Format

Send to

Choose Destination
Psychopharmacol Bull. 2003 Spring;37(2):50-8.

What is a significant response in drug studies of attention-deficit/hyperactivity disorder: statistical significance is necessary, but is it sufficient?

Author information

1
Attention Disorders Institute, Troy, Michigan 48085, USA. sangalb@sbcglobal.net

Abstract

The methods used to report data of clinical treatment trials can influence the interpretation of the results. Reporting that a study drug achieves a statistically significant result does not provide the absolute likelihood and magnitude of benefits. Reporting results using an absolute definition of clinical expectation from a pharmacological intervention may be useful, and characterization of placebo response may help to frame such an absolute definition. The objective of this report was to characterize placebo response in patients with attention-deficit/hyperactivity disorder (ADHD). The design was a double-blind, placebo-controlled trial of a study drug that did not produce a statistically significantly different response from placebo in a multi-center trial. Data from placebo and study drug groups at a single site were collapsed into one group to evaluate response to placebo (or ineffective drug) in a double-blind fashion. Sixty-four children aged 6 to 12 years with a Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, diagnosis of ADHD (combined type), were enrolled; 57 completed the study. Subjects were randomized to receive either transdermal buspirone patches or matching placebo patches. Mean ADHD Rating Scale (ADHD-RS) score was 44.3 (SD 5.4) at baseline, and 37.4 (SD 10.9) after 6 weeks. Mean ratio of week 6 to baseline ADHD-RS was 0.84 (SD 0.22). Less than 5% of the sample had a 60% or greater decrease in ADHD-RS in response to placebo or ineffective medicine. A 60% decrease in the ADHD-RS seems to constitute one clear definition of clinical expectation from a pharmacological intervention, with minimal possibility of such a decrease occurring as a placebo effect.

PMID:
14566214
[Indexed for MEDLINE]

Supplemental Content

Loading ...
Support Center