Skip to Content
Interactive Textbook on Clinical Symptom Research Logo


Home button

Trial Design: Pain Sections
Author Bio
Introduction
Currently selected section: Placebo Effects
Single Dose Trials
Repeated Dose Trials
Explanatory Versus Pragmatic
Dose-Response
Parallel Group Versus Crossover
Conclusion
 

 

Chapter 1: Clinical Trials of Pain Treatment: Placebo Effects: Problem 2.2
 
  What is your interpretation?

You answered:

Selection C   The study should be replicated with the inclusion of a questionnaire to examine whether patients and research staff can guess what the treatment assignment is.

PERHAPS. THIS IS CONTROVERSIAL

Many experts have recommended that patients and research staff be asked to guess the identity of treatments as a test of blinding (Moscucci et al., 1987). Few investigators have done this, however, perhaps because of fears that a finding that a study is incompletely blinded will be used to discredit their positive results. In addition, an accurate test of blinding is hard to obtain. For example, patients may correctly guess whether they are active treatment or placebo by noting the presence or absence of therapeutic response. This would occur in a perfectly well blinded study, and it would not be fair to discount a positive study because patients identified the medication because of therapeutic effects, not side effects.

Moscucci et al., 1987 suggested that one ask patients whether their guess is based on the occurrence of therapeutic effects or side effects. However, this distinction is difficult in pain studies because with most analgesics, pain relief and side effects begin at the same time. Furthermore, post hoc explanations of why a person made a certain guess are notoriously unreliable. Even if the patient says he guessed "active medication" because of side effects, the answer may have been influenced by pain relief. Shlay et al., 1998 suggested a method to partly protect against this bias by stratifying patients' correct and incorrect guesses of treatment according to the grade of relief they reported on a category scale, followed by a contingency table analysis to test whether correct guesses were more common than chance.

We suggest that you click here to return to the problem and select each of the four possible answers. Here and in many of the following problems, new material is presented in each of our responses.