Screening Proprietary Drug Names for Similarities: Research Design and Questionnaire Structure .

1
Screening Proprietary Drug Names for
SimilaritiesResearch Design and Questionnaire
Structure.

• Shari Diamond, J.D., Ph.D.
• Northwestern University
• June 26, 2003

2
Challenges to Test Design

• Products are not yet on the market
• Difficulty of simulating conditions under which
  prescriptions are
• written
• delivered orally
• filled by pharmacists and hospital personnel

3
Expert Panels

• Knowledgeable about currently marketed drugs
• Familiar with drug pairs that have generated
  errors
• Can use source lists to generate potential
  candidates with confusing name similarity
• Tacit knowledge

4
But Limits on Ability of Experts to Predict Errors

• May generate similars that dont pose a threat
• May miss potential errors because
• E.g., may fail to generate mispronunciations that
  cause error
• E.g., may not anticipate similarities generated
  by handwriting

5
Need to Test Expert Predictions

• Individuals are often bad predictors of their own
  (or others) reactions
• Testing phase gauging actual reactions is crucial
• Need sample drawn from relevant population
• Responding to appropriate stimuli

6
Assume We Want to Test the Name Taxol

• Respondents are told they will see a series of
  drug names, one at a time.
• Some will be drugs currently on the market, and
  some will be drugs not yet available.

7
Procedures for Testing

• Respondents will receive a set of handwritten
  drug names, one at a time, including (but not
  limited to) Taxol
• Self-administration is possible if respondents
  are hooked up to the Internet
• Can be timed exposure to reflect usual time spent
  in examining a prescription
• Order of presentation of a series of names can be
  rotated

8
Instructions

• You will be shown the names of several drugs, one
  at a time.
• Some of these drugs may be currently on the
  market and some may not be.
• For each drug, the name of the drug will be
  followed by several questions.
• These questions will ask about your reactions to
  the drug name you just saw.

9
Questions Asked After Each Name Is Shown

• Please type in the name of the drug you just saw.
• (What is the name of the drug?)
• Have you seen this name before today?
• If yes, do you happen to recall what condition(s)
  it is used to treat?

10
Other Cues

• Testing name alone maximizes likelihood of name
  confusion
• Other cues (e.g., dosage, directions for use) can
  reduce it
• In fact, best prevention of error is to provide
  multiple cues (e.g., both brand and generic
  names)
• Including cues in screening tests may reduce
  apparent likelihood of error, but wont reflect
  reality if cues are inconsistently provided

11
A Potential Approach When the Expert Panel
Identifies A Particular Similarly Named Drug

• The respondent is shown the new drug.
• The respondent then views a line-up of
  pharmaceutical products (or a picture of them)
  and is asked

12
Line-up Instructions

• The drug whose name you were just shown may or
  may not be displayed here.
• Please indicate whether or not it is in the
  display.
• If it is, please indicate which number it is.

13
Caveats With The Line-up Approach

• Despite instruction, need to control for guessing
• Reserve for
• situations where similarly named drugs are likely
  to be stored side by side

14
When Should the Questions Be Closed-Ended? IT
DEPENDS

• Not when testing for comprehension/recall of name
  
• Potentially in providing list of condition(s)
  the drug may be used to treat
• Line-up is essentially a multiple-choice question
  a recognition task

15
Focus Groups as a Substitute?

• Good for generating ideas (the expert panel)
• Weak for evaluating individual reactions to
  specific stimuli
• Interdependence of responses from group members
  Low N
• Crucial role of moderator

16
Problems in Validating

• One-sided and potentially incomplete feedback
• Are approvals followed or not followed by
  reported medical errors?
• But disapprovals never tested

17
Future

• Computerized communication (no handwriting
  problems)
• Bar codes to permit computer reading of
  prescriptions
• In the present,
• need to proceed with caution