Background: The potential impact of participation in a study on participant behavior (even in the control group) has been long recognized. The so-called Hawthorne effect may reduce the observed impact of interventions to increase adherence with medical recommendations. Methods: A randomized trial of methods to encourage completion of treatment for latent tuberculosis infection among adolescents was conducted at two sites in Los Angeles County. Charts were reviewed to determine completion of care for patients for a 15- month period prior to the randomized trial initiation (n=482), for trial control group participants (n=195), and for patients for a 6-month period after the trial (n=148). Demographic variables were also abstracted. Data were analyzed to determine if completion of care was related to study participation. Results: In bivariate analyses, treatment completion was related to study site, age, race/ethnicity, and study participation; 22% of study control group participants failed to complete treatment compared to 31% of patients who initiated therapy before or after the trial (non-participants). After controlling for study site and demographic variables in multivariate logistic regression analysis, the odds ratio for failure to complete therapy was 1.7 (95% CI 1.1-2.5) for the non-participants compared to the study participants. Conclusion: In analyzing the impact of adherence interventions, it is important to consider the effect of study participation on participant behavior. It is sometimes possible to construct comparison groups to investigate this potential effect on specific studies.
Learning Objectives: At the conclusion of the session, the participant (learner) in this session will be able to: 1) discuss the potential impact of study control group participation on outcome measures, and 2) identify potential methodological procedures to investigate this impact.
Keywords: Adherence, Behavioral Research
Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: None
Disclosure not received
Relationship: Not Received.