Applying Predictive Modeling to Identify Patients at Risk to No-Show

Patients not attending appointments is a common problem in healthcare. This often results in compromised patient care, wasted clinical resources, and limited access for other patients. Most of the current strategies to manage "no- shows" are proven to be ineffective and costly when applied...

Full description

Saved in:
Bibliographic Details
Published inIISE Annual Conference. Proceedings p. 2370
Main Authors Dravenstott, Ronald, Kirchner, H Lester, Strömblad, Christopher, Boris, Derek, Leader, Joseph, Devapriya, Priyantha
Format Journal Article
LanguageEnglish
Published Norcross Institute of Industrial and Systems Engineers (IISE) 01.01.2014
Subjects
Clinics
Datasets
Electronic health records
Endocrinology
Engineering
Health care access
Marital status
Medical practices
Network topologies
Neural networks
Patients
Population
Predictions
Primary care
Sensitivity analysis
Online AccessGet full text

Cover

More Information
Summary:Patients not attending appointments is a common problem in healthcare. This often results in compromised patient care, wasted clinical resources, and limited access for other patients. Most of the current strategies to manage "no- shows" are proven to be ineffective and costly when applied to the entire population, and few efforts have been made to accurately target and intervene on patients likely to no-show. This research aims to construct models using Artificial Neural Networks to predict the risk of a no-show for a given appointment. The dataset for this study contains 3 million appointments over a two-year period and includes clinic, provider, and patient specific attributes. Models have been created for Primary Care and Endocrinology services separately. A validation cohort was randomly selected to include 25% of the data for each service. The final models yielded an area under the ROC curve of 0.78 and 0.81 on the validation cohorts for Primary Care and Endocrinology, respectively. The models provide the opportunity to identify patients with a high likelihood to no-show, enabling targeted interventions that have the ability to reduce no-show rates, and in turn improve patient care, provider utilization, and patient access to care.