Computers and Behavioral Assessment

Description of the Strategy

Computerized behavioral assessment tools include hardware and software applications that enable clinicians to monitor client behavior in ways that would have been impossible 20 years ago. Applications include functional interviews, behavioral questionnaires, clinical case modeling tools, self-monitoring programs, behavior-rating scales and checklists, psychophysiological assessment, and observational systems that utilize symbolic or virtual reality environments. Delivery platforms range from handheld computers (i.e., personal digital assistants, or PDAs) to desktop applications to immersive virtual worlds.

Although the term computerized behavioral assessment refers to a broad class of software and hardware applications, the technology rests on a functional foundation. Two variables are functionally related if changes in one variable are associated with, but do not necessarily cause, changes in a second variable. For example, a client's anger outbursts may be exacerbated when she consumes alcohol. Thus, quantity of alcohol consumed may be functionally related to frequency and intensity of the client's anger. One would expect reduction in alcohol consumption to have therapeutically desirable effects on anger modulation.

Computerized behavioral assessment applications distinguish themselves by facilitating assessment of functional relations. For example, a method that has gained increasing acceptance involves use of handheld computers to conduct ecological momentary assessment (EMA). In an EMA assessment strategy, clients use handheld computers to record in their natural environment frequency or intensity of selected target behaviors (e.g., anger intensity) and variables hypothesized to be functionally related to the target behavior (e.g., alcohol consumption). Recording sessions are usually brief and conducted multiple times per day. Advantages to the EMA approach include the following: Handheld computers are ubiquitous and unobtrusive, data are stored in a file for easy analysis, the computer can prompt the client for recording sessions, and feedback can be automatically generated. With regard to the last point, some researchers have gone so far as to develop cognitive behavior treatment modules that are delivered by handheld computer. However, researchers have noted drawbacks to the EMA approach as well: Clients fatigue over time and may stop entering data, behaviors that occur at high rates may be difficult to code or self-monitor, the value of the data depends upon a priori specification of variables hypothesized to be functionally related, and handheld computers are expensive to replace if lost or stolen.

Computers can also be used to conduct behavioral interviews. Researchers have developed programs to identify client problems and track dimensions of those problems over time. Other researchers have developed structured interviews that are symptom based and focus on diagnostic categorization rather than functional relations. In both cases, the computer is used to collect client information and generate a report that can be used by the clinician to plan and monitor treatment.

Most clients respond well to computerized interviews. While computerized interviews facilitate easy data entry, evidence has accumulated that clients may be more likely to divulge sensitive or embarrassing information to a computer than to a human interviewer. For example, several studies have shown that clients often prefer computerized interviews to human interviews.

In addition, computerized interviews are more reliable in delivering question prompts. Studies have shown that human interviewers may omit 5% or more of questions in a structured interview. However, critics have pointed out that such efficiency is not without cost. For example, a computer cannot monitor nonverbal client behavior, detect subtle inconsistencies in client reports, and can employ only rudimentary data analytic techniques in place of the inferences seasoned clinicians may make.

...