Research project Congested but underutilized infrastructures:
How information can be used to regulate demand in healthcare
Many of the resources in our society (hospitals, roads, etc.) are congested at certain times but underutilized at others. This project set out to investigate the potential of regulating demand over time and space to create a more even utilization and a more efficient use of existing resources.
In particular, this project aimed at exploring how demand (patient arrivals) can be influenced at the Emergency Department (ED). Previous research indicates that patients often wait hours or days, until a convenient time, before visiting the ED. This leads to demand peaks at the ED with congestion and long waiting times as a consequence. In cooperation with Google and Vårdguiden (Swedish health care guide) we planned to use forecasting models to present predicted and current waiting times to patients online.
Taking this new information into account, patients would be given the opportunity to reduce their expected waiting time at the ED by coming at a less congested hour or by choosing another hospital. The anticipated outcome of this policy was a reduction in demand variability over time and over hospitals in the Stockholm region, ultimately leading to reduced costs, improved patient care and a more efficient use of resources.
From a research perspective the objective of this project is to improve our understanding of how demand can be regulated in a congested system (stores, hospitals, roads, etc.): an overarching question with strong implications for research, policy and practice.
Project description
One of the main proposed methodologies in this project was to run a large-scale field experiment where waiting time information was to be presented online for the seven emergency departments in the Stockholm region. These experiments were approved by SPESAK (the expert council for emergency care), Stockholm Regional Council and the Ethical Approval Board. Unfortunately, Stockholm Regional Council decided, during the course of the project, that the experiment could not be performed.
Due to this decision, the project focus was shifted to more observational studies analyzing care seeking behavior, what factors influence the behavior and show through the development of a real-time web investigate the feasibility of impacting this behavior.
Results
Study 1 structurally estimates and simulates patient care seeking behavior and shows that wait times can be reduced by as much as 14% just by communicating wait time information to potential patients.
Study 2, a survey of ED patients, supports the conclusions from the simulation and show that patients are indeed willing to wait at home in order to reduce wait times at the ED.
Study 3 & 6 show that there are many other factors than disease urgency that influence the decision if and when to seek medical care at the ED.
Study 4 develops a new model that accurately predicts ED inflow on an hourly level. Such forecast could have great importance in reducing wait times if implemented in the planning of resources.
Study 5 investigates how well potential patient follow the advice given by 1177 Vårdguiden and show that most, but not all, comply (65%). This would indicate that 1177 could play a pivotal role in coordinating patient flows.
Finally, we developed a tool that continuously collects ED inflow data, makes forecast and presents them in real-time through a web interface. This shows that it is practically and financially feasible to make this information available to end users.
Conclusions
Both our empirical and simulation studies support our thesis that there is a large potential to reduce waiting times by communicating current and forecasted emergency department workload and/or wait times. We believe that 1177 Vårdguiden should play an integral role in communicating such information.
Project members
Project managers
Olov Isaksson
Associate Professor, Docent
Members
Fredrik Eng Larsson
Associate Professor
Martin Nordberg
MD, PhD
