Mixed Method Approaches to Understanding Cancer-Related Dietary Risk Reduction among Public Housing Residents

Improving diet is one important pathway for addressing cancer disparities. We conducted mixed-method analyses of 468 24-h dietary recalls from 156 African-American women residents of Washington DC public housing to better understand dietary patterns. Recalls were rated for five cancer-related preventive characteristics (adequate fruits/vegetables, moderate fat, moderate calories, no alcohol, and adequate Healthy Eating Index score), combined as a scale. Bivariate and multivariate analyses identified psychosocial and dietary characteristics associated with scale scores. Qualitative analyses of dietary records identified contextual aspects of food patterns within and across score groups. Sixty-one percent of respondents met zero or one dietary goal; alcohol abstention was most common (64%). Only 12% achieve either three (6%), four (4%), or all five (< 1%) goals; five fruit and vegetable servings were least common (15%). The underlying scalar structure of responses suggests that fruit and vegetable consumption is seldom achieved in this population without other scale components. Poorer scores were associated with younger age, depressive symptoms, stressful life events, smoking, and food-purchasing practices. Qualitative analyses identified eight themes related to differences between dietary patterns. Findings reinforce the value of nonreductionist approaches to cancer-related nutrition intervention.