Patient safety culture at neonatal intensive care units: perspectives of the nursing and medical team 1

OBJECTIVE: to verify the assessment of the patient safety culture according to the function and length of experience of the nursing and medical teams at Neonatal Intensive Care Units. METHOD: quantitative survey undertaken at four Neonatal Intensive Care Units in Florianópolis, Brazil. The sample totaled 141 subjects. The data were collected between February and April 2013 through the application of the Hospital Survey on Patient Safety Culture. For analysis, the Kruskal-Wallis and Chi-Square tests and Cronbach's Alpha coefficient were used. Approval for the research project was obtained from the Ethics Committee, CAAE: 05274612.7.0000.0121. RESULTS: differences in the number of positive answers to the Hospital Survey on Patient Safety Culture, the safety grade and the number of reported events were found according to the professional characteristics. A significant association was found between a shorter Length of work at the hospital and Length of work at the unit and a larger number of positive answers; longer length of experience in the profession represented higher grades and less reported events. The physicians and nursing technicians assessed the patient safety culture more positively. Cronbach's alpha demonstrated the reliability of the instrument. CONCLUSION: the differences found reveal a possible relation between the assessment of the safety culture and the subjects' professional characteristics at the Neonatal Intensive Care Units.


Introduction
Patient safety is related to factors like human fallibility, weaknesses in health organizations, problems with technological devices, communication and inappropriate dimensioning between teams and professionals, as well as with the task burden and limited knowledge about safety (1) .
One of the safety pillars is the institutional culture, which is based on good communication, trust, organizational learning, commitment of the hospital management to safety, leadership, non-punitive approach of errors and a shared perception of the importance of this theme (2)(3) . Therefore, the health institutions need to promote a culture based on these values to improve the patient safety. thus avoiding individual accountability processes (4) .
The literature shows that the assessment of the patient safety culture differs among the hospital organizations depending on the size, accreditation, specialty, and can vary among the professionals, depending on the function, length of experience and age (5)(6)(7)(8)(9) .
In order to promote a culture that sets patient safety as a priority, however, first, it is fundamental to assess the patient safety culture present in the health organizations, in order to plan specific intervention in search of better results (3) .
Identifying the factors that further the patient safety culture and the barriers that interfere in it permits the development of strategies according to the particularities of each place of work, obtaining more satisfactory results (10)(11) .
Therefore, as the discussion of the patient safety culture is recent in Brazil, the scientific knowledge on this theme is incipient, mainly in specialized services authors were used, Translation Guidelines for the AHRQ Surveys on Patient Safety Culture (12) . Although this study was not aimed at validating the instrument, in order to avoid any noise due to the translation, the coherence of the produced translation was verified by consulting the translation to Portuguese produced by other authors (13)(14) , showing that the questions had the same meanings.
The data collection started when the research project was presented to the professional. Next, they received the Informed Consent Form (ICF) and an envelope with the data collection instrument, which the participants completed and placed in a box.
For the descriptive analyses, the answers were recoded as positive: I completely agree, I agree, always and frequently; negative: I completely disagree, I disagree, never and rarely; neutral: I neither agree nor disagree (3) . For inferential statistical analysis, the software R 3.0.1 (15)

Results
In this study, the statistical difference was verified years. Most of the subjects, that is, 93(66%) had more than 10 years of experience in the profession (Table 1).   Table 2, it was observed that, when the subjects had less than one year of experience at the hospital, the percentage of positive answers tended to be higher (59%) than for participants with more than one year of experience, who obtained an average 40% of positive answers.
Also according to Table 2  In addition, the results also showed that some   (Table 5).

Discussion
The patient safety culture at health services is represented by the set of knowledge, thoughts, beliefs, habits, customs and routines followed and shared among the team members (19) . Thus, many individuals can develop good or bad habits according to the safety climate in the environment they are inserted in (2) . In view of this aspect, it was considered that the length of professional experience, as well as the length of experience at the institution or unit, can provide distinct perceptions of the safety culture.
Hence, in view of the results, it was observed that, as the professionals gained experience, at the institution as well as the NICU, they chose less positive answers with regard to the safety culture. In contrast, a study involving 136 professionals assessed the relation between the answers obtained on the HSOPSC and the professionals' age, detecting that the youngest nurses (between 23 and 43 years) chose less positive answers in terms of safety when compared to the oldest ones (between 44 and 65 years) (6) .
Another professional characteristic that influences the number of positive answers was the professional function, where the physicians and nursing technicians showed a higher percentage of positive answers than nurses and auxiliary nurses at the NICU. In that sense, a study of 1113 professionals from Spanish hospitals also showed that the function was a factor associated with the assessment of the safety culture, but the nurses obtained a more positive perception (20) . In addition, according to another study undertaken in different countries, while the surgeon believed that teamwork in the surgery room was strong and, hence, assessed the culture more positively, the other members disagreed (21) .
In addition, individuals from the same professional category may perceive the safety culture differently.
According to a study of 550 nurses from the United States and Canada, despite belonging to the same profession, the professional position inside the organization influenced the perceived safety. The nurses in management or administrative functions assessed the safety culture more positively than the clinical nurses (8) .
As regards the patient safety grade, a difference In a study at Spanish hospitals, on the other hand, the nurses were more positive with regard to the grade than the physicians (20) .  (23) .
The low compliance with error reporting may also be related to the conducts aimed at the professionals, mainly to the punitive error approach, as shown in a study of 70 Intensive Care nurses, where 52(74%) reported that punishment happens in most cases when an error is reported (24) .
Besides these results, as the data were collected before the validation of the HSOPSC for the Portuguese language (14) , the reliability of the internal consistency Low Alpha coefficients were also found in the validation of the Turkish (n=309), Spanish (n=174), Dutch (n=583) and Japanese (n=6395) (25) versions of the instrument though. These studies highlight that these results may be influential due to the sample size as, the larger the sample, the greater the chances of repetition in the analysis of the Alpha will be and, finally, the higher the Alpha coefficient may be (18) . Therefore, the HSOPSC should be used in other studies in Brazil, as it is only by using the instrument in different samples that its validity and reliability can be confirmed (25) .

Conclusion
The Through error reporting and the non-punitive culture, the causal factors can be identified and barriers can be implemented that reduce the risk situations.
The size of the population and sample to assess the reliability of the instrument was considered a limitation in this study. Nevertheless, it is considered that the HSOPSC permits analyzing the safety culture at institutions and identifying its characteristics.
Therefore, this instrument can be used in the Brazilian health context, through the replication of research in NICUs, with a view to exploring the patient safety culture in these environments in further depth, as well as through its application in other hospital populations.