TABLE 3.

Adjusted model of anemia at follow-up among adolescent schoolgirls in the Northern and Volta regions of Ghana (n = 1307)

Variable	Adjusted OR (95% CI)
Cumulative number of IFA tablets consumed	0.95 (0.90–0.99)*
Cumulative number of IFA tablets consumed, squared1	1.00 (1.00–1.00)*
Baseline hemoglobin concentration, g/dL	0.33 (0.25–0.44)***
Age, years	1.17 (1.02–1.35)*
Wealth tertile
Middle vs. lowest	0.60 (0.34–1.05)
Highest vs. lowest	0.34 (0.17–0.67)**
Overweight vs. normal weight	1.21 (0.63–2.31)
Obesity vs. normal weight	0.78 (0.19–3.23)
Practicing vs. not practicing geophagy	1.32 (0.77–2.27)
Positive vs. negative malaria test	1.40 (0.59–3.31)
Rich source of heme iron vs. not consumed2	1.23 (0.72–2.09)
Fair source of heme iron vs. not consumed3	1.07 (0.55–2.09)
Rich source of nonheme iron vs. not consumed4	1.54 (0.87–2.73)
Foods and beverages fortified with iron vs. not consumed5	1.50 (0.93–2.40)
Citrus fruits vs. not consumed6	0.89 (0.50–1.59)

Estimates are weighted to be representative of all eligible girls in the school. Estimates, 95% CIs, and P values are calculated using multivariable logistic regression with Taylor series variance and account for clustering in school and intra-individual covariance. Follow-up occurred 9 months after baseline.

^*P < 0.05;

^**P < 0.01;

^***P < 0.001.

Abbreviation: IFA, iron and folic acid.

¹Model also accounted for cumulative IFA consumption, squared to address nonlinearity of the association.

²Red meats, such as beef, lamb, goat, or wild game and organ meats.

³Other animal-source foods, including eggs, poultry, and fish

⁴Dark green leafy vegetables, legumes, nuts, and seeds.

⁵Cereal and beverages fortified with iron Milo, Ovaltine, Cerelac, Yumvita, or Nido.

⁶Citrus fruits, including oranges, lemons, pineapple, mango, etc. This dietary variable was measured over the previous 24 hours.