Skip to main content
. 2020 Jan 31;150(5):1093–1099. doi: 10.1093/jn/nxaa016

TABLE 3.

Endline work efficiency of Rwandan women by treatment group (Fe-Bean or Control-Bean) or change in iron biomarker1

Outcome Predictor n LS means (95% CI) for group P
ITT analysis
 WE at 25 W Treatment group 97 Fe-Bean: 48.2% (44.0%, 52.4%); Control: 47.2% (43.2%, 51.3%) 0.74
 WE at 40 W Treatment group 119 Fe-Bean: 37.0% (34.4%, 39.5%); Control: 36.6% (34.1%, 39.0%) 0.82
Secondary analyses
 Full sample (n = 125)
  Outcome
   WE at 25 W Change Hb2 97 0.17 (−0.23, 0.58)3 0.41
   WE at 25 W Change Fer4 97 −0.29 (−0.75, 0.16)3 0.21
   WE at 40 W Change Hb2 119 0.24 (0.01, 0.48)3 0.044
   WE at 40 W Change Fer4 119 0.03 (−0.25, 0.32)3 0.83
 IDNA at baseline (n = 60)
  Outcome
   WE at 40 W Change Hb2 60 0.24 (−0.10, 0.58)3 0.16
   WE at 40 W Change Fer4 60 0.50 (0.06, 0.95)3 0.027
 IDA at baseline (n = 43)
  Outcome
   WE at 40 W Change Hb2 43 0.64 (0.05, 1.23)3 0.036
   WE at 40 W Change Fer4 43 0.10 (−0.76, 0.96)3 0.82
1

All models adjusted for baseline WE at 25 W or 40 W, as appropriate, baseline CRP, and endline CRP. Control-Bean, intervention arm receiving conventional beans; CRP, C-reactive protein; Fe-Bean, intervention arm receiving iron-biofortified beans; Fer, serum ferritin; Hb, hemoglobin; IDA, iron-depleted anemic; IDNA, iron-depleted nonanemic; ITT, intent-to-treat; LS, least square; WE, physical work efficiency.

2

Hb is reported in units of g/L.

3

Values are β estimates (95% CI).

4

Ferritin is reported in units of μg/L. Ferritin was not transformed for any secondary analysis because the residuals for the models were normally distributed in models using the untransformed variable but were not when using the log-transformed variable.