Article Preview
Top1. Introduction
“No-show,” a term characterizing a client who has an appointment of some sorts with another entity and who ultimately fails to show up to that appointment for some reasons, was originally used by airline companies. Today, the no-show term has gained popularity with a variety of other businesses, including hotels and other entities in the hospitality industry as well as the health care sector. As a no-show will always impact negatively on particular businesses, in the health care sector, its impact has been increasingly significant. A patient who misses an appointment, or just fails to arrive on time, will waste that clinic’s and health professional’s time slot, which could otherwise be gainfully deployed to care for another patient (customer).
This article focuses on generating a no-show algorithm. The primary objective is to build a model that offers reliable predictions of a patient’s future behavior on the basis of historical data of appointment characteristics of several healthcare centers and the use of selective parameters from the patients’ data. Hence, not only will the proposed approach account for the behavior of the entire sample population, but the specific behavior of the patient being assessed will also be considered, as this latter behavior can drastically change the no-show probability. Ultimately, results obtained from the algorithm can be used to identify, contact as well as notify patients, thereby verifying the attendance of the respective patient being contacted. If the system detects that a patient may not show (including, patients who do not respond to confirm the notification), the algorithm finds a replacement. When the algorithm ends, whether a replacement is or is not found, the appointment information is updated accordingly, and the healthcare provider is notified of the change.
The rest of the paper is organized as follows. Section 2 covers the background, including related extant literature regarding population-based and individual-based approaches. The developed algorithmic solution is described in section 3. Section 4 then highlights the results of the evaluation performed to validate the algorithm. Finally, section 5 provides the conclusions by discussing briefly the limitation of this work and its practical implications while offering insights into future work.