There are also various alternatives for the location in which an intervention is delivered. Location may include community-based settings, the home, a clinical setting, a residential facility, the workplace, or a research laboratory. Each of these settings has a unique set of characteristics that warrant consideration. In a research laboratory, the setting is much more controlled with respect to extraneous influences (e.g., pets, other family members), and thus one has more confidence about the internal validity of the study. However, it is artificial and there may be logistical problems for the participants in terms of travel, which may impact on recruitment and retention. Participants may also be uncomfortable in laboratory settings. Conducting interventions in home settings may be more comfortable, less stigmatizing if the intervention concerns a highly sensitive matter (e.g., depression, HIV prevention, sexual practices), and easier for study participants. However, this can also create logistical problems for the interventionist and add to trial costs (e.g., travel time of interventionists and travel costs). Home contexts also vary considerably in terms of factors such as clutter, cleanliness, pets, other family members, and safety that can impact on the delivery of the intervention. These factors can create unique challenges, especially if the intervention involves the use of some form of technology (Gitlin,
2003). In the PRISM trial, it was sometimes challenging to install computer systems in homes of participants owing to clutter and space constraints.
A general benefit of community settings is that they represent the context in which the intervention will be actualized and are familiar to participants. However, potential threats to both internal and external validity need to be carefully considered. In health care settings, the representativeness of patients and providers and settings is important (Glasgow, Bull, Gillette, Klesges, & Dzewaltonwski, 2002). Threats to internal validity include potential contamination (e.g., people in the same nursing home assigned to different treatments); confounding factors such as unanticipated changes in the community (e.g., weather events) or changes in organizational policy, missions, or practices; and issues with treatment fidelity. If a health care provider in a busy clinic is responsible for implementing the intervention or assessment protocols, he or she may take shortcuts or may not understand the importance of following standard procedures to assure the integrity of the delivery of an intervention. Thus strategies such as intervention manuals and booster-training sessions, to enhance treatment fidelity, need to be in place. CABs can also provide valuable input on how to best meet the challenges associated with implementing intervention programs in community settings. Community-based participatory research approaches are also helpful as they foster “buy-in” from the community and commitment to the project/intervention (see Melnyk & Morrison-Beedy, 2012 for more in-depth discussions of implementation in specific types of community settings).
The involvement of multiple sites at different institutions and locations poses a challenge to intervention delivery considerations. Expanding studies to more than one site is quite common in both efficacy and effectiveness trials (e.g., REACH I and II, Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE). For the most part, multisite trials involve separate research teams but common research protocols and data collection strategies. Sometimes recruitment strategies and possibly the way in which an intervention is delivered may need to vary at the research sites to accommodate differences in the characteristics of target populations. Multisite studies often involve a separate data coordination site to provide oversight regarding data collection and also the delivery of the intervention. For example, in the REACH trials, the University of Pittsburgh served in this role. In the PRISM project, the University of Miami Miller School of Medicine assumed this role.
Decisions to engage in multisite trials require careful consideration as to how the delivery characteristics of an intervention will be implemented and monitored. These types of trials add considerably to cost and logistical considerations. Other important issues include establishing protocols for the research teams, treatment fidelity, data collection, storage, and transfer. In addition, the mechanisms for communication among team members within and across sites need to be clearly articulated, as do protocols for decision making, conflict resolution, and dissemination and publication activities.