"How long until I am seen, doc?" Modelling paediatric emergency department waiting times to make personalised predictions.

Abstract

Background: ED patient wait times have been progressively increasing leading to patient dissatisfaction in ED. Managing patient expectations towards wait times in ED may be more effective at decreasing dissatisfaction than shortening actual wait times. Models for predicting wait times have been made for general EDs but not for solely paediatric departments. We aimed to create a model that could predict the personalised wait time of a child presenting to paediatric ED after triage.

Methods: This was a single-centre retrospective study analysing all ED attendances to the Bristol Royal Hospital for Children between 1 January 2022 and 31 December 2022. From anonymised routinely collected administrative data, we created a multiple linear regression model to predict wait times. We developed the model by randomly assigning 80% of the data to a training set and used the remaining 20% as a validation set to assess the accuracy of our model. CIs were calculated using 500 bootstrap iterations sampled from the validation set. Understanding that patients are satisfied being seen sooner than their predicted wait time, we considered the result to be unsuccessful if their actual wait time was 30 min over their predicted wait time.

Results: From 40 828 ED presentations, the median patient wait time was 65 min (IQR 34-122). Our model was able to predict wait times for 84.2% (95% CI 83.42% to 84.91%) of attendances successfully. Triage category, number of patients waiting, number of patients in the department, time of presentation, length of wait of last patient and day of week all had a significant impact on prediction of wait times (all p<0.001).

Conclusion: Tailored models created using routine data can be used to give individualised predictions for wait times in paediatric ED, which could be given to patients with the aim of managing expectations and improving patient satisfaction.