Virtual Warehouse Research Articles

Plenary III-01: A Methodology for Quantitative Measurement of Quality and Comprehensiveness of a Research Data Repository

Background and Aim: With the existence of research over federated repositories, it is desirable to utilize high quality integrated data repositories (IDRs). IDR can be defined as a data warehouse optimized for research purposes rather than clinical care, which contains clinical, administrative, trial, and -omics data. In this work, we focus on the quality of clinical and administrative components of IDRs. There is no standardized methodology which could quantitatively evaluate the quality of an IDR (e.g., ‘Does a given IDR have at least 2000 adult diabetic patients (type 1) with complete pediatric history?’). With the increased interest in research of existing data (such as comparative effectiveness research) and increasing number of institutions with an comprehensive IDRs, it is important to have a mechanism for selecting quality IDRs. Methods: Our poster will present a set of IDR quality measures which can compare IDRs in size and completeness. We considered the following criteria for a good measure: the measure is intuitive to interpret; facilitates monitoring improvement; and does not place any arbitrary value on individual measure components. Our methodology proposes a hierarchy of definitions of minimum EHR elements and uses a simple count of each level to quantitatively evaluate an IDR (e.g., count of patients with at least one diagnosis and one laboratory result). Results: We have applied our methodology to an IDR at Marshfield Clinic. Our poster will list all measures and results. Selected results were: 1.7M unique patients (level G1), 0.4M patients with at least one diagnosis, lab and prescription (level D3). To facilitate evaluations at other institutions, we have created an ANSI-SQL script which can compute all measures in a single execution. Conclusion: Our evaluation methodology provides a quick way to compare IDRs at different institutions. It can be applied to institutions contributing to a virtual warehouse. Our goal was to arrive at a pragmatic set of measures operating on an easy-to-implement event schema. The limitations are: focus on general research idea and event types and criteria included in some level definitions. We plan to conduct a Delphi study involving informatics experts to arrive at an improved consensus set of measures.

A virtual warehouse system for production logistics

Rapid change of product demand impacts the master production schedule and affects the stability of production. Traditional inventory planning generated by advanced planning and schedule system is inadequate to handle the effect of variability, non-linear relations, and other abrupt factors on product demand. This paper presents a virtual warehouse system (VWS) which is capable of refining inventory planning so as to streamline product planning and control if there is an abrupt change in the product demand. It can also serve as an advisory tool for supporting decision making and monitoring the inventory allocation of WIP when supply shortage occurs. The VWS is based on the concept of the postponement of product differentiation, a simulation model and knowledge-based system. The inventory, which was originally planned for current and future demands, will act as a buffer to be reallocated and to fulfil a sudden demand. The flexibility of inventory stirring is of key significance in this study. A prototype VWS was built and a successful trial implemented in a prestigious manufacturing enterprise. The results show that the inventory level across the supply chain among different parties can be minimized and the manufacture enterprise more predictable and adaptable to an abrupt change in product demand.

Designing Across Borders

This article discusses that many companies are installing a companywide product data management (PDM) system, also known as an engineering data management system. These systems do several things, besides acting as a central data repository. Much of the data that overwhelms companies comes from engineers using a variety of software applications, such as computer-aided design, manufacturing, or engineering programs. The PDM systems manage engineering data that may come from a myriad of sources and applications but pertain to the development of one product. The chosen PDM system would have to track the design and manufacture of each part that would be assembled to form the finished printer. It would have to tell users what stage in the process each part had reached. But it would also have to serve as a virtual warehouse for all digital versions of a part and product, including older versions that could not be lost because they might be needed for reference at any time.

The virtual warehousing concept

The virtual warehouse (VW) is a state of real-time global visibility for logistics assets such as inventory and vehicles. The efficiency of the VW is comparable to that achieved in a local warehouse facility. We assess the enabling technologies, develop a conceptual framework, and identify the essential decision-support modules of a virtual warehouse. A case study for field repair service illustrates potential benefits of perfect information, in terms of reduced transportation, labor costs and service times. The case study pertains to field logistics support for business telephone systems. The productivity and utilization of highly skilled field technicians are improved substantially. A field scrap policy supported by VW reduces the volume of returns to the repair depot by over 20%.