Implementation science studies the use of strategies to adapt and use evidence-based interventions in targeted settings (e.g., schools, workplaces, health care facilities, public health departments) to sustain improvements to population health. This nascent field of research is in the early stages of developing theories of implementation and evaluating the properties of measures. Stakeholder engagement, effectiveness studies, research synthesis, and mathematical modeling are some of the methods used by implementation scientists to identify strategies to embed evidence-based interventions in clinical and public health programs. However, for implementation science to reach its full potential to improve population health the existing paradigm for how scientists create evidence, prioritize publications, and synthesize research needs to shift toward greater stakeholder input and improved reporting on external validity. This shift will improve the relevance of the research that is produced and provide information that will help guide decision makers in their selection of research-tested interventions.

Objective: Diet is an important determinant of health outcomes, but physicians have few ways to identify persons with suboptimal diets. The purposes of this study were to examine the reproducibility of a short dietary assessment questionnaire (PrimeScreen) and to compare its results with those of a longer food frequency questionnaire and with plasma levels of selected nutrients. Design: Each subject completed two PrimeScreen questionnaires at an interval of 2 weeks and one full length, 131-item, semiquantitative food frequency questionnaire (SFFQ), and had a sample of blood drawn. We compared the PrimeScreen with two reference standards, the SFFQ and plasma levels of selected nutrients. Setting: A large managed care organization in New England. Subjects: A total of 160 men and women, aged 19±65 years, participated. Results: For foods and food groups, the mean correlation coefficient (r) was 0.70 for reproducibility and 0.61 for comparability with the SFFQ. For nutrients, the mean r was 0.74 for reproducibility and 0.60 for comparability with the SFFQ. No substantial differences were evident by sex, race, body mass index, occupation or education. Correlation coefficients for the comparison of vitamin E, b-carotene and lutein/ zeaxanthin intakes from the PrimeScreen with plasma levels were 0.33, 0.43 and 0.43, respectively. These values were similar to those comparing the SFFQ with plasma levels. The median time to complete PrimeScreen was 5 min; 87% of participants required fewer than 10 min. Conclusions: A quick way to assess quality of diet among adults, PrimeScreen has adequate reproducibility and its results compare well with a longer food frequency questionnaire and biomarkers.

The intervention led to significant improvements in health behaviors among a working class, multiethnic population, regardless of race/ ethnicity and socioeconomic status. Interventions that respond to the social context of working class individuals across racial/ethnic categories hold promise for improving cancer-related risk behaviors.