Our paediatric nephrology centre provides comprehensive acute and chronic dialysis, apheresis and kidney transplantation services to Southeast England serving a population of over 1.7 million children and adolescents. A core team of haemodialysis specialist nurses provides dialysis care. We also have a small number of children’s kidney nurses, less experienced in dialysis treatments, working on the unit regularly. There is usually one experienced dialysis nurse for every two patients receiving haemodialysis and one nurse per patient for complex therapies, clinically unstable, or young children.

An HD Trigger Tool (HTT) to detect and monitor harm events specific to HD has been developed for adults receiving in-centre HD in Derby, UK [7]. The adult HTT was discussed extensively at our departmental dialysis staff meeting, with input from dialysis nurse specialists. The adult HD trigger tool was adapted for a paediatric HD population and also includes additional local safety practices. We produced a ‘per-dialysis session’ pHTT, referred to as the ‘The Evelina London pHTT’, that includes 54 trigger events across six domains: general care, vascular access, dialysis, medication, laboratory tests, and other (Table 1). Individual triggers in each domain were adapted to the paediatric population (for example, post-HD weight targets; ultrafiltration rate). We also incorporated local operating procedures into the HTT — for example, our standard practice to review medications and blood results with the parent or caregiver weekly, and our daily ‘safety huddle’ of all dialysis staff.

The ‘safety huddle’ is a mid-morning update between all nurses on the unit, ideally once all the patients are on their machines and before any staff go on breaks. Typically, this is a 15-min meeting (‘huddle’) to discuss the patients’ weight, access, investigations completed or due, social concerns, and time planned for the session. This provides an opportunity for the nurse-in-charge to review the dialysis session plan, ensure medications are given and blood results are checked, and any concerns are escalated appropriately. It also informs the rest of the nurses about each patient’s plan for the session so they can work as a team/cover breaks.

Each trigger was evaluated for level of physical harm and categorised as shown in Fig. 1. We defined harm as anything caused by a medical intervention that caused discomfort or symptoms for the patient or required intervention to correct. We categorised harm as (i) no harm (category A); (ii) temporary harm requiring intervention (categories B and C); or (iii) permanent harm (categories D, E, and F). Preliminary use of the pHTT tool was conducted by two dialysis nurses who confirmed ease of use, minimal time commitment, and no significant addition to their workload. All HD nurses at our centre were trained to complete the pHTT, and completed the pHTT at the end of each HD session in < 5 min per patient. Allocation to harm categories was undertaken independently by two authors and any differences between them discussed, and then a single category was agreed upon. An example of how different harm categories were allocated to a trigger event is illustrated in Fig. 2.

The categorisation of each trigger for level of physical harm using the ‘per dialysis session’ paediatric Haemodialysis Trigger Tool (pHTT) adapted from the Derby HD Trigger Tool (HTT) [7]

Example of the application of the ‘per dialysis session’ paediatric Haemodialysis Trigger Tool (pHTT) during a typical in-centre haemodialysis session. This example is of one single haemodialysis session, showing how the harm rating was arrived at for that session. The level of harm was assigned based on the outcome at the end of that specific session. It is though logical to assume that a series of such “no harm events” are likely to result in a “harm event” in the future

Prospective data collection for all children receiving in-centre HD was performed over an initial 8-week period. A second 8-week cycle of using the pHTT was repeated after a 3-month interval. The results of both cycles are reported here. The results of the pHTT were reviewed by a member of the medical team on a weekly basis.

Data are displayed as total number (and percentage), mean with 95% confidence interval, or median and interquartile range. We calculated the ‘trigger rate per session’ over each cycle and defined it as the number of triggers divided by the number of HD sessions. We calculated the ‘harm rate per session’ over each cycle and defined it as the number of triggers resulting in harm (temporary harm requiring intervention; or permanent harm (categories B–E)) divided by the number of HD sessions over each cycle. Statistical significance was considered if P ≤ 0.05.