Article Text

Download PDFPDF

Critical incident reporting in neonatal practice
  1. Nimish V Subhedar,
  2. Heather A Parry
  1. Neonatal Intensive Care Unit, Liverpool Women's NHS Foundation Trust, Liverpool, UK
  1. Correspondence to Dr N V Subhedar, Neonatal Intensive, Care Unit, Liverpool Women's Hospital, Crown Street, Liverpool L8 7SS, UK; nim.subhedar{at}


Patient safety incidents (PSIs) occur relatively frequently in healthcare. Incident-reporting systems are designed to systematically collect information relating to PSIs in order to identify and rectify problems in the delivery of clinical care. This review provides an overview of the incident reporting process and summarises local and national data of PSIs reported in a neonatal population.

Statistics from


It is estimated that an adverse event or ‘incident’ occurs in one in 10 of all NHS admissions, accounting for approximately 850 000 adverse events annually.1 Neonates are a unique group of patients who are particularly vulnerable by virtue of their immaturity and need for prolonged and/or complex intensive care support.2 3 Although relatively new in the field of healthcare, incident-reporting systems have been successfully used in other high-risk industries, particularly aviation. In this article we review the basic components of successful incident-reporting systems and provide examples of how local reporting might help to improve clinical practice.

What is an ‘incident’?

The terminology preferred currently is ‘patient safety incident’ (PSI), defined as ‘any unintended or unexpected incident which could have or did lead to harm.’4 Other terms commonly used to describe such episodes include ‘critical incident,’ ‘clinical incident,’ ‘adverse clinical event’ and ‘iatrogenic event.’ PSIs comprise actual harmful events as well as ‘near-miss’ events, where a problem is intercepted before any harmful effect occurs. A medication error where a baby becomes unwell because the wrong dose of drug has been administered would be considered a harmful event, whereas a prescription error recognised and intercepted before the wrong dose could be administered would be a near-miss; however, both are PSIs.

Incident-reporting systems

Critical incident-reporting systems aim to identify problems in the delivery of healthcare and provide an opportunity for an organisation to learn from errors in order to improve patient safety. They can operate at a local level (within individual units, clinical departments or hospitals) or at a national level (eg, the National Patient Safety Agency (NPSA) National Reporting and Learning System (NRLS) in England and Wales).5 Most systems capture raw data about the nature of the incident, its location, those involved, any impact on the patient and potential contributory factors.

Internal (local) reporting systems enable healthcare staff to:

  • ▶. identify the type and frequency of incidents that occur;

  • ▶. categorise incidents according to seriousness of risk;

  • ▶. perform appropriate investigations into the underlying ‘root causes;’

  • ▶. monitor trends and identify emerging patterns of clinical problems;

  • ▶. learn from errors and devise and implement local solutions to identified problems.

External (national) reporting systems allow:

  • ▶. capture of information about rare errors;

  • ▶. identification of new emerging problems on a national scale;

  • ▶. healthcare staff to report anonymously and confidentially to a central, independent body;

  • ▶. organisations to learn from each others' errors;

  • ▶. development and dissemination of solutions across healthcare systems.

Incident reporting forms and trigger lists

Most hospital-based incident-reporting systems have traditionally adopted a voluntary, non-punitive approach. Anonymous reporting allows individuals to notify the organisation of an incident confidentially. While this may promote higher levels of reporting, it leads to difficulties in obtaining further detailed information which is essential before an in-depth review can be carried out (see below). Staff are encouraged to report incidents using paper forms (or more recently, online using a web-based proforma6) guided by an incident ‘trigger list’—a list of key events prompting staff to complete an incident form, triggering further investigation.7

Trigger lists are developed locally to suit local requirements but typically include specific adverse clinical events (eg, unexpected neonatal death), general areas of concern (eg, equipment failure), organisational issues that have the potential to impact on patient care (eg, staffing problem), clinical complications that may be related to inappropriate clinical management (eg, pneumothorax) and markers of poor quality of care (eg, admission temperature <36°C). One difficulty in defining a list of incident triggers is distinguishing an event that might be the result of a preventable error from one which is a recognised clinical complication: whereas a baby receiving the wrong drug is clearly an avoidable problem, necrotising enterocolitis in a preterm baby may be unavoidable despite provision of the highest standards of clinical care. In our view, triggers should focus on identification of discrete adverse events that are potentially preventable through better clinical practice rather than clinical diagnoses which may have a multifactorial aetiology. It is essential for staff to be able to submit information about incidents quickly and easily; forms which are inaccessible or unduly complex and time-consuming to complete are likely to act as a barrier to effective incident reporting. Detailed information gathering, analysis and review can be undertaken subsequently and should not impede the reporting process.

Risk scoring

Grading the risk associated with an incident is useful for two reasons. First, it allows prioritisation so that the most serious incidents are investigated, reviewed and learning points addressed with the greatest urgency. Second, it introduces a degree of consistency with semiobjective assessment of the level of seriousness of incidents which is useful in evaluating trends in incident reporting over time. In the UK, the most commonly used tool for risk scoring is a matrix which combines an assessment of the likelihood of a particular event occurring (again) with the likely consequence of such an event. A numerical and colour score is generated to quantify the level of seriousness (see figure 1).8 For example, a minor skin injury related to extravasation of an intravenous infusion would be assigned a score of 5/yellow (low risk) by multiplying the likelihood (5, almost certain to happen again) by the consequence (1, negligible). In contrast, a fatal hospital-acquired infection in an extremely preterm baby may be considered to be high risk because of a score of 20/red (likelihood score of 4 multiplied by a consequence score of 5).

Figure 1

Risk scoring system (adapted from National Patient Safety Agency8).

Within an organisation, the same risk matrix is often used for both grading risks prospectively (risk assessment) and retrospectively (incident scoring).

Review and analysis of incidents

The process of reviewing incidents is central to the critical incident-reporting system. It involves gathering further information about what happened in order to develop an understanding of why the problem occurred. Clearly, this is an essential step prior to being able to develop and implement a solution to prevent the error recurring. The procedure for reviewing incidents is likely to vary between individual units. In our practice, each incident, whether a harmful event or a near-miss, is initially reviewed by a risk coordinator or nursing shift leader in order to grade the seriousness of risk associated with it (see above). A detailed and thorough investigation of all incidents is labour-intensive and time-consuming, and therefore not practicable in the context of a busy neonatal unit. Our practice is to focus limited resources on structured review of those incidents considered to be moderate or high risk. A multiprofessional panel is convened to perform an in-depth analysis of the incident focusing on identifying ‘care delivery problems’ and associated contributory factors to allow identification of the underlying ‘root causes’ (fundamental issues).9 The next step is to formulate and implement an action plan designed to make specific changes to the way care is delivered. This action plan is then formally reviewed quarterly to ensure that the necessary changes have been made. Boxes 13 outline examples of critical incidents and changes in practice that have resulted from this formal review process. Certain types of incident (such as those with potential to cause serious harm and often categorised as high risk) are designated ‘serious untoward incidents’ (SUIs). These incidents are handled differently in that they are reported directly to Strategic Health Authorities through the Strategic Executive Information System for SUI reporting, while undertaking an urgent detailed review and root cause analysis.

Box 1 A misplaced endotracheal tube


An endotracheal tube (ETT) was inadvertently malpositioned with its tip in the right main bronchus resulting in hypoxaemia.


Detailed investigation of the incident revealed a number of contributory factors including a lack of knowledge and training about the appropriate depth of ETT insertion and deficiencies in patient identification and filing of x-rays.


A baseline audit showed that an ETT was malpositioned in 54% of postintubation x-rays. Following a review of published evidence, a new guideline was developed providing written guidance on the depth of ETT insertion, replacing the rule of thumb in general use at that time (weight (kg)+6 cm).17 Education and training sessions were held for medical staff and nurse practitioners who were given a convenient pocketsized ‘ready reckoner’ for assessing ETT length according to postmenstrual age and weight.


A reaudit demonstrated an absolute decrease of 23% in the frequency of malpositioned ETTs following the implementation of these interventions.

Box 2 A tenfold medication error


A preterm ventilated baby received an overdose of intravenous morphine which was being administered as an infusion. The error was identifi ed during a routine nursing check 12 h after the infusion had been started. The baby required reintubation and ongoing ventilatory support following a failed extubation attempt.


Investigation identifi ed that the rate of morphine infusion had not been decreased from 1 to 0.1 ml/h after the initial loading dose had been administered. Other contributory factors included the fact that the infusion pump displayed the fl ow rate but not the drug dose being administered causing confusion among staff regarding the amount of morphine being infused.


The use of medication safety software (Guardrails, Cardinal Health, UK) was implemented in all intravenous infusion pumps. The software prevents inadvertently large doses of drugs being administered by limiting the maximum dose/fl ow rate for an individual drug, selected from a preprogrammed drug library. In addition, information about the loading and maintenance doses of morphine to be infused was clarifi ed in the local pharmacopeia.


Since the introduction of medication safety software, no tenfold dosing errors have been reported for any neonatal medication.

Box 3 Medication errors with insulin


A series of medication errors were reported related to the prescription of intravenous insulin infusions for treating hyperglycaemia in preterm infants. Errors included ambiguous prescriptions, preparation of the wrong concentrations and incorrect labelling of syringes.


Contributory factors identifi ed during review of individual incidents included poor prescribing practice, the need for a double dilution method of preparation and the lack of a process for checking and documenting infusions at the start of nursing shifts.


Various changes in practice were implemented in response to these incidents. Preprinted prescribing and administration stickers were introduced to simplify the prescription of insulin and labelling of syringes; a policy of checking each infusion fl uid type and strength was made mandatory at the start of each nursing shift; and education and training related to the use of intravenous infusion pumps with inbuilt medication safety software was enhanced.


No medication errors involving insulin were reported during the 1-year period following the introduction of this package of interventions.

How common are PSIs in neonatal care?

There is a growing body of literature describing incidents and errors in neonatal practice. A recent systematic review identified 10 studies of which seven specifically reported medication errors.10 Three further studies have subsequently been published.11,,13 Interpretation and comparisons of data relating to neonatal incidents are hampered by differences in definition of what constituted an incident, whether information was collected prospectively or retrospectively, whether reporting was mandatory or voluntary, the method of ascertainment and the population being studied. In contemporary practice, the overall rate of incidents in neonates ranges from 19 to 74 per 100 admissions or 20 to 32 per 1000 patient days.11,,13

National reporting and learning system

The NPSA was set up in 2001 to improve the safety of patient care in England and Wales. It subsequently established the NRLS to collect reports of PSIs and information about their investigation and to formulate and disseminate solutions to enhance patient safety at a national level. It was hoped that the NPSA would be able to help healthcare organisations learn lessons from each others' experiences through publication of good practice bulletins such as Safer Practice Notices and Safety Alerts. Unfortunately, to date, there has been little ‘neonatal-specific’ information issued directly as a result of reports received through the NRLS. It has been proposed that individual Neonatal Networks or a national professional body such as the British Association of Perinatal Medicine might be in a better position to collate and disseminate information relating to critical incidents. However, although desirable, there is currently no such national specialty-based system in neonatal medicine, and few neonatal networks are in a position to perform this task effectively. In 2007, 12 073 neonatal incidents were reported by 265 NHS trusts to the NRLS. This is likely to be a significant underestimate of the true picture, since the NPSA definition of a ‘neonatal’ incident is one that occurs in a patient aged <28 days of age and ignores ‘neonates’ who have experienced an adverse event beyond this period. The majority of incidents caused no harm (64%, including near-misses) or low harm (24%). However, 11% of incidents were considered to be associated with moderate or severe harm, or death. Table 1 shows a breakdown of the total number of neonatal incidents reported to the NRLS according to incident type (personal communication, Jenny Mooney, NPSA).

Table 1

Categories of neonatal incident* reported to National Patient Safety Agency in 2007

Liverpool experience

There are very few published reports describing the nature and frequency of critical incidents from neonatal units in the UK. Liverpool Women's Hospital is one of the largest neonatal centres in the UK, with around 8300 deliveries and 1100 neonatal admissions per annum. Approximately 500 PSIs related to neonatal care are reported annually (table 2); during 2005–2007 there were 1521 incidents equivalent to 44 per 100 admissions and 32 per 1000 patient days. One-quarter of reported incidents were harmful events, and three-quarters were near misses. The vast majority (96%) of incidents were categorised as very low/low risk. High- and moderate-risk incidents comprised almost exclusively significant hospital-acquired infections. Overall, the commonest categories of incident related to medication (24%), equipment (13%), staffing levels (10%) and patient records/identification (10%).

Table 2

Breakdown of incidents reported at Liverpool Women's Hospital

Limitations of incident reporting

Incident reporting is just one way of identifying error in clinical practice. Traditional systems focusing on voluntary reporting detect only a minority of errors that have occurred, usually those that have had a harmful effect on the baby and/or are considered to be more significant or serious. In one study, incident reporting identified only 8% of all adverse events that occurred in the neonatal intensive care unit.11 Other methods include safety ‘WalkRounds,’14 case note reviews with or without the use of focused trigger tools,11 15 16 direct observation of staff performing procedures and a review of clinical negligence claims.

Effective incident reporting is an expensive, labour-intensive and time-consuming process. In our unit, a full-time risk coordinator with a background in neonatal clinical practice has the responsibility for overseeing the smooth running of the incident-reporting system including receiving and reviewing incident forms, convening multiprofessional formal reviews, monitoring progress against action plans and providing education, training and feedback for clinical staff. Support is provided by members of the trust infection team, biomedical engineers, patient data-management team, nurse educators and a departmental pharmacist. Representatives of the different grades of medical and nursing clinical staff attend formal reviews and participate in implementing any resulting action plan. Participation of staff is essential if incident reporting is to be an effective means of changing practice. Individual and group feedback is an important component in educating staff about what should be reported, how incidents are investigated and how action plans are developed and executed. Staff need to be reassured that the process is focused primarily on learning from errors and not designed to assign blame. By definition, learning from incidents is essentially a reactive process. Ideally, effective risk management should also involve processes where prospective risk assessment identifies areas where action is required to protect patient safety. This approach might be preferable in situations where, for example, a new piece of equipment is being introduced or a new clinical policy being implemented.

Critical incident reporting as a measure of patient safety

Incident reporting is not an end in itself. Its ultimate purpose must be to improve patient safety through promoting safer and more effective clinical practice. The quality of a clinical service cannot be assessed by monitoring and reporting numbers of critical incidents alone because voluntary incident-reporting systems usually grossly underestimate the true extent of patient safety problems. Furthermore, incident reporting is a crude and inaccurate tool for measuring patient safety: for example, large numbers of incidents may merely reflect a greater willingness of staff to report (something associated with a stronger learning culture) rather than poor clinical practice. Demonstrating that incident reporting has improved clinical practice is difficult. Ideally, one should be able to show that specific, measurable adverse clinical outcomes (eg, neonatal bloodstream infection rate) have been successfully reduced rather than to point to a reduction in the number of incident reports relating to neonatal infection.

Future challenges

Although incident reporting is now well established in many areas of healthcare including neonatal practice, a number of challenges remain if it is to become an effective means of improving clinical practice.

  • ▶. Healthcare organisations are improving how information about incidents is collected, reported and recorded. However, the lessons learnt through systematic review and analyses of these incidents are poorly shared between different departments within an organisation and between different neonatal units.

  • ▶. It is vitally important for staff to see that incident reporting has an impact on clinical practice. Feedback to staff about incidents that have been reported and the outcome of any detailed review are essential to maintain their enthusiasm for, and commitment to, incident reporting. However, dissemination of information and the lessons learnt from incident reviews remain a challenge.

  • ▶. Staff need to be reassured that the primary goal of incident reporting is not to assign blame for an event, but to investigate and analyse why an error has occurred and how best to prevent it recurring. Involvement of all types and grades of staff in the review process is an important way of gaining their confidence in the process.

  • ▶. In any organisation, the goal is to establish a culture where patient safety is considered paramount, but it must be recognised that this is likely to take several years to achieve.

Examples of improvements in clinical practice following Patient Safety Incidents (PSIs)



  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.