Data source

The data source for this study is the National Reporting and Learning System (NRLS). This is a national repository of anonymised patient safety incident reports from National Health Service (NHS) organisations across England and Wales.22 Individuals, organisations and NHS Trusts can voluntarily submit reports to the national repository. They are encouraged to report any ‘patient safety incidents’, defined as ‘any unintended or unexpected incident that could have or did lead to harm for one or more patients receiving NHS-funded healthcare’.22 They contain categorical information which includes patient demographics, level of harm and location and date of the incident. There are also open text boxes for information about what happened and why it happened. More information about the NRLS data is available on their website.22

Sample selection

A sample of incidents relating to gastrostomy, jejunostomy and NG feeding in paediatrics were requested from NRLS to include incidents reported between 1st August 2012 and 31st July 2017. The following free text search terms were used to identify the incidents: Gast* button, G-button, mickey button, enteral feed, NGT, NG tube, NG feed, naso-gastric feed, naso-gastric tube, jejunostomy feed, jejunostomy tube, jejunal feed, jejunal tube and gastrostomy. A total of 9327 incidents were received from NRLS.

The incidents were first filtered by reported incident location to identify reports occuring in the home and then by age to remove incidents involving patients over 18 years. These 349 incidents were manually reviewed to exclude incidents without a clear description, not relating to enteral feeding, not relating to home care and any remaining reports involving adults. This produced a final sample of 268 incidents for analysis. Figure 1 shows the flow diagram illustrating the steps taken to identify the sample.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Flow diagram showing the steps taken to identify the final sample of incidents for inclusion.

Analysis

The selected incidents were imported into NVivo, V.12. The free text descriptions were coded to identify the problems in care, any stated contributory factors and patient outcomes. An adapted framework approach was used.23 Two authors (BP and RN) familiarised themselves with the data and noted relevant themes. BP and RN coded 10% of the incidents identifying the problems in care, contributory factors and patient outcomes. Initial agreement was over 90%. Discrepancies were resolved through discussion and the two authors then coded half of the remaining incidents each. Incidents which were unclear were discussed with author SH, a community paediatrician, who assisted with the analysis of the clinical circumstances described in these cases. The problems in care, contributory factors and patient outcomes were then grouped into themes and subthemes through an iterative process by adapting existing general incident frameworks to fit the narrow field of enteral feeding in the home.24–26 A sample of 10% of the incidents were coded independently by author AL, who has significant clinical expertise in surgical enteral feeding. Agreement was 100% for outcomes and care problems and 95% for contributory factors. Author CV, with expertise in incident analysis, carried out a final check of the frameworks. Example incidents and coding are shown in online supplementary file 1. A more detailed description of the analysis process is available in online supplementary file 2.

Ethics

The incidents from the NRLS were anonymised and made available by NHS Improvement through a data sharing agreement with the Oxford Academic Health Science Network (AHSN). This was part of a service improvement project and conducted under the auspices of the Patient Safety Collaborative at Oxford AHSN as part of their regional Specialised Paediatric Care programme.

Problems identified in the processes of care

At least one problem in care was identified in each incident, with some incidents having two or three problems. The problems in care fell into nine different categories (see table 1). The most common categories were equipment and devices (n=98, 28%), procedures and treatments (n=86, 24%), information, training and support needs of families (n=54, 15%), feeds (n=52, 15%) and discharge from hospital (n=31, 9%). Incidents occurred across the age span, with 32% occurring in children under 1 year, 26% in children 2–4 years, 27% in 5–11 years and 9% in children 12–17 years (in 6% of incidents the age was unknown).

Table 1 shows the specific problems identified within each of the categories. Faulty and damaged equipment was a very common problem (n=62, 18%). It is unclear to what extent the problems with faulty and damaged equipment are underpinned by poor design or by inappropriate use of equipment. Many of the problems highlighted may relate to inadequate training or knowledge of nurses and paid carers. For example, medications and feeds were administered through the wrong port of a gastrojejunostomy tube in 12 incidents, and in two incidents, medication was wrongly inserted into the balloon part of the gastrostomy button device.

Outcomes for the child

Table 2 shows the breakdown of outcomes for each incident. There was a clearly stated harm to the child in 52 (19%) incidents, including 17 (6%) incidents which resulted in a hospital admission or accident and emergency visit. Some of the incidents in the potential harm category may have resulted in harm to the child which was not stated by the reporter. In some of the incidents classified as ‘potential harm’, there was a clear potential for harm but no actual harm occurred. An example of this would be where a child was fed through an NG tube despite being unable to obtain aspirate and a suitable pH value.

Factors contributing to the incidents

There were 97 contributory factors identified in the incidents. In the majority of incidents, no contributory factors were mentioned. Contributory factors fell into five broad categories: organisational factors (32%, n=31), staff factors (21%, n=20), family carer factors (20%, n=19), feeds, equipment and medication factors (12%, n=12) and patient factors (15%, n=15). Table 3 gives definitions and example quotes for each category. Organisational factors such as poor communication between services, lack of service availability and evening and weekend discharges were common. These factors highlight the transition from hospital to home as a particularly risky period and concerns regarding the availability of community services to support families. The circumstances of the family carer, such as the involvement of a secondary family carer (eg, grandparent), ongoing child protection issues or a parent experiencing distress, all affected the provision of care. The training needs of individual staff members was also a common problem and raises questions about the safety of care in the community in some services.

Implications for clinical practice and organisation of services

If children with complex medical needs are to be cared for at home, the provision of services to support them needs to be strengthened in a number of respects. First, handovers between hospital and community need improving so that all children are safe in the critical first few weeks at home. Second, family carers need consistently good quality training. Third, sufficient provision of community services is needed with the required expertise to support these families, and fourth, equipment needs to be reliable, with back-up equipment available in the community.

The weeks following discharge from hospital are a high-risk period. We found a number of instances where community teams had not been informed of the child’s discharge, and cases where children were discharged without the required equipment, medication or feeds. It is vital that there is continuity of care between hospital and community services. Pressure to discharge patients due to bed shortages may be increasing the risks. Standardised checklists could help address some of the problems identified at discharge. It can be tempting to think the problems identified in the incidents are mostly related to the transition from hospital and home, but the range of ages of the children involved suggests there are also considerable problems in long-term care.

Inadequate training of family carers was a frequently reported concern. Other studies have found evidence of safety concerns in the practices of some family carers.16–18 Many families also worry about making mistakes or feel inadequately prepared.29 30 The adequacy of training and information for parents needs to be viewed as a system issue and vital to the safety of care at home.31

Underlying a number of the problems identified, is inadequate provision of services in the community to support families. Lack of expertise and availability of services for specialised paediatric care have been identified by others previously, including parents.32 33 Our study indicates that there are varying levels of expertise among those who provide care, whether that be parents, Community Children’s Nurses, paid carers or school or respite staff. Established nutritional support teams need sufficient time to train families and other community professionals, to ensure that as much as possible, day-to-day care and minor complications are safely managed at home. There needs to be cross-pollination of expertise across services.

Different surgical/gastroenterological specialists and feeding healthcare professionals use different devices, which creates a complex landscape for parents and community services to navigate. Increased standardisation is needed. Improvements are needed to the reliability of equipment and provision of back-up equipment, or children will miss vital feeds and medications.

Box 1 gives a list of recommendations for action and further investigation. Key recommendations are given for the four major themes identified in this study. If children with complex needs are to be cared for safely at home, services to support families must be strengthened in these four areas.

Strengths and limitations

Incident reports are excellent tools for learning and for generating improvements to current systems. Patient safety incidents have been extensively studied in the hospital setting, but to date there has been limited research in other care settings.23 34 Our study is, to our knowledge, the first to examine incidents relating to the safety of enteral feeding at home. It documents a range of problems that need attention. More broadly, this study begins to examine a new area of research in the field of patient safety: care in the home and the involvement of family members in providing this care.35

The limitations of incident reporting have been discussed in detail elsewhere.36 Incident reporting generally only detects a small proportion of the total number of adverse events occurring.27 Our study therefore cannot comment on the frequency of safety problems with enteral feeding devices at home. We note that the terms ‘PEG’ and ‘PEJ’ were not included in the search terms so it is possible that some relevant incidents were missed. The majority of these reports were written by healthcare professionals and, as a result, our study cannot adequately explore the perspective of parents. Ideally, families should be more involved in incident reporting as they are the primary caregivers.