Clinical vs Actuarial Judgment

METHODS OF JUDGMENT AND MEANS OF COMPARISON

In the clinical method the decision-maker combines or processes information in his or her head. In the actuarial or statistical method the human judge is eliminated and conclusions rest solely on empirically established relations between data and the condition or event of interest. A life insurance agent uses the clinical method if data on risk factors are combined through personal judgment. The agent uses the actuarial method if data are entered into a formula, or tables and charts that contain empirical information relating these background data to life expectancy.

Clinical judgment should not be equated with a clinical setting or a clinical practitioner. A clinician in psychiatry or medicine may use the clinical or actuarial method. Conversely, the actuarial method should not be equated with automated decision rules alone. For example, computers can automate clinical judgments. The computer can be programmed to yield the description "dependency traits," just as the clinical judge would, whenever a certain response appears on a psychological test. To be truly actuarial, interpretations must be both automatic (that is, prespecified or routinized) and based on empirically established relations.

Virtually any type of data is amenable to actuarial interpretation. For example, interview observations can be coded quantitatively (patient appears withdrawn: [1] yes, [2] no). It is thereby possible to incorporate qualitative observations and quantitative data into the predictive mix. Actuarial output statements, or conclusions, can address virtually any type of diagnosis, description, or prediction of human interest. . . .

Although some research appeared on clinical and actuarial judgment before the mid-1950s, Meehl (1954) introduced the issue to a broad range of social scientists in 1954 and stimulated a flurry of studies. Meehl specified conditions for a fair comparison of the two methods.

First, both methods should base judgments on the same data. This condition does not require that clinical judgment and statistical method, before comparison, use the same data to derive decision strategies or rules. The clinician's development of interpretive strategies depends on prior experience and knowledge. The development of actuarial methods requires cases with known outcome. The clinical and actuarial strategies may thus be derived from separate or overlapping data bases, and one or the other may be based on more or fewer cases or more or less outcome information. For example the clinician may have interpreted 1,000 intelligence tests for indications of brain dysfunction and may know the outcome for some of these cases based on radiologic examination. The actuarial method may have been developed on the subset of these 1,000 cases for which outcome is known.

Second, one must avoid conditions that can artificially inflate the accuracy of actuarial methods. For example, the mathematical procedures (such as regression analysis or discriminant analysis) used to develop statistical actuarial decision rules may capitalize on chance (nonrepeating) relations among variables. Thus, derivation typically should be followed by cross-validation, that is, application of the decision rule to new or fresh cases, or by a standard statistical estimate of the probable outcome of cross-validation. Cross-validation counters artificial inflation in accuracy rates and allows one to determine, realistically, how the method performs. Such applications are essential because a procedure should be shown to work where it is needed, that is, in cases in which outcome is unknown. If the method is only intended for local use or in the setting in which it was developed, the investigator may partition a representative sample from that setting into derivation and cross-validation groups. If broader application is intended, then new cases should be representative of the potential settings and populations of interest.