Table 3.
Opportunities to advance the potential of multilevel interventions
| Design |
| Greater attention to identifying, conceptualizing, and reporting important contextual factors across levels and over time |
| More use of dynamic, adaptive, emergent rapid learning designs that evolve and learn over time, rather than static designs |
| Greater attention to the interfaces across and among levels |
| Being explicit in all research about the potential effects of which levels are being studied and which levels might influence the phenomenon under study, even if they are not the focus of the study |
| More focus on effectiveness trials in real-world practical (typical) settings to inform practice and policy, and on more transparent reporting of adjustments |
| Conceptually based strategic interventions using pragmatic designs that inform practice and policy |
| Analysis |
| Reconceptualize reliability, sustainability, and fidelity to allow for intervention evolution and local and temporal adaptation |
| Capture, but also move beyond, only measures of central tendency and study of subgroups including biologic, economic, and environmental factors |
| Use of multimethod approaches that integrate quantitative modeling across multiple levels where relevant qualitative data can be generated and qualitative methods to evaluate levels with small numbers and to identify specific interlevel processes that are important for the outcomes of interest |
| Complex systems and dynamic simulation modeling may provide additional insights where data are sparse |
| Reporting on unintended consequences, factors emergent during studies, and how they were addressed |
| More thorough and transparent reporting of resources and costs expended, including program promotion and supervision, and sensitivity analyses to estimate the impact of variations in setting, staff, patients, etc. |
| Translation |
| More transparent reporting of setting, site, and clinician selection and representativeness; context and range of application are needed |
| Moving beyond fidelity to interventions that are locally adapted, evolutionary, and participatory, and publication of implementation lessons learned and guidelines |
| Study of emergent properties, as well as identification of how multiple levels interact in context rather than in isolation; we need to move beyond conceptualizations of “maintenance” to evolvable and capacity enhancing interventions |
| Funders, review groups, and training programs need to expand beyond the currently dominant reductionistic approaches; instead, we must value and support transdisciplinary methods, theories, and empiric approaches needed to conduct research across the multiple levels affecting health care and health |