A Process Model to Support Automated Measurement and Detection of Out-of-Bounds Events in a Hospital Laboratory Process

Article excerpt

Abstract

Business Activity Monitoring (BAM) allows organizations to capture enterprise events from their source systems and utilize these to detect non-compliant business situations. Similar concepts may be leveraged in the healthcare domain to improve the quality of patient care and the efficiency of clinical processes. This paper introduces a generic set of constructs for formally specifying threshold values relevant for cycle time and utilization calculations. It also describes a mechanism to capture information, including thresholds, about important business parameters for Six Sigma measurement. This full set of constructs are the basis for automated measurement and monitoring and are incorporated into the process model during the definition or capture phase thereby linking the definition and monitoring phases through a common underlying process model. Bespoke software is also described which uses the constructs contributed by this research to manage and monitor process models and enterprise events. A process performance module provides automated measurement and monitoring capabilities. At an aggregate level, this is achieved through the provision of process cycle time data for selected time periods on demand and the examination of business processes at frequent intervals with alerts generated for exceptional scenarios. At a more granular level, this solution uses a rules-based approach to evaluate individual events and generate alerts for out-of-bounds business parameters. This paper demonstrates the benefits of these capabilities for health informatics through application to a Laboratory Testing process observed at a local hospital. The paper also suggests recommendations for the extension of current modelling languages with respect to the constructs detailed herein.

Key words: Business activity monitoring, Process performance measurement, Process modelling and definition languages, Event modelling, Six Sigma

1 Introduction

Cantara et al. of Gartner Research predict that by 2012, "organizations that don't take an end-to-end process view will compromise their business performance" [5]. This process view requires clear definition and understanding of business processes. A business process is one of the core elements of organizational 'being'. Many definitions exist, but two are selected here. A process "contains purposeful activity", "is carried out collaboratively by a group", "crosses functional boundaries", and is invariably driven by the outside world" [25]. A process is designed, foremost, to deliver value to its customers (internal or external) and comprises a sequence of steps (actions, tasks, activities or functions) potentially spanning multiple departments, and even organizations. A business process is "a collection of interrelated work tasks, initiated in response to an event, that achieves a specific result for the customer of the process" [28]. From these two definitions, certain characteristics of business processes emerge. They are customerfocused, cross-functional, have a specific scope, require a precise definition of each task and have a measurable output. Despite this seemingly logical design, organizations are typically compartmentalized into departments making it difficult to manage cross-functional processes. Organizations find themselves functionally-oriented instead of process-oriented [28].

The wider context for the research detailed in this paper is process improvement. Organizations typically apply the Define, Measure, Analyse, Improve, Control (DMAIC) approach to (continuous) process improvement. This is a closed feedback loop and of these five phases, the first three relate directly to this research. The measure and analyse phases may be facilitated through the implementation of Business Activity Monitoring (BAM) solutions based on event-driven architectures for delivering operational-level monitoring capabilities. Although containing the terms "business activity" in the title, this approach can be applied to any domain. …