Possible Problems or Issues That Could Affect Intervention Studies
Intervention research is important in clinical practice because it improves care by implementing evidence-based programs. It provides information on practices, which improve medical outcomes for patients and families. According to Grove, Burns, and Gray (2013), the problems that affect intervention studies relate to internal and external validity. Those aspects are threats to the integrity of intervention research. Challenges to external validity involve the particular nature of participants and study settings that affect generalizability. For example, a study carried out in a pediatric ICU in one hospital might be highly specific to the setting and the participants and might be difficult to generalize to other settings. Issues relating to construct validity emanate from the possibility that measurement properties differ from one setting to another, which affects the validity of inferences regarding the relationship between the intervention (treatment) and results (outcome). Inferences become completely or partially wrong because of threats to validity.
In randomized controlled trials, internal validity is affected by various issues, including changes in personnel at the research setting, such as a hospital or school, mobility of participants (patients), community mistrust of the study, and mastery testing rescheduling. Some events, including major changes in the course of the treatment, could affect the outcome. Other issues that might affect the outcome of an intervention study include the possible bias in assignment to intervention and control groups as well as inadequate blinding. Various internal validity threats are likely to impact on the inferential quality, but the most likely include “ambiguous, temporal precedence, history, attrition, treatment fidelity, and compensatory equalization of treatments” (Buckwalter et al., 2009, p. 112). Therefore, researchers in intervention studies should be aware of the issues and design them in a way that does not affect the validity of the outcome.
