The effects of publishing emergency department wait time on patient utilization patterns in a community with two emergency department sites: a retrospective, quasi-experiment design

Lengthy wait time for access to healthcare services is a persistent issue in many jurisdictions [1, 2], and to make matters worse, wait times for different providers in the same area can vary significantly [3, 4]. While solutions to these problems usually involve capacity increases or efficiency improvement, with the adoption of information technologies in the healthcare system, an alternative or supplementary approach has been proposed and implemented in some jurisdictions: publishing wait time information of different providers so that patients can make informed choices on which provider to use [5–8].

There are good reasons to believe that such an approach might help reduce lengthy wait times. For example, patients are known to be willing to travel or to switch healthcare providers in order to achieve shorter wait times [9–14], and if more patients choose providers with shorter wait times, then lengthy wait times might be less likely to occur. Such effects, however, have not been empirically demonstrated.

One of the most promising areas of such application is emergency department (ED) care, as lengthy waits in EDs are widespread in many communities [15–18]. Wait times in different ED sites in the same community can differ significant [8, 19, 20], and patients' selection of providers is less constrained by factors such as referral, continuity of care, and familiarity with providers, as patients usually do not require a referral to visit an ED, and patients usually do not have the ability to select care providers in an ED.

In this study, we aim to test the hypothesis that publication of wait time information would lead to more patients visiting the ED site with shorter published wait time, and that such a change in utilization pattern would lead to reduction in lengthy wait times in a community with two ED sites. This study was reviewed and approved by the Research Ethics Board at the University of Western Ontario.

Figure 1a presents the wait time for those discharged without admission at VH, Figure 1b presents the differences in wait times (VH-UH), Figure 1c presents the combined volume of these two sites, and Figure 1d presents the portion of patients who went to VH.

The wait times in these two sites can vary significantly, and there seemed to be no clear pattern in the relative wait times in these two sites. This variation and lack of clear pattern suggest that wait time information provides valuable information, as patients would not be able to predict the wait times of these two sites using previous experiences and/or using external factors such as time of the day or day of the week.

We found that the rates of wait times exceeding 4 h and the 95th percentile of wait times in the two sites decreased after the publication of wait time information, even though the average wait times experienced a slight increase. We also found that after controlling for other factors, the site with shorter wait time had a higher likelihood of being selected after the publication of wait time information, but there was no such relationship before the publication. These findings were consistent with the hypothesis that the publication of wait time information leads to patients selecting the site with the shorter wait time.

Due to the lack of randomization or a control community, it is difficult to establish conclusively that publication of wait time information caused patients to select the site with shorter wait time, as there may be alternative explanations of the relationship between publication of wait time information and shift in patient utilization patterns.

Nevertheless, we have reasons to believe that the results were not due to alternative explanations. More specifically, a range of factors, including local disease outbreak, construction work that made one of the ED sites more difficult to reach, or capacity changes in one or both sites, could have happened coincidently with the publication of ED wait time information and could have altered patients' selection of ED sites. However, no significant capacity change was made to either site during the study period.

Moreover, changes caused by one or a combination of these factors would persist for the entire duration during which these factors were present, but the daily volume graph did not show evidence of any sustained change.

Given the variability and the lack of clear pattern of the relative wait times of these two sites, it seems unlikely that the relative wait time would be associated with any of the aforementioned factors. Our findings provide support to the usefulness of providing wait time information to the public in addressing lengthy wait times for healthcare services.

Beyond the immediate impact on patient flows and changed wait times for those patients who use the web site, the publication of wait time information may also have broader implications.

For example, in communities with alternative providers to EDs, such as walk-in clinics or urgent care centers, publication of ED wait time information may have implications on relative patient volumes among ED sites and these alternative care providers if the patients use published ED wait times to decide not only which ED site to visit, but also whether or not to visit ED or alternative providers. This possibility is especially important in jurisdictions with universal insurance coverage such as Canada, where ability to pay is not an issue affecting patients' choices of healthcare provider. Our study does not provide any information on such implications, but it is a worthwhile topic to pursue.

Surveys also show that it is the perceived wait time, not the actual wait time, that influences patient satisfaction with EDs [25–27]. Publishing ED wait time information could enhance patient satisfaction if such published wait time is a reasonably accurate prediction of actual wait time; if the published wait time does not reflect the patient's experiences, on the other hand, patient satisfaction could be negatively affected. Given the variability in wait times of individual patients, it is unclear how accurately a summary statistic can reflect the experiences of these patients, but the implications for patient satisfaction is a topic that warrants further investigation.

The findings of our study are consistent with the hypothesis that publication of wait time information would lead to more patients visiting the ED site with shorter published wait time. The impacts of such a change in utilization pattern on ED wait times were mixed: rates of wait times exceeding 4 h and the 95th percentile of the wait times in both sites had a statistically significant decrease after the publication of wait time information, although the average wait times slightly increased in both sites (such increases did not achieve statistical significance).

The patient consent requirement was waived by the Research Ethics Board for this study, as only de-identified data were used and it was not practical to obtain patient consent for such a retrospective study.