In the past 25 years, following the publication of the study To err is human: Building a safer health system,1 national and international health authorities have made considerable efforts to improve patient safety.

There have been advances in safety in the field of paediatrics, too,2–4 but the safety of neonatal patients has not been adequately studied, despite the high incidence of adverse events (AEs) in this group.5,6 Neonatal units, and especially those devoted to intensive care, are risk areas for the development of AEs on account of their complexity and the characteristics of their patients.5–8 Thus, it is necessary to introduce and implement tools and strategies that allow the detection of incidents and that protect against and help reduce AEs. Voluntary AE reporting systems underestimate the prevalence of these incidents, so the use of other tools for the active search of AEs, such as safety checklists (SCLs) can complement reporting and improve detection, which allows us to learn more from our mistakes. The WHO has promoted the use of SCLs both in surgery9 and childbirth.10 Within the framework of a plan for improving patient safety in the department of neonatology of our hospital, we decided to assess the potential usefulness of SCLs for the detection and correction of incidents and the prevention of AEs.

We conducted a prospective, quasiexperimental study in the Department of Neonatology of the Hospital General Universitario Gregorio Marañón between March and September of 2015. The study was approved by the Ethics and Research Committee of the hospital, and received no funding.

The neonatal department offers care at the IIIC level and serves patients with any type of neonatal disease. It has two separate inpatient areas—an intermediate care unit with 34 beds, and a neonatal intensive care unit (NICU) with 16 beds—and a staff comprising 91 nurses, 55 nursing assistants, 19 physicians and a department chief, in addition to staff in training. The department has electronic health records and prescription systems, although physicians occasionally make adjustments to treatments by means of hand-written prescriptions.

Two SCLs were created, one for each unit (Appendices A and B). They were developed by a working group consisting of neonatologists and preventive medicine physicians based on the AEs and incidents described in the literature and the voluntary reports submitted the previous year, in addition to their own personal experience.

We collected data for variables pertaining to:

• -

  Patient identification: due to the serious consequences that errors in identification may have. We considered that a patient was identified correctly if the patient was wearing an identification band whose information was the same as the information in the crib, the treatment sheet and the nursing chart.

• -

  Medication: because medication errors are among the most frequent errors described in the literature. Checking the medication prescription sheet (drug, dose, route of administration, interval between doses, formulation) and the correct transcription to the nursing chart.

• -

  Respiratory support: due to the significant consequences that errors in this area may have. Type of support, checking of respiratory parameters, humidifier, temperature, secure attachment of endotracheal tube, and integrity of nasal mucosa and skin.

• -

  Checking that the settings for the alarms in monitoring devices and other equipment are correct: each patient has a different problem and requires personalised care, and therefore the thresholds for alarms, both in vital signs monitoring and other devices, must be set individually. For instance, it is essential to monitor oxygen saturation in preterm newborns, in whom hyperoxygenation must be prevented, or to monitor cardiac patients for the presence of pulmonary hypertension or low cardiac output. Similarly, adequate settings in drug infusion pumps can help detect dosing errors early on, and adequate pressure threshold presets in pumps allow the detection of problems such as thrombosis or extravasation.

• -

  Nutrition: due to the severe consequences that errors in nutrition may have on occasion. Type of nutrition, amount and rate of infusion, route of administration, verification of correct functioning and adequate placement and secure attachment of feeding tube.

• -

  Vascular access lines: with the purpose of reducing the risk of nosocomial infection and catheter-related complications. Verification of correct identification (following the protocol of our hospital), of continued need for access line, integrity of the skin at the site of insertion and along the catheter trajectory, catheter functioning and tissue perfusion. Infusion of medication using the line specified in the prescription.

• -

  Urinary catheter: with the purpose of reducing the incidence of urinary tract infections. Indication of sustained catheter use, verification of closed system.

• -

  Watching for ulcer development.

The entire staff participated in information sessions in the month prior to the introduction of the SCLs.

The SCLs were filled out in cooperation by the medical and nursing staff in charge of the patient. In the NICU, SCLs were completed 3 times a day (once per shift), as there are frequent changes in treatment and supportive care in its patients. In the intermediate care unit, it was completed once a day during the morning shift, as changes are less frequent throughout the day compared to the NICU.

The use of SCLs is widespread in several health care fields, such as surgery, in which the WHO has promoted the use of SCLs in the operating theatre as part of the Safe Surgery Saves Lives initiative,9 which has been associated with a reduction in surgery-related mortality and complications.12 Furthermore, the WHO has developed a safe childbirth checklist, conceived as a tool to improve maternal care during birth.10

However, SCLs are a poorly studied tool in the field of neonatology. It is estimated that there are between 30 and 74 adverse events per 100 patients hospitalised in neonatal units.13,14 It is of vital importance that we investigate tools that allow us to detect these incidents more thoroughly in order to improve the safety of neonatal patients.

To our knowledge, this is the first published study on the usefulness of SCLs for the detection and prevention of incidents that may impact patient safety in a neonatology setting.

According to our results, SCLs were particularly useful in the NICU, where its use allowed the detection of 1 incident per 5.3 completed SCLs, compared to the intermediate care unit, where detecting a single incident required performance of 29.3 SCLs.

In the intermediate care unit, during the 4 months when the SCL was being used, adherence was low and the interim analysis at 4 months demonstrated that only 28 incidents had been detected, and none of them were serious. The incidents detected most frequently involved medication, which is consistent with the literature published to date.6,7,15,16

In the NICU, however, there was a higher degree of adherence to SCLs (56.5%). Once again, the incidents detected most frequently were those associated with medication (44%). Within this type, the most frequent incidents were prescription errors, which amounted to 79%. This finding is particularly relevant, as it demonstrates that despite all the measures implemented in our hospital to reduce these errors (electronic prescription, protocols, automated infusion pumps), this continues to be a problem deserving further investigation and the implementation of new preventive tools and strategies. In a recent study conducted in Spain,7 prescription errors amounted to 39.5% of all detected medication errors, and administration errors were the most frequent type of medication error (68.1%). There are barriers to comparing the findings of this study with our own, as voluntary incidence reports were the source of data in the former, while we used SCLs to obtain data for our study. Other published studies that used data from voluntary reporting have also found a greater proportion of administration errors.15,17

The detected incidents were analysed, and corrective measures were implemented to prevent or reduce the likelihood of their recurrence.

The higher yield of SCLs in the NICU compared to the intermediate care unit was reinforced by the consequences of the detected incidents. While in the intermediate care unit 100% of the consequences were rated as potential harm that did not ultimately occur, 26.1% of incidents in the NICU resulted in temporary or minor harm. Furthermore, only 0.1 incidents per admitted patient were detected in intermediate care, compared to 4.3 in the NICU. Given the low yield of the SCL in the intermediate care unit, its use was discontinued 4 months after its introduction. A new SCL is currently being developed to better fit the needs of this unit.

In our opinion, the greater usefulness and adherence to SCLs in the NICU may be due to several factors. On one hand, to the greater complexity and instability of NICU patients, who also require more complex treatments and techniques. On the other, these patients frequently require treatment changes. Furthermore, the higher nurse-to-patient and doctor-to-patient ratios increases the availability of staff for conducting SCLs. Establishing the causes of the differences between the units would require a more detailed analysis and a different approach to the one adopted in this study, and may be the subject of future studies.

Another of our objectives following the introduction of the SCLs was to make a survey of the attitudes of the NICU staff regarding their use. We ought to highlight that 68.75% of the staff that responded to the anonymous satisfaction survey considered that the use of SCLs had succeeded in preventing AEs, and 87.5% felt it did not add significantly to their workload. Although overall satisfaction was lower—as only 54.16% felt satisfied or very satisfied—100% responded that if a family member were to be admitted to the unit, they would want SCLs to be used.

One salient finding was the difference in the overall satisfaction reported based on the years of work experience, as the proportion of satisfied or very satisfied respondents was nearly double in the staff with fewer than 5 years’ experience compared to more experienced staff (83.3% vs. 44.4%). Less experienced staff perceived SCLs as more useful, which means that these lists can be a useful tool for improving patient safety in units with high staff turnover rates.

In our experience, SCLs have proven useful in the detection and prevention of incidents in the NICU setting. Furthermore, their use has generated a specific culture and systematic approach to everyday's work, establishing the habit of checking on every patient's treatment and monitoring. On the other hand, their use forces health care staff to regularly reassess the need to continue using peripheral or central access lines and urinary catheters, removing them once they become unnecessary, which may contribute to reducing the risk of nosocomial infection. Their use promotes a patient safety culture, which is particularly important for staff in training or newly hired.

We are aware of the limitations of our study, for while we demonstrated the usefulness of SCLs for the purpose of detecting incidents in our unit, it would be important to assess whether this had a direct impact on our patients, with a reduction in the incidence of health care-associated complications. This would be difficult to establish, as many factors influence the development of these complications, and variations in the latter could not be solely attributed to the use of SCLs. It is important that we continue refining this safety tool to achieve greater adherence and increase the satisfaction of the health care staff.

In conclusion, SCLs have proven to be a useful tool for the detection of incidents, especially in the NICU, that increases our knowledge on patient safety and is perceived positively by the staff of the unit.

The authors have no conflict of interests to declare.