|
Session Title: Unintended Interventions
|
|
Panel Session 835 to be held in International Room on Saturday, November 10, 1:50 PM to 3:20 PM
|
|
Sponsored by the Quantitative Methods: Theory and Design TIG
|
| Chair(s): |
| Melinda Davis,
University of Arizona,
mfd@u.arizona.edu
|
| Abstract:
Evaluators measure the strengths and limitations of programs and policies using a broad array of methods. However, even the best designed investigation can go awry, and study protocols can result in surprising and unintended effects. It is these unintended effects that can inform future research. Non-specific effects of treatment are usually treated as nuisance variables, to be eliminated or at least controlled. However, they can be a rich source of new interventions. A 'failed' study may not be a failure at all, if it identifies a new approach for a difficult problem. Vignettes will be presented from a variety of studies; the Consent Form as a potent treatment, useful mistakes in randomization, assessment as intervention, and the unexpected effect of a seemingly minor part of the study protocol. Each demonstrates a novel way to learn from evaluation results; effective interventions may be hidden in the non-specific effects of treatment.
|
|
Non-specific Effects of Treatment: Vignettes
|
| Melinda Davis,
University of Arizona,
mfd@u.arizona.edu
|
| Souraya Sidani,
Ryerson University,
s.sidani@utoronto.ca
|
|
Discovering serious design flaws at the end of an evaluation is an experience that new and seasoned evaluators share. Every so often, there is a non-specific effect of treatment that may be greater than the intended study intervention. In and of themselves, study protocols can have surprising results, and these non-specific effects of treatment can be a rich source of new interventions. This presentation will provide a series of vignettes, including the Consent Form as a potent treatment, and useful mistakes in randomization. Non-specific effects of treatment are often treated as nuisance variables, to be eliminated or at least controlled. Evaluators try to minimize problems arising in recruiting, the composition of the comparison group, and the nature of the placebo condition, study assessments, retention, and follow-up procedures. However, we may be able to learn from these unwanted effects. A 'failed' study may not be a failure at all, if it identifies a new approach for a difficult condition.
|
|
|
Non-specific Effects of Treatment: Assessments
|
| Andrea Chambers,
University of Arizona,
aschambers@virginia.edu
|
| Melinda Davis,
University of Arizona,
mfd@u.arizona.edu
|
| John Mark,
Stanford University,
jmark@stanford.edu
|
|
Asthma symptoms may be related to panic or fear, and children with asthma are at special risk for problems in psychological functioning. We tested the effectiveness of two brief behavioral treatments to reduce anxiety in children with moderate asthma. We hypothesized that the treatments would reduce anxiety, and help maintain asthma control while tapering corticosteroids. While there were no significant effects of the treatments; all groups significantly lowered their use of steroids without compromising their asthma symptoms. We hypothesize that the decreases in steroid use were due to a non-specific effect of treatment; the assessment protocol. The effective ingredient appeared to be the very brief monthly visits by a pediatrician to assess each child's pulmonary function. Inhaled corticosteroids carry significant health risks, and brief pulmonary evaluations may be an effective method to reduce their use.
| |
|
Non-specific Effects of Treatment: Biomarkers
|
| Melinda Davis,
University of Arizona,
mfd@u.arizona.edu
|
| Dan Shapiro,
University of Arizona,
shapiro@u.arizona.edu
|
|
At any point in time, the majority of smokers are not actively planning to quit, and most will not make quit attempts without some sort of treatment or prompting. We tested the effectiveness of two brief counseling interventions for smoking cessation in smokers who were not ready to quit. We did not find either treatment to have the advantage. However, there was a non-specific effect of treatment. The study protocol included spirometry, and participants who learned that they had marked decreases in their lung capacity were more likely to reduce their smoking. While spirometry was not the focus of our study, the results are consistent with effects of catastrophic and life changing health events. Biomarkers are a useful, although underutilized technique to encourage smoking cessation.
| |