Learning from mistakes: analyzing incidents in a neonatal care unit *

ABSTRACT Objective: to analyze incidents reported in a neonatal care unit. Method: a quantitative, cross-sectional and retrospective study with a sample of 34 newborns. Data were collected through a structured form, composed of two parts: sociodemographic/clinical characteristics of the newborns, and characteristics of the reported incidents. Data were collected from the institution’s computer system, in a period corresponding to 13 months, and analyzed by means of descriptive statistics. Results: the majority of the newborns were preterm (70.6%), male (52.9%) and born through caesarean section (76.5%). During the study period, 54 incidents were reported, totaling a frequency of 1.6 incident per newborn. It was found that 61.1% of incidents were related to medicines, 14.8% to accidental loss of tracheal tube and 9.3% to catheter obstruction. Conclusion: analysis of the reported incidents has shown that most incidents refer to the drug process. Information about the incidents can increase the perception of health professionals regarding the impact of their actions.

authors report that "in neonatal intensive care units a single patient, sometimes an extreme premature, is manipulated by several professionals, which predisposes to an increased chance of suffering the consequences of an error" (6) .
With regard to neonatal care, some studies have been published with the purpose of measuring the occurrence of adverse events in this type of service.
A study published in Argentina, with the objective of describing the epidemiology of adverse events in a neonatal population of Buenos Aires, found a relative frequency of clinical histories with the presence of at least one adverse event of 16.9%, being the occurrence of adverse event associated with hospitalization in the Intensive Care Unit, prolonged hospitalization, lower gestational age and lower birth weight (5) .
In the United States of America, in a study conducted in a Neonatal Intensive Care Unit, the authors described a 74% of incident rate in hospitalized newborns, with the most frequent occurrences the infections associated with health care, accidental extubations, intravenous catheter infiltrations, skin rupture and intraventricular hemorrhage (7) .
Another study, conducted in Brazil, reported that 183 (84%) of the 218 newborns included in the investigation had suffered some type of adverse event.
Of the 579 identified adverse events, with a rate of 2.6 adverse events per patient, 29% were thermoregulation disorders, 17.1% were glycemia disorders, 13.5% were hospital healthcare-related infections and, lastly, 10% referred to unscheduled extubation (8) .
The incidence of errors and adverse events instigates health organizations worldwide to promote a culture based on the best and safer care, seeking to reach professionals of the various levels of care. It is essential to understand that safety practices need to be adapted to different populations, and economic, social and cultural contexts. There is a scarcity of investigations into the occurrence of incidents in the neonatal care unit, and a greater knowledge about this theme is necessary for the quality of care (6) .
In daily practice, it can be seen that the occurrence of incidents during the care process directly reflects the patient's safety indicators, the quality of care and motivation of the professionals involved, although these facts are still devalued by many managers in health institutions. In view of the above, and the imminent importance of conducting studies that clarify this issue, the following question emerged: "What are the incidents that occur in a neonatal care unit of a private hospital in southern Brazil?".
The objective of the present study was to analyze incidents reported in a neonatal care unit. This enabled knowing the incidents that occur in a neonatal care This study were included the newborns who had a length of stay of more than 24 hours and who had at least one reported incident during their hospitalization. Newborns with of incomplete reports of incidents were excluded.
Thus, the population was constituted by 340 newborns and the sample by 34 newborns.
The form, developed by the researchers, was composed of two parts, namely: Part 1) data referring to the sociodemographic characteristics of the newborns, such as sex, age at birth (gestational age), age of hospitalization and age of discharge, type of delivery, birth weight and weight at discharge, Apgar score (in the first and fifth minutes of life), length of stay in the unit, date of birth, hospitalization date and discharge date, reason for hospitalization, origin and outcome of hospitalization; Part 2) data corresponding to incident reports, namely type of incident (care to which it relates), its classification, severity, avoidability, description of the incident and actions taken after the incident. In the first fortnight of data collection, a pilot test of the tool developed by the researchers was carried out and there was no need to change the instrument, so the data collected were used in the study.
The institution's information technology team was asked to provide a list of newborns who had been hospitalized during the study period and who had at least one incident through recorded electronically.
After providing this list, the data on the newborn and on incident reports were consulted in the institution's computer system, allowing the form to be completed.

Results
The present study found that newborns who had suffered at least one incident during admission to the neonatal care unit represent 10% (n = 34) of the study population. Table 1   The incidents were also analyzed on the type of associated care (Table 2).   26.6% very low birth weight (<1500 grams); also, 42% of these newborns received an Apgar score lower than seven in the first minute (9) . These same characteristics  (15) . In a previously cited study, the authors analyzed 749 records using a record review procedure and found a total of 554 adverse events, which represents a rate of 0.74 events per records analyzed (7) . In another publication, researchers examined incidents reported voluntarily over a one-year period in eight neonatal care units and one pediatric unit of Dutch institutions and found 5225 incidents, of which 4846 were considered eligible for analysis, in 3859 hospitalizations, totaling 1.25 incidents due to hospitalization (16) . In another study, conducted in Brazil, 183 (84%) of the 218 newborns included in the investigation reported some type of adverse event.
A total of 579 adverse events were identified, resulting in a rate of 3.16 adverse events per newborn (8) . This study presents higher results than those found in this investigation.
In relation to publications on neonatal incidents, there is a scarce number of studies on this subject, which highlights the need for further investigations with this approach and a deeper understanding of the characteristics of these incidents. There is a predominance of studies related to the occurrence of adverse events targeting the adult population, but a shortage of data focused on the pediatric population, especially newborns. In a recent study, the researchers

Discussion
Research on the epidemiological profile of hospitalizations in neonatal units also found data related to the characteristics of newborns. In a Brazilian study, the authors described that 70% of the newborns had been born through cesarean delivery, 70% were premature (gestational age <37 weeks), 41.6% of the Rev. Latino-Am. Enfermagem 2019;27:e3121.
reported that there is a gap in investigations on the occurrence of incidents, especially adverse events, in neonatal intensive care units (6) .
According to the results presented in a study conducted in Argentina, 65% of the adverse events found in the clinical histories of the newborns produced transient sequels without risk of death; however, half of the deaths that occurred were considered very likely to be preventable. Regarding the category, the incidents evidenced in 50% of the cases were related to the errors occurred during the monitoring of the clinical state or with the nursing care required by the neonates during hospitalization, for example the handling of catheters, accidental extubations, retinopathy of the preterm newborn, hemorrhages, transfusions, among others (5) .
In a study that investigated incidents involving with 55% of incidents classified as human error (17) . In a current study that prospectively analyzed intubations in a neonatal care unit, the authors found during the investigation period 273 intubations with available data, of which 107 were intubations with adverse events.
The increase in the number of intubation attempts and emergent intubations were predictors of adverse events (18) .
Most of the incidents analyzed in the present study were related to the medication, totaling 61.1% of the notifications. In a study recently published, the researchers found in a neonatal unit 511 reports of adverse drug-related events over a seven-year period, resulting in an incidence of 32.2 drug-related adverse events per 1000 days, with 39.5% of prescription errors, 68.1% of administration errors and 0.6% were adverse drug reaction (19) . 17.5% to employees who had been involved in the error, and only 8.7% cases the physician was informed (20) .
A reliable way of knowing the factors that cause the errors and that reduce the quality and safety of the care provided is through a detailed analysis of the incidents that occurred. A more in-depth knowledge of the incidence and characteristics of incidents, as well as the continuous monitoring of the occurrence of these errors, could help improve the quality of health care for the neonatal population (12) . In addition, actions are needed to prevent incidents, which include the continuous training of all professionals and the development of practices directed to the whole system, including the technical and organizational environment (17) .
The present study chose to perform the search for incidents retrospectively and through electronic, anonymous and voluntary notifications in the computer system of the chosen institution. In a recent Brazilian publication, the researchers also chose to use information directly extracted from the databases of the studied institutions in order to avoid errors resulting from manual transcripts of information (21) .

Although voluntary reporting of incidents is not
considered the most effective way to detect adverse events, it is still the mechanism used by most health institutions. This is due to the fact that this tool is easily available to professionals, is a source of information that sometimes provides detailed descriptions of the facts, and assists in the review of processes. Underreporting of incidents is one of the reasons that prevents its effective voluntary use as a research tool on patient safety. and that corroborate the other few published studies.
In addition, knowing the reported incidents provided professionals and managers with support in choosing priority areas and actions for the development of improvements, since they could reflect on the mistakes most commonly made and valued by the teams.
Voluntary notifications were the only source of incident identification, being a limitation of this study.
Thus, it may have restricted the amount of information about them and reduced the scope of the investigation.

Conclusion
It was evidenced that 10% of newborns admitted to the unit had undergone at least one incident during the investigation period, which points to the existence It is believed that the number of incidents that occurred in the neonatal unit may be greater than that reported, taking into account that there are errors that were not perceived by the professionals or were not recorded in the institution's notification system. Thus, a combined approach of incident detection methods is considered to be the most complete and effective since these methodologies, when used alone, have some shortcomings. In order to achieve greater numbers of voluntary notifications, it is necessary to develop an effective safety culture, in which not only the institutional administration but also the care professionals are aware of their role in the development of harm reduction.