What did we do?

Manchester University NHS Foundation Trust has implemented and integrated an innovative Electronic Patient Record solution called Hive. The Hive Programme is bringing about wide-spread change, and improvement, in every part and process in the organisation, transforming how they work for the continued benefit of their patients.  

This project aimed to describe how a predictive model within Hive that supports the identification of patients who are likely to not attend their outpatient appointment can be used to help staff manage outpatient no-shows in the trust.

 

 

Why was it important?

Outpatient no-shows (i.e., missed appointments) are a substantial problem for healthcare systems. The no-show rate across the NHS is 8%, or 5,656,365 no-show outpatient appointments per year, at a cost of £225 million.

 

One of the solutions that has been developed to manage outpatient no-shows is based on the use of predictive modelling. Predictive models could be used to identify patients at high no-show risk, allowing clerical and administrative staff to take action to prevent or manage no-shows. Many such predictive models have been developed, but very few have been evaluated in clinical settings – and none of these studies took place in the UK. There is also very little information about how hospital staff and patients view the use of these predictive models.
 

 

How did we do it?

Throughout this process, we collaborated with our Public and Community Involvement & Engagement (PCIE) panel. In our discussion sessions with the panel, we explored the panel’s views of outpatient no-shows and predictive modelling, as well as interpreting the findings of our study.

 

Before undertaking primary research, we completed a rapid systematic review of interventions aiming to reduce outpatient no-shows by using predictive models, examining the effectiveness of these interventions, as well as the associated costs, acceptability to staff and patients, and effect on health inequities.  

 

Following on from this review we conducted primary research both before, during and after the predictive model was implemented. We observed and interviewed staff, to help us understand how they managed outpatient no-shows, how they intended to use (and then used) the predictive model, and what barriers, facilitators, benefits, and consequences they expected (and then experienced) in the implementation of the model. Drawing on quantitative data also allowed us to examine call volume and changes to non-attendance rates.

 

 

Findings

Our review found that several promising interventions can be used in combination with predictive models. Specifically, predictive model-based reminders and predictive model-based patient navigator phone calls are probably effective at reducing no-shows.  However, it is uncertain whether predictive model-based overbooking is effective. Additionally, the researchers concluded that more evidence is needed regarding the cost-effectiveness, acceptability, and equity of all identified interventions. Quantitative outpatient appointment data revealed that the outpatient no-show rate dropped by around 2% after the implementation of HIVE. However, further analysis is needed to assess whether this reduction was sustained in subsequent years and whether it can be attributed solely to HIVE and the HIVE predictive model, or whether other factors also contributed to the decline.

 

The full rapid systematic review is available from the Journal of the American Medical Informatics Association here.

 

The initial evaluation of the pilot phases is available from BMJ Open here.

 

 

Who did we work with?

• Manchester University NHS Foundation Trust 

  

 Our team included:

• Professor Dawn Dowding
• Dr Norina Gasteiger
• Dr Akbar Ullah
• Dr Louise Laverty
• Professor Roman Kislov
• Dr Anastasia Rousaki
• Dr Anthony Wilson
• Sam Evans
• Amy McCawley
• Thomas Jones
• Cassian Butterworth
• James Laybourne 

  

 

Downloadable resources

 

 

More information

 

 

 

 

Head of ARC-GM

 

Sue Wood