Application of Revised Nursing Work Index and the Maslach Burnout Inventory for Registered Nurses from Public Health Institutions, Sao Paulo, Brazil

Nogueira, Lilia de Souza

Purpose: The aim of this research was to verify the relationship between the results of the Revised Nursing Work Index (NWI -R) and the Maslach Burnout Inventory (MBI).

Methods: The Revised Nursing Work Index (NWI-R) (1) is a tool that measures the practice environment. It consists of 57 items, in which 15 of the 57 items were divided in 4 subscales: autonomy, control over the practice setting, doctor nurse relationship and organizational support (1).

The autonomy subscale consists of five items and the control over the practice setting subscale comprises of seven items. The doctor nurse relationship subscale is formed by three items. The organizational support subscale consists of ten items derived from the subscales cited above (1).

The scale used in the NWI-R is the Likert and can vary from one to four. The lower the score, the greater the favorable attributes of the nursing practice. The scores for the subscales are obtained by averaging the scores of the subjects' responses. They can vary from one to four points (1).

The MBI aims to measure the physical and emotional exhaustion of workers by assessing their feelings towards work. It has 22 items divided into three domains: emotional exhaustion reduced personal accomplishment and depersonalization. The sum of the responses for each topic determines the variation of each domain: nine to forty-five points for emotional exhaustion, eight to forty points for personal fulfillment and five to twenty-five points for depersonalization and (5). The scores for the emotional exhaustion and depersonalization items are considered as follow: the higher the score, the greater the emotional exhaustion feeling and depersonalization perceived by nurses. Regarding the decrease in the personal fulfillment item, higher scores depict high personal achievement Professional (2.5).

In this study five categories of response (never, rarely, sometimes, often, always) were used. They differ slightly from the seven original categories. The choice of the five categories was due to the fact that Brazilian professionals had difficulties to discriminate between the seven original categories of response (8).

This study was conducted in hospitals and public clinics under direct supervision of the Coordinator of Health Services of the State Secretariat of São Paulo, Brazil Health, and it is part of a larger study whose objective was to know the nursing records and variables associated with them.

Among the 43 eligible health institutions invited, three institutions did not take part in the study. Thus, 40 health institutions, whose collection took place between January 2011 and January 2012, were included in the study.

The questionnaires were completed by nurses who worked or were responsible for sectors where there were patients. The project was submitted to the Ethics Committee of the School of Nursing for evaluation of the ethical aspects. Participants who provided information by self-report were informed about the study and its procedures and, if agreed, signed the Informed Consent Form.

The Pearson correlation coefficient was used to verify the correlation between variables. The correlation strength used was that proposed by Levin, Fox (2004 ) in which perfect correlation was attributed for those that reached 1.0 , strong correlation for those between 0.9 and 0.6 , moderate correlation for those between 0.59 and 0.3, and weak correlation for those below 0.3.

Analysis of clusters was used to identify groups of similar institutions. The method of K-means was employed and it was based on the Euclidean distance (6-7). The results of the MBI and these groups of institutions were compared by Analysis of Variance (ANOVA). In the case of significant difference between groups, the identification of these groups was done by multiple comparisons using the Bonferroni test. In all analyzes, the significance level of 5 % was established.

Results: The NWI profile represented 40 institutions. The MBI was completed by 745 nurses with an average age of 43 years (SD = 9.6), average training time of 14 years (SD = 8.5) and on average 9 years of working in the participating institution (SD = 9.4). The study was comprised of 23 general institutions, 11 institutions and 6 psychiatric outpatient services.

The NWI-R showed reliability of 0,962 for the total items, 0,695 for autonomy items, 0.794 for control over the practice setting, 0.773 for doctor nurse relationship and 0,815 for organizational support. Regarding the MBI, the reliability was 0.602 for the total items, 0,877 for emotional exhaustion, 0.549 for the depersonalization and 0.723 for personal fulfillment.

The mean NWI-R was 2.27 in the total sample, standard deviation (SD) 1.03, median two, minimum and maximum values of 1 and 4, respectively. As for the MBI, the mean and SD of domains found in this population were: emotional exhaustion 22.37 (± 6.47), depersonalization 8.98 (± 3.04) and decreased job satisfaction 31.74 (± 3.98).

Table 1. NWI-R (total and subscales) and MBI (total and domains) correlation. São Paulo, Brazil - 2010-2011

	NWI-R
MBI	Total Score	Autonomy	Control over the Practice Setting	Doctor Nurse Relationship	Organizational Support
Total Score	0,26	0,25	0,24	0,19	0,21
Total Score	p=0,107	p=0,123	p=0,140	p=0,250	p=0,202
Emotional Exhaustion	0,46	0,46	0,39	0,24	0,40
Emotional Exhaustion	p=0,003	p=0,003	p=0,013	p=0,134	p=0,010
Depersonalization	0,28	0,33	0,19	0,13	0,27
Depersonalization	p=0,081	p=0,039	p=0,229	p=0,439	p=0,087
Decreased Job Satisfaction	-0,41	-0,44	-0,26	-0,11	-0,40
Decreased Job Satisfaction	p=0,008	p=0,004	p=0,104	p=0,516	p=0,011

Pearson correlation coefficient

Table 1 shows that the correlation between emotional exhaustion and NWI-R total score, autonomy, control over the practice setting and organizational support were moderate and positive. The same results were found between the depersonalization and autonomy variables. Moderate negative correlation was present in the analysis between decreased job satisfaction and the total score of NWI-R, autonomy and organizational support. The results of the analysis of the remaining correlations NWI-R and MBI were weak, both among total scores, and between the domains of MBI subscales and the NWI-R.

Table 2. Health institution distribution according to NWI-R homogeneous groups. São Paulo, Brazil - 2010-2011

Groups	N	%
1	13	32,5
2	16	40,0
3	3	7,5
4	8	20,0

In cluster analysis, four groups of hospitals with similar profile were identified from the four subscales of the NWI -R. In characterizing these groups, we found that group 4 had the lowest scores on all subscales of the NWI -R. Group 3 received the highest scores on three subscales of the instrument (autonomy, control over the practice setting and organizational support). The highest values of subscale doctor nurse relationship were identified in group 1 (Table 2).

There was no difference between groups in relation to the total MBI and decrease the area of personal fulfillment. Difference between groups was observed for emotional exhaustion ad depersonalization (p< 0.05) domains.

Considering multiple comparisons by the Bonferroni test, it was found that there were differences only between groups 3 and 4 in relation to emotional exhaustion, and groups two and three, on the depersonalization domain. Group 4, as the NWI -R, gathered the institutions with the most unfavorable for the nursing practice, which might have contributed to the physical and emotional exhaustion of nurses in these institutions.

Conclusion: The group of public hospitals in São Paulo, Brazil with the best nursing practice showed significant statistical difference for the emotional exhaustion and depersonalization domains compared to groups of hospitals with less favorable characteristics.