Risk-Based Questionnaires Fail to Detect Adolescent Iron Deficiency and Anemia

Abstract

To evaluate the predictive ability of screening questionnaires to identify adolescent women at high-risk for iron deficiency or iron deficiency anemia who warrant objective laboratory testing. Cross-sectional study of 96 female individuals 12-21 years old seen at an academic medical center. Participants completed an iron deficiency risk assessment questionnaire including the 4 Bright Futures Adolescent Previsit Questionnaire anemia questions, along with depression, attention, food insecurity, and daytime sleepiness screens. Multiple linear regression controlling for age, race, and hormonal contraception use compared the predictive ability of 2 models for adolescent iron deficiency (defined as ferritin <12 mcg/L) and anemia (hemoglobin <12 g/dL). Model 1, the Bright Futures questions, was compared with model 2, which included the 4 aforementioned screens and body mass index percentile. Among participants, 18% (17/96) had iron deficiency and 5% (5/96) had iron deficiency anemia. Model 1 (Bright Futures) poorly predicted ferritin and hemoglobin values (R2 = 0.03 and 0.08, respectively). Model 2 demonstrated similarly poor predictive ability (R2 = 0.05 and 0.06, respectively). Mean differences for depressive symptoms (0.3, 95% CI -0.2, 0.8), attention difficulty (-0.1, 95% CI -0.5, 0.4), food insecurity (0.04, 95% CI -0.5, 0.6), daytime sleepiness (0.1, 95% CI -0.1, 0.3), and body mass index percentile (-0.04, 95% CI -0.3, 0.2) were not significantly associated with ferritin in model 2. Mean differences for hemoglobin were also nonsignificant. Risk-based surveys poorly predict objective measures of iron status using ferritin and hemoglobin. Next steps are to establish the optimal timing for objective assessment of adolescent iron deficiency and anemia.