Development and Evaluation of MADDIE: Method to Acquire Delivery Date Information from Electronic Health Records

Objective: To develop an algorithm that infers patient delivery dates (PDDs) and delivery-specific details from Electronic Health Records (EHRs) with high accuracy. Materials and Methods: We obtained EHR data from 1,060,100 female patients treated at Penn Medicine hospitals or outpatient clinics between 2010-2017. We developed an algorithm called MADDIE: Method to Acquire Delivery Date Information from Electronic Health Records that infers a PDD for distinct deliveries based on EHR encounter dates assigned a delivery code, the frequency of code usage, and the time differential between code assignments. We validated MADDIE's PDDs against a birth log independently maintained by the Department of Obstetrics and Gynecology. Results: MADDIE identified 50,560 patients having 63,334 distinct deliveries. MADDIE was 98.6% accurate (F1-score 92.1%) when compared to the birth log. The PDD was on average 0.68 days earlier than the true delivery date for patients with only one delivery (± 1.43 days) and 0.52 days earlier for patients with more than one delivery episode (± 1.11 days). Discussion: MADDIE is the first algorithm to successfully infer PDD information using only structured delivery codes and identify multiple deliveries per patient. MADDIE is also the first to validate the accuracy of the PDD using an external gold standard of known delivery dates as opposed to manual chart review of a sample. Conclusion: MADDIE infers delivery dates and delivery-specific details from the EHR with high accuracy and relies only on structured EHR elements while harnessing temporal information and the frequency of code usage to identify accurate PDDs.

A Hybrid Statistical Method for Accurate Prediction of Supplier Delivery Times of Aircraft Engine Parts

Aircraft engine assembly operations require thousands of parts provided by several geographically distributed suppliers. A majority of the operation steps are sequential, necessitating the availability of all the parts at appropriate times for these steps to be completed successfully. Thus, being able to accurately predict the availabilities of parts based on supplier deliveries is critical to minimizing the delays in meeting the customer demands. However, such accurate prediction is challenging due to the large lead times of these parts, limited knowledge of supplier capacities and capabilities, macroeconomic trends affecting material procurement and transportation times, and unreliable delivery date estimates provided by the suppliers themselves. We address these challenges by developing a statistical method that learns a hybrid stepwise regression — generalized multivariate gamma distribution model from historical transactional data on closed part purchase orders and is able to infer part delivery dates sufficiently before the supplier-promised delivery dates for open purchase orders. The hybrid form of the model makes it robust to data quality and short-term temporal effects as well as biased toward overestimating rather than underestimating the part delivery dates. Test results on real-world purchase orders demonstrate effective performance with low prediction errors and constantly high ratios of true positive to false positive predictions.