tive components of interventions. Many of the appraisals have a narrow focus (e.g., schools, local environments) and fail to recognize the larger systems context in which the interventions were implemented. This lack of focus on contextual considerations (such as feasibility [including legal and political aspects], sustainability, effects on equity, potential side effects, acceptability to stakeholders, and cost-effectiveness) and other dimensions of generalizability (such as the intervention reach and the representiveness of participants, implementation and adaptation of the intervention, and its maintenance and institutionalization) limits the usefulness of the appraisals to inform decision making in diverse settings. Furthermore, the challenges that arise when interventions are undertaken in the real world lead the conclusions drawn to be negative or obsolete. In fact, many appraisals exclude comprehensive multilevel, multisector interventions (the most promising type of interventions for addressing obesity, as discussed in Chapter 2) because the specific effects on individuals cannot be disaggregated, and the results are considered “messy.” When such interventions target multiple behaviors simultaneously, it is often difficult to disentangle their impacts. Although such trials yield some positive results, the many gaps in the evidence suggest that obesity prevention lacks the vast array of effective interventions that are available for many other public health issues (e.g., tobacco control, immunization). While efforts are under way to fill these gaps, the gravity of the obesity epidemic calls for immediate action, suggesting that, as noted in an earlier IOM report, decision makers should rely on the best available evidence, instead of waiting for the best possible evidence (IOM, 2005).
The choice of outcomes on which to focus affects conclusions about the effectiveness of interventions across this set of appraisals. The outcome measures for selecting which individual studies to include in the appraisals were similar, and in most cases were cause for excluding a large percentage of studies initially identified for consideration. As stated in Chapter 2, the ultimate objectives of obesity prevention efforts are lowering the mean BMI level and decreasing the rate at which people enter the upper end of the BMI distribution. Accordingly, most individual-level studies focus on distal measures of weight status (e.g., percent overweight or obese, BMI or BMI z-score, body fat percentage), risk factors, or disease presence rather than the more proximal psychosocial and behavioral outcomes and the pathways that can alter the conditions in which eating, physical activity, and weight control occur (see Figure 2-2 in Chapter 2). Moreover, the period of follow-up in many of the studies included in the appraisals may not have been long enough to detect significant, sustainable changes in weight status. Indeed, the comprehensive changes called for to combat the obesity epidemic—changes in the physical environment, social norms, cultural practices, and policy—often require years to accomplish and are not always tracked or funded at adequate levels for sustained periods. In the appraisals that specify duration of study or length of follow-up as an inclusion criterion, the vast majority required a period of