Inpatient census, or occupancy, is a primary driver of resource use in hospitals. Fluctuations in occupancy complicate decisions related to staffing, bed management, ambulance diversions, and may ultimately impact both quality of patient care and nursing job satisfaction. We describe our approach in building a computerized model to provide short-term occupancy predictions for an entire hospital by nursing unit and shift. Our model is a comprehensive system built using real hospital data and utilizes statistical predictions at the individual patient level. We discuss the results of piloting an early version of the model at a mid-size community hospital. The primary focus of the paper is on the development and methodology of a second generation of the predictive occupancy model. The results and accuracy of this new model is compared to a variety of other predictive methods based on tests using 2 years of actual hospital data.
One of the major obstacles to using organizational data for mining and knowledge discovery is that, in most cases, it is not amenable for mining in its natural form. Using a data set from a large tertiary care hospital, we provide strong empirical evidence that data enhancement by the introduction of new attributes along with judicious aggregation of existing attributes results in higher quality knowledge discovery. Interestingly, we also found that there is a differential impact of data set enhancements on the performance of different mining algorithms. We define and use several measures including entropy, rule complexity, and resonance to evaluate the quality and usefulness of knowledge discovered.
Simulation studies of outpatient clinics often involve significant data collection challenges. We describe an approach for data collection using sensor networks which facilitates the collection of a large volume of very detailed patient flow data through healthcare clinics. Such data requires extensive preprocessing before it is ready for analysis. We present a general data preparation framework for sensor network generated data with particular emphasis on the creation and analysis of patient path strings. Several examples of the analysis of sensor network data are also presented. Our approach has been used in two large outpatient clinics in the United States.
Increases in the rate of births via cesarean section and induced labor have led to challenging scheduling and capacity planning problems for hospital inpatient obstetrical units. We present occupancy and patient scheduling models to help address these challenges. These patient flow models can be used to explore the relationship between procedure scheduling practices and the resulting occupancy on inpatient obstetrical units such as labor and delivery and postpartum. The models capture numerous important characteristics of inpatient obstetrical patient flow such as time of day and day of week dependent arrivals and length of stay, multiple patient types and clinical interventions, and multiple patient care units with inter-unit patient transfers. We have used these models in several projects at different hospitals involving design of procedure scheduling templates and analysis of inpatient obstetrical capacity. In the development of these models, we made heavy use of open source software tools and have released the entire project as a free and open source model and software toolkit.
Spiraling health care costs in the United States are driving institutions to continually address the challenge of optimizing the use of scarce resources. One of the first steps towards optimizing resources is to utilize capacity effectively. For hospital capacity planning problems such as allocation of inpatient beds, computer simulation is often the method of choice. One of the more difficult aspects of using simulation models for such studies is the creation of a manageable set of patient types to include in the model. The objective of this paper is to demonstrate the potential of using data mining techniques, specifically clustering techniques such as K-means, to help guide the development of patient type definitions for purposes of building computer simulation or analytical models of patient flow in hospitals. Using data from a hospital in the Midwest this study brings forth several important issues that researchers need to address when applying clustering techniques in general and specifically to hospital data.
We describe a loosely integrated set of teaching modules that allow students to explore decision support systems (DSS) topics ranging from data warehousing to model based decision support. The modules are used as part of a spreadsheet based modeling approach to teaching DSS. We describe one particular module in detail in which students are forced to deal with a muddy dataset as a prelude to analysis, modeling and decision support system development.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.