Nonmedical influences on medical decision making: an experimental technique using videotapes, factorial design, and survey sampling

Health Services Research, August, 1997 by Henry A. Feldman, John B. McKinlay, Deborah A. Potter, Karen M. Freund, Risa B. Burns, Mark A. Moskowitz, Linda E. Kasten

[TABULAR DATA FOR TABLE 1 OMITTED]

In the second videotaped scenario ("postdiagnosis"), the consulting physician sees a different patient, who has been referred for a second opinion after a positive biopsy that has established the presence of a 0.8-cm infiltrating carcinoma, with clean margins on the specimen and equivocal hormone receptors. Staging information, if requested by the viewer, indicates that 2 of 29 tested nodes are positive for tumor, and all metastatic evaluation is negative. The patient thus has Stage IIA disease, the area of least consensus among practitioners with regard to the appropriate choice of primary therapy or the need for adjuvant therapy.

The script for each scenario was developed by the three clinical investigators from Boston University Medical Center (BUMC) and was based on cases provided by two experienced clinicians. The dialogue was reviewed for authenticity by a panel of practicing physicians and edited for continuity by a professional video writer-producer. The scenes were rehearsed and performed by professional actors and videotaped under professional direction in an actual hospital office. Strict quality control procedures were followed during taping to ensure that the script was followed faithfully and that each actor playing the patient maintained her assigned characteristics, including both verbal and nonverbal behavior. Variant sequences within each scenario (see further on) were executed in a standardized fashion for later editing into the tape. One male actor played the consulting doctor in all variant scenarios and remained largely off-camera.

Patient Characteristics. The patient was portrayed by eight different female actors, each presenting a certain combination of age, race, and socioeconomic status (SES). Age was set at either 65 or 80 years, as listed in a character synopsis at the beginning of the videotape and enacted by an actor of approximately that age. Race was represented by casting an actor from the appropriate group. SES was expressed visually in the style of dress and verbally by minor grammatical variants of the script. SES was also implied by the specification of insurance coverage (either Medicaid or Medex/Medicare) in text form at the beginning of the tape. With all three characteristics thus dichotomized, the possible combinations numbered [2.sup.3] = 8, and each combination was assigned to a different actor-patient. Age-appropriate pseudonyms were assigned to the patients for convenience in presentation (Table 2).

Table 2: Patients Portrayed on Videotape: Half-Factorial Design

Actor(*)      Age     Race       SES      Scenarios     Variants

Alice         65      White     Upper     Ere, Post        A
Annie         65      White     Lower     Ere, Post        B
Bess          80      White     Upper     Ere, Post        B
Blanche       80      White     Lower     Pre, Post        A

Edith         65      Black     Upper     Pre, Post        B
Henrietta     65      Black     Lower     Pre, Post        A
Lenore        80      Black     Upper     Pre, Post        A
Reva          80      Black     Lower     Ere, Post        B

* Pseudonym.