We thank Issenberg and colleagues for highlighting the potential of computer simulation to assess clinical competence. We focused on the most common methods used in residency training programs currently, but we agree that computer simulations may become a widespread method to assess the clinical competence of residents. A recent survey of medical schools found that 73 of 124 schools are already using some form of computer simulation for evaluation (1).

Advantages of computer-based methods include uniformity in assessment; opportunity to assess skills without possible harm to live patients; exposure to specific and important medical content; and ability to provide immediate feedback to the learner, assess decision making over time, and design cases suitable to the different levels of trainees (1, 2). Advances in microcomputer technology have also greatly reduced the cost of computer simulation by eliminating the need for the mainframe computers that were necessary for initial simulation efforts (2).

However, despite our own interest in computer simulation, this "new" method has several important limitations as an assessment tool. Despite years of intensive work on computer-based simulation at the National Board of Medical Examiners, Melnick (2) noted that he still could not convincingly define which distinct attributes of competence are assessed by this method (3). Even with the explosion of multimedia technology, computer simulations are an incomplete representation of a total patient and cannot capture all the nuances of both the verbal and nonverbal aspects of a complex live patient encounter (3). Furthermore, because reliability and validity have not been convincingly established, use of computer simulations for summative evaluations of residents' clinical skills requires further investigation (4).

Finally, we wish to re-emphasize the need for a multidimensional approach. As the use of "patient surrogate" measures of competence, such as standardized patients and computer simulations, become more available, affordable, and accepted, we are concerned that medical educators will substitute these tools for the more traditional method of judging competence: observing residents interacting with patients. The growing financial pressures placed on today's teaching faculty may accelerate this process. We agree that computer simulation can be an integral part of an institution's multifaceted program to evaluate the competence of its residents, but computer assessment methods should augment, not substitute for, the important evaluation and feedback provided by seasoned faculty to trainees.