A conversation with the Argonne team
In the Winter 2008 issue of JHIM, Dariusz Blachowicz, John H. Christiansen, Archana Ranginani and Kathy Lee Simunich (the Argonne team), explore agent-based modeling and simulation (ABMS), a technique that may help healthcare organizations more effectively determine the future return-on-investment of regional healthcare delivery and electronic healthcare record systems. [You can read their article “How to Determine Future EHR ROI: Agent-Based Modeling and Simulation Offers a New Alternative to Traditional Techniques” in the JHIM-Digital Edition by logging on to the Member Center—Ed.]“Successfully testing, validating and communicating the expected consequences of alternative business practices, processes, protocols and policies requires an objective analytical approach,” the teams writes. “ABMS of healthcare delivery can provide actionable guidance for decision-makers by enabling healthcare experts to define the individual, agent-level rules of operation; allowing them to see how the agent rules play out over time in a detailed real-world context; providing them with the tools to assess the consequences of alternative plans; and giving them a clear method for communicating results to the broader stakeholder community.”
Since President Bush’s April 2004 executive order calling for the establishment of a national system of interoperable EHR for all Americans by 2014, many communities have begun the process of transforming from paper-based to electronic-based records, but their level of success has varied, according to the authors. For many organizations creating a sustainable business model is a major challenge for health information exchange efforts.
“Quantifying the value of EHR networks on healthcare delivery involves understanding the impact of many interrelated factors involving patient care, clinical practices, medical outcome and organizational structure. Some healthcare organizations have successfully employed modeling and simulation to gain a broader perspective on their business processes across multiple areas before they invested a lot of time and money implementing changes, and have thus avoided costly and ineffective process changes,” the authors note.
Agent-based modeling and simulation represents a new approach. According to the authors, “ABMS allows decision-makers to experiment using a simulated healthcare system that closely mimics the real world. With very little upfront investment, policymakers can use agent simulations to pose various health delivery and use scenarios, and examine the outcome in terms of expected cost and long-term system dynamics.”
In this Web-exclusive interview, the Argonne team discusses how ABMS works; why it may be a more effective alternative for measuring ROI; and how using this model on the local level will help interoperability on a national level.
JHIM Online: Why are clinical- and decision-support applications more complex to analyze in terms of ROI?
Argonne: The extreme diversity of participants and their business practices in the healthcare market require an especially careful approach in analyzing future return on investment. Each stakeholder pursues their own goals; those goals will frequently conflict with those of other players. To complicate the issue, a stakeholder’s success in achieving goals may be strongly affected by “soft” system performance measures tied to clinical outcomes, such as quality of life and patient satisfaction; these can be very difficult to quantify. Clinical- and decision-support applications are designed to view how agents interact on several layers, including the electronic health record network, business, regulatory matters and in an overall system that may contain a fairly fine-scale representation of clinical processes and outcomes. The objective is to find the best compromise solution that would satisfy everyone with minimum investment upfront.
JHIM Online: Your article says that agent-based modeling and simulation is a new alternative to traditional modeling techniques. How was agent-based modeling and simulation developed? What are its benefits for measuring ROI, and how does it differ from traditional models
Argonne: Agent-based modeling and simulation has its historical roots in the study of complex adaptive systems (CAS). CAS was originally motivated by investigations into the adaptation and emergence of biological systems. ABMS provides a platform to study how both system- and individual-level patterns of behavior emerge from rules defined by the stakeholder. ABMS can also leverage traditional modeling techniques, such as embedded systems dynamics, statistical modeling, risk analysis, and optimization, in order to link micro-level behavior with macro-level effects that deduce or predict future events and behaviors.
ABMS builds upon traditional methods of simulation modeling by employing elements of discrete event modeling and statistical methodologies – even learning techniques such as genetic algorithms – but focuses on the interoperability of agent behaviors within the system. This allows for a much more flexible modeling paradigm where alternative courses of actions can be explored without having to encode every interconnection between the various subsystems of the simulation, such as social rules, government policies, insurance, and payment systems.
The ROI in ABMS systems stems from the ability to analyze alternative policies or implementations of policies within a flexible framework and to then intelligently choose the most promising path for the constraints involved.
Page: 1 2
