A Focus on Simulation in Medical Informatics

This special issue of JAMIA addresses simulation, a methodology that holds great promise for the advancement of information technology in health care. Simulation has been termed a third domain, complementing both natural language and mathematical/statistical approaches. The process of constructing a simulation model introduces a new way of thinking about medical informatics. It also provides tools to build understanding and generate insights into complex systems and processes. Simulation has many potential applications in medical informatics including research, evaluation, education, training and policy making.

A number of developments, including the Next-Generation Internet, high-bandwidth communication, object-oriented software design, distributed and parallel processing, and scientific visualization techniques, have greatly enhanced the power and expressiveness of simulation. The Internet permits world-wide access to resources not widely available for medical education. Simulation technology that utilizes high performance computers and graphics permit realistic real-time displays of virtual body structures that can be used to teach anatomy and surgery among others.

Three papers in this issue report on the development of simulation-based learning environments. Temkin et al.¹ describe a virtual-reality-based anatomical training system for teaching human anatomy over the Internet. The system’s interface allows the user to navigate through the Visible Human and to touch geometric structures with a haptic device. A second paper by Dev et al.² describes simulated medical environments that can be used to support the teaching of anatomy and basic surgical skills over the Internet. The anatomical learning environment provides learning resources for hand anatomy including a three dimensional model of the hand that can be rotated and viewed in stereo at different depths of dissection. The surgical environment provides a model of pelvic anatomy and surgical tools that can be used to carry out a set of basic maneuvers including probing, cutting and suturing.

Pugh et al.³ report on the development and evaluation of performance assessment measures for a physical examination simulator used in medical education. They utilized sensors and data acquisition technology to capture performance data from mannequins used to teach medical students to perform simulated clinical female pelvic examinations. The study derives a number of new performance measures and shows that these measures provide objective, reliable and valid measures that can be used to assess physical examination techniques using the pelvic examination simulator.

In many instances, the application of an objectivist approach, such as randomized control trials, is difficult to adapt to research and evaluation in medical informatics.⁴ In real practice settings information systems are complex involving not only information technology but the organizational environment as well. The system under investigation is dynamic and changes over time. Seldom can the investigator employ randomization and control many aspects of the environment in which the information technology has been implemented. Moreover there are relatively few studies over time of processes involving information technology. One reason is the inherent difficulty of conceptualizing these processes, but another is that tools for doing so have been limited.

One approach that provides flexibility is computer simulation.⁵ First, a model that is an abstraction of the real system under investigation is developed. The act of developing a model and translating it into a simulation requires making system components and their relationships explicit. Using simulation, an investigator can express ideas about the structure of complex systems and their behavior in a precise way. Simulation can be used even in situations where the behavior of a system can be observed but the exact processes that generate the observed behavior are not fully understood. Second, models that represent a health care system or components of the system can be manipulated without disrupting the real system or practice setting. Once validated, the model yields accurate estimates of the behavior of the real system over time. The effects of variations in system inputs, different initial conditions, and change in the structure of the system can be observed and compared.⁶

For example, modeling and simulating complex molecular pathways can provide a better understanding of physiological as well as pathological processes. Tsavachidou et al. ⁷ simulate the underlying molecular process involved in the menstrual cycle both pre- and post-menopause. Their study has implications for developing estimates of the individualized risk of post-menopausal disorders such as breast and ovarian cancer, cardiovascular disease, and osteoporosis.

Three papers demonstrate the use of simulation to evaluate information technology applications. The first paper⁸ reports an evaluation of a web environment developed to support research by an academic internal medicine faculty. In a usability laboratory experiment, simulated scenarios were presented to participants in order to evaluate the ability of the Web environment to improve the ability of clinicians to perform clinical research. Overall users were able to locate the appropriate information resources over 80% of the time.

The second paper reports a simulation model that represents the medication delivery system in a hospital.⁹ The model is used to estimate medication errors that result in adverse drug events (ADEs) and related additional days of hospitalization. Five separate interventions are evaluated to determine their potential ability to reduce ADEs and related hospital costs. The results suggest that prevention efforts that focus on a single stage of the process have limited impact on the overall medication error rate. This kind of insight is important because the successful introduction of information technology depends upon demonstrating its cost effectiveness in real practice settings.

The growth of network applications in informatics has created not only a demand for network bandwidth but a demand for consistent, reliable performance as well. Simulation provides a tool that can be used to generate synthetic internet traffic and to test different network protocols. Shiffman et al.¹⁰ report a simulation-based study of quality of service related to the Next Generation Internet using PathMaster, a prototype image database application that takes cell images from cytology specimens and compares them to images in a database. They developed a simulated Internet and used it to evaluate several strategies for managing congestion and shaping traffic.

The message of this special issue is that simulation provides a powerful tool for the medical informatics community. The new generation of simulation software incorporates graphical interfaces. This facilitates the use of the software to model complex systems and processes. Investigators, in general, no longer have to deal with complex mathematical expressions and programming languages. These models provide a powerful means of exploring model assumptions, the effects of structural changes and their effects on the dynamic behavior of the system and performance measures. The application of simulation to medical informatics has profound implications for the way we conceptualize information systems and processes, the kinds of questions we ask, the types of data we collect and the conclusions we draw. Simulation provides tools that can inform practical actions as well as contribute to policy and theory in informatics. In the future we will see more simulation-based studies and a widening of the scope of simulation research in medical informatics.