Article #30
Newsletter list 


Previous Article 

Next Article 


Scripts as an Alternative Framework for Describing Medical Decision Making

Robert Hamm
Oklahoma City, OK

This year we had a paper published in the volume on medical decision making in the Cambridge JDM series: Hamm, R.M., Scheid, D.C., Smith, W.R., and Tape, T.G. (2000). Opportunities for applying psychological theory to improve medical decision making: Two case histories. In G.B. Chapman and F. Sonnenberg (Eds.), Decision Making in Health Care: Theory, Psychology, and Applications (pp. 386-421). New York: Cambridge University Press.

The first section reviews the various efforts that have been made to apply decision theoretic ideas or research products in order to improve medical practice. It describes projects that have applied decision analysis to individual patients, made evidence-based guidelines available to physicians, trained physicians to reason with analytical principles, provided decision-relevant information to physicians, or provided computerized decision aids.

The next section presents two interventions in detail, and analyzes the reasons they did not accomplish their expected goals. The first intervention was an attempt to make physicians reduce the proportion of sore throat patients to whom they prescribe antibiotics by training them to judge more accurately the probability each patient has a sore throat due to streptococcus. Their probability judgments become more accurate (and lower), but they still prescribed antibiotics to the same proportion of patients. Perhaps the assumption "Decisions are suboptimal because judgments are suboptimal" is not correct. Or maybe a judgment of diagnostic probability did not actually control the physicians' prescribing behavior.

The second intervention attempted to make physicians start talking about "end of life" issues with hospital patients and patients' families soon enough that the families could decide not to have "everything possible" done to revive the patient as he or she died. A central element of the study's intervention was to provide accurate (model based) estimates of the probability that the patient would die during this hospital admission. The intervention had essentially no effect on the use of cardiopulmonary resuscitation (CPR). The study assumed that if physicians were given information that decision theory says is relevant for decision making (survival probabilities with and without CPR, and patient preferences), they would use it. Perhaps the typical physician's end-of-life decision making strategy makes little use of such information.

The final section presents a descriptive metaphor: that physicians follow scripts when they make decisions about patients. We argue that interventions will be more successful if they are based on a clear understanding of the scripts physicians currently use, and if physicians are explicit about the alternative scripts that are proposed to improve their decision making.

For the Brunswik community, we would note that the systematic framework that we use for studying judgments or teaching judgment strategies (e.g., "What is the probability that a patient's sore throat is due to streptococcus?" as a function of N features) can as easily be applied for studying action strategies (e.g., "What would you do for this patient?" as a function of N features; or, "What is the probability that you would do X?" as a function of N features).

Contact Robert Hamm

Previous Article                Next Article

Newsletter list