational leadership increased the nurses' satisfaction in all institutional areas, reduced burnout, and provided a favorable work environment (2) .
Another study (3) designed a test model in which the authentic leadership of managers was combined with the nurses' perception regarding structural empowerment, performance, and personal satisfaction. The sample consisted of 280 nurses who responded to instruments for data collection on each of the assessed variables.
Leadership had a significant and positive influence on all the investigated attributes.
A study that investigated the role of organizational power and the leader's personal influence in the creation of a high-quality professional practice environment for nurses (4) found a direct and positive relationship.

Method
The present correlational study involved a systematic investigation of the nature of the relationship or association between variables (5) (6) ; and 3) Grid & Leadership in Nursing: ideal behavior (7) .
The selected nursing technicians were given an envelope that contained the following: 1) the aforementioned instrument for characterization; and 2) Grid & Leadership in Nursing: actual behavior (7) to complete relative to the corresponding randomlyassigned nurse.
As a result, pairs were formed between a nurse and the nursing technician randomly selected by him or her to evaluate the former's perception on ideal leadership behavior and the assessment made by the latter on the actual performance of the nurse as his or her immediate leader.
The participants responded to the instruments away from the work environment. A date was scheduled to hand the questionnaire directly to the principal investigator. The principal investigator did not hold a leadership position at any of the investigated ICUs.
The characteristics of each instrument are described in the following. The B-NWI-R (6) is based on the Nursing Work Index (NWI), which was drafted in 1989. It consists of 65 items. The Nursing Work Index Revised (NWI-R) (8) was formulated to make the NWI shorter and also to measure particular characteristics of the work environment that are favorable to nursing professional practice.
The NWI-R consists of 57 items, 15 of which were conceptually distributed across three subscales: autonomy, control over the practice setting, and the nurse-physician relationship. A total of 10 of these 15 items were clustered to form a fourth subscale: organizational support (8) .
Conceptually, the subscales are defined as follows (8) : autonomy (five items) and control (seven The score of each subscale is calculated based on the average scores of the responses given by subjects and ranges from one to four (8) .
Because the NWI-R was translated and adapted for the Brazilian population (9) and the subscales were validated (B-NWI-R) (6) , they were used in the present study.
Grid & Leadership in Nursing: ideal and actual behavior (7) consists of two instruments based on the grid theory to assess the ideal leadership behavior of nurses and the actual views of staff members in this regard. One instrument is thus responded to by the staff leader, and the other by a staff member. As both instruments are appropriate for the Brazilian setting and were subjected to apparent and content validation (7) , they are used in the present study.
The instrument consists of 25 statements with four answer options, attributed different scores and categorized as follows: entirely desirable (four points), desirable (three points), undesirable (two points), and entirely undesirable (one point) (7) .
Each statement concerns one style of leadership (1.1; 1.9; 5.5; 9.1; and 9.9) (7) . The leadership style attributed the highest score represents how a nurse exerts leadership as a function of his or her conception of ideal behavior and the view of a member of his or her staff regarding the actual situation (7) .  Table 1 shows the intersection between the average scores on the B-NWI-R and the actual leadership styles of nurses independent of the particular ICUs.
The data presented above show that there was no difference between the work environment and actual nursing leadership (p = 0.852). Styles 9.9 and 9.1 were associated with the lowest average scores on the B-NWI-R. In comparison, the means for styles 1.1 and 1.9 tended to be higher but not significantly.

Actual leadership style was not significantly related
to the various B-NWI-R domains, as shown in Table 2.   were lower at the public hospital ICUs. The scores on the B-NWI-R relative to leadership style 1.9 were lower at the private hospital ICUs. Thus, the most favorable (6) correlation between the leadership style in which the professional is most involved (9.1) (7) and the mutual respect between nurses and doctors corresponded to the public hospital ICUs.

Discussion
The interest in studies assessing the impact of the work environment on the outcomes of care and management arose in the 1980s, when the American Academy of Nursing conducted a nationwide study to establish which characteristics of the work environment favored nursing professional practice. The institutions known as magnet hospitals shared the following features: management, professional practice, and professional development (8) .
Those attributes are assessed by the NWI-R (8) , which is the basis for B-NWI-R (6) and which is anchored in the Nursing Professional Practice Model (10) . The latter is defined as a system (structure, process, and values) that lends support to nurses to control the care provided to patients and the setting in which care is provided.
On those grounds, the results of the present study show that the work environment at the investigated ICUs is favorable for daily nursing practice (mean = 1.95; 95% CI: 1.85-2.05), regardless of the hospital being analyzed.
One might thus infer that the ICU work environment is attractive for nursing professionals. This claim is corroborated by a study conducted in South Korea (11)  The average score of 17 rural ICUs in the state of São Paulo (SP), Brazil (12) , on the B-NWI-R was 2.13, which was higher than the score found in the present study. Thus, the work environment at the four investigated ICUs was found to be healthy.
However, no significant association was found between the average score on the B-NWI-R and the actual nursing leadership styles (p = 0.852). The same held true for the scores on the B-NWI-R subscales.
Thus, the work environment was not found to influence the nursing leadership styles in the present study. However, only four of the B-NWI-R domains were validated for use in Brazil (6) . Once validated, the remaining domains might allow for a broader assessment of the work environment characteristics than the domains available at the present time.
When the ICUs at the public and private hospitals were analyzed separately, no significant difference was found in relationship between the average B-NWI-R score and the nurses' actual leadership styles (p = 0.437). Only the nurse-physician relationship domain exhibited p = 0.001. It should be noted that the ICUs at the public hospitals exhibited the most favorable correlation between the leadership styles in which the professional is most involved, 9.1 (7) , and the nurse- Leadership has been associated with several variables, including patient satisfaction. There is strong evidence (12) of an association between positive leadership behaviors and increased patient satisfaction and reduced adverse events (12) .
In one study (13) , 29% of job satisfaction, 22% of organizational commitment, and 9% of productivity were explained by the use of leadership behavior.
Those findings are in accord with the results of three studies (14)(15)(16) in which the variables of interest were job satisfaction, organizational climate (14)(15) , and burnout (16) .
Based on those findings and the fact that the work environment might attract and retain excellent professionals, according to the Nursing Professional Practice Model (10) , it was hypothesized that such strong characteristics might influence nursing leadership styles,