Knowledge discovery tools for extraction and analysis of practice patterns from labor and delivery data

In this study, knowledge discovery tools were implemented to identify and extract practice patterns from structured clinical data. The implementation process consisted of the selection of a care domain, data collection, identification of patient attributes and outcomes, identification of care setting, modeling clinical and operational processes, and enabling data navigation. This process allowed identification of variation in practice among patient subgroups, and among care facilities. The developed techniques were demonstrated within the Labor and Delivery (L&D) setting supported by electronic nursing documentation.; Retrospective data from 2,611 L&D patients admitted during the 2-month period of January and February 2006 were selected for analysis. A tool was developed to support navigation among patient profiles that enabled the user to view patterns of nursing practice and outcomes associated with specified patient characteristics.; Nursing processes were represented using five scoring models to measure the hourly quantity of nursing care, referred to as Nursing Effort, received by each patient in the study. Scoring models considered nurse-to-patient ratios, the number of concurrent patients being treated by a given nurse, patient acuities, and the number of activities documented for each patient as represented by computerized nurse-entered documentation. Nursing Effort scores, along with patient attributes, were used in an analysis of patient outcomes including the duration of labor stages, the occurrence of various labor events, and cost of care.; Significant differences were detected in the amount of Nursing Effort received according to patient acuity and the facility where care was provided. Using survival analysis techniques, significant relationships were identified between Nursing Effort scores and the durations of various labor stages. Regression models facilitated the identification of significant relationships between particular patient attributes and outcomes.; The development of knowledge discovery tools in this study demonstrates the feasibility of a computerized care process model that provides real-time case profiling and outcome forecasting through the identification of variations in nursing care and patient characteristics. Additionally, generalized benefits of the practice pattern extraction capability, which extend to other clinical domains, are presented with supporting examples from the L&D setting.