Advertisement-Induced Prescription Drug Requests Patients' Anticipated Reactions to a Physician Who Refuses

Journal of Family Practice, June, 1999 by Robert A. Bell, Michael S. Wilkes, Richard L. Kravitz

METHODS

We conducted the survey after gaining approval from the University of California-Davis Human Subjects Review Committee. Oral informed consent was obtained from each subject before beginning an interview.

SAMPLE AND DESIGN

Our sample, drawn from Sacramento County, was generated using a standard random telephone survey strategy.[19] Specifically, an equal number of computer-generated 4-digit suffixes were attached to all of the exchanges (prefixes) in use in the target population. As expected, most of these numbers were unassigned or nonresidential numbers. When a number was found for a household, that number was used as a seed number to create a block of 9 additional numbers that were added to the number pool. For instance, if the number 234-5678 was for a household, the numbers 234-5670 through 234-5679 were called. This strategy has the property of being serf-weighting, because the proportion of interviews attempted for each exchange is proportional to the number of working residential numbers for that exchange.[20]

Interviews were conducted in the spring of 1998 by 5 undergraduate interviewers who had participated in an extensive training seminar. No attempt was made to conduct an interview if a phone number was for a business, government office, or a household where English was not spoken. The member of each household selected for the study was determined using the Hagen-Collier randomized respondent selection procedure.[21] With this approach, the target respondent in each household is randomly determined to be the youngest female, youngest male, oldest female, or oldest male. This individual was considered unreachable after 6 unsuccessful call attempts. The supervisor made random call-backs to validate calls. The survey completion rate was 69% for households for which contact with the eligible party was made.

The final sample was composed of 201 women and 128 men. The greater proportion of women (61%) reflects both a higher refusal rate among men and more difficulty in reaching male targets (54% of adults in the survey population are women). Approximately 77% of the sample was white. The age profile was as follows: 18 to 29 years = 21%; 30 to 39 years = 18%; 40 to 49 years = 22%; 50 to 59 years = 16%; 60 to 69 years = 11%; 70 years and older = 12%; and less than 1% declined to answer. Nineteen percent of respondents reported a yearly household income of less than $30,000; 25% were in the $30,000 to $44,999 range; 29% were in households making more than $45,000 to $59,999; 26% reported incomes exceeding $60,000; and slightly more than 1% declined to answer the income question. Approximately 58% of respondents had a high school education or less, and 42% were college graduates. At the time of the survey, 58% of respondents were taking at least 1 prescription drug, and 93% were covered by a health plan.

OUTCOME MEASURES

Respondents were instructed to imagine that they have asked their doctor to provide a prescription for a drug after seeing an advertisement for it, but the physician refused to provide the prescription. They were given descriptions of 4 possible responses to this refusal, and were asked to indicate how likely it was that they would experience or initiate each response. The possible responses were: (a) become disappointed in their physician (disappointment); (b) try to change their physician's mind by convincing the physician of their need for the drug (persuasion); (c) talk to a different physician about getting a prescription for the drug (prescription shopping); and (d) quit going to the physician and switch to a new physician (doctor switching). Three response categories were presented: not at all likely, somewhat likely, and very likely.