Skip to main content

Table 3 Measures of association and variation between individual- and community-level factors and childhood anemia

From: Maternal anemia is a potential risk factor for anemia in children aged 6–59 months in Southern Africa: a multilevel analysis

Characteristic Malawi Mozambique Namibia Zimbabwe
  aOR 95% (CI)/p aOR 95% (CI)/p aOR 95% (CI)/p aOR 95% (CI)/p
Exposure variable
Maternal anemic status
 Nonanemic 1.00 1.00 1.00 1.00
 Anemic 1.69 (1.35–2.12)a 1.71 (1.37–2.13)a 1.55 (1.08–2.22)c 1.52 (1.25–1.84)a
Maternal Hb concentration 0.64 (0.55–0.55)a 0.67 (0.58–0.78)a 0.70 (0.54–0.91)b 0.75 (0.65–0.86)a
Individual-level
Biofuel smoke exposureβ
 No NA 1.00 1.00 NA
 Yes NA 0.93 (0.50–1.73) 0.92 (0.58–1.45) NA
Sex of child
 Male 1.00 NA 1.00 NA
 Female 0.93 (0.78–1.11) NA 0.76 (0.59–0.99)a NA
Child age in months
 6–11 1.00 1.00 1.00 1.00
 12–23 0.28 (0.17–0.44)a 0.58 (0.39–0.88)a 1.08 (0.70–1.69) 0.99 (0.74–1.33)
 24–35 0.17 (0.11–0.27)a 0.36 (0.24–0.54)a 0.51 (0.33–0.79)b 0.46 (0.35–0.62)a
 36–47 0.12 (0.07–0.19)a 0.30 (0.20–0.46)a 0.31 (0.20–0.49)a 0.31 (0.23–0.42)a
 48–59 0.09 (0.06–0.15)a 0.22 (0.14–0.33)a 0.26 (0.16–0.43)a 0.25 (0.18–0.34)a
Recent episodes of fever
 No 1.00 1.00 1.00 1.00
 Yes 1.57 (1.29–1.90)a 1.68 (1.19–2.37)a 1.25 (0.93–1.68) 1.09 (0.84–1.43)
Recent episodes of diarrhea
 No 1.00 1.00 1.00 1.00
 Yes 0.95 (0.73–1.24) 1.07 (0.74–1.55) 1.21 (0.86–1.70) 1.07 (0.84–1.35)
Children are stunted
 No 1.00 1.00 1.00 1.00
 Yes 1.33 (1.10–1.61)b 1.49 (1.17–1.89)a 1.50 (1.05–2.15)c 1.17 (0.97–1.42)
Children are underweight
 No 1.00 1.00 1.00 1.00
 Yes 1.22 (0.91–1.62) 1.19 (0.84–1.69) 0.65 (0.42–1.02) 1.26 (0.94–1.69)
Mother’s age in years
 15–24 1.00 1.00 1.00 1.00
 25–34 0.92 (0.75–1.14) 0.90 (0.69–1.16) 0.70 (0.51–0.95)c 0.89 (0.74–1.08)
 35–49 0.96 (0.74–1.25) 0.92 (0.68–1.24) 0.66 (0.46–0.95)c 0.83 (0.65–1.06)
Mother’s educational level
 No education 1.00 1.00 1.00 NA
 Primary 0.87 (0.66–1.13) 1.09 (0.85–1.40) 0.82 (0.46–1.45) NA
 Secondary and above 0.73 (0.50–1.06) 0.66 (0.42–1.04) 0.83 (0.47–1.44) NA
Household wealth index
 Poorest 1.00 1.00 1.00 1.00
 Poor 0.88 (0.50–1.06) 0.89 (0.64–1.25) 1.53 (1.01–2.30)c 0.96 (0.75–1.22)
 Middle 0.96 (0.73–1.27) 0.80 (0.56–1.13) 1.09 (0.72–1.66) 0.82 (0.64–1.06)
 Richer 0.81 (0.60–1.08) 0.84 (0.56–1.27) 1.08 (0.62–1.88) 0.91 (0.67–1.23)
 Richest 0.69 (0.48–0.99)c 0.68 (0.38–1.24) 0.74 (0.36–1.53) 0.76 (0.53–1.11)
Community-level (clusters)
Community maternal anemia
 Low anemia 1.00 1.00 1.00 1.00
 High anemia 1.19 (0.93–1.54) 1.52 (1.19–1.94)a 1.478 (0.97–2.25) 0.778 (0.63–1.03)
Place of residence
 Urban 1.00 1.00 1.00 1.00
 Rural 1.07 (0.71–1.60) 1.14 (0.80–1.63) 0.87 (0.61–1.25) 0.78 (0.58–1.03)
Community wealth††
 Low 1.00 1.00 1.00 NA
 High 0.91 (0.69–1.19) 0.78 (0.53–1.14) 0.84 (0.53–1.34) NA
Community female education§
 Low 1.00 1.00 1.00 NA
 High 0.82 (0.65–1.03) 1.14 (0.83–1.55) 0.801 (0.50–1.29) NA
Community distance to HF¥
 Low 1.00 1.00 1.00 NA
 High 1.33 (1.07–1.65)b 0.98 (0.73–1.33) 0.99 (0.72–1.36) NA
Community safe water access
 Low 1.00 1.00 NA 1.00
 High 0.90 (0.72–1.12) 0.78 (0.57–1.07) NA 0.91 (0.73–1.14)
Measures of variation or clustering
Community level
 [τ(SE)] 0.193 (0.084)c 0.418 (0.120)a 0.157 (0.126) 0.086 (0.048)c
 [τ(SE)] 0.056 (0.080) 0.280 (0.114)b 0.070 (0.129) 0.064 (0.050)
 MORð 1.52 1.85 1.46 1.32
 MOR 1.25 1.66 1.29 1.27
 ICC (%) 5.54 11.27 4.55 2.55
 ICC (%) 1.68 7.84 2.09 1.90
 PCV (%) 70.94 32.98 55.31 26.07
Model fit statistics
 DIC (−2log likelihood) 2935.72 2220.25 1397.00 3282.89
  1. Notes: Note: aP < 0.0001; bP < 0.001; cP < 0.05; aOR, adjusted odds ratio; CI, confidence interval; [τ (SE)], community-level variance; SE, standard error; MOR, median odds ration; ICC, intraclass correlation; PCV, proportional change in variance; the proportional change in variance expresses the change in the community level variance between the null model and the individual level model, and between the individual level model, and the model further including the community level covariate; DIC, deviation information criterion. Model I: unconditional model with random intercepts and had no predictors. Model II: contained a random-intercept fixed-slope and adjusted for individual-level factors. Model III: contained a random-intercept fixed-slope and adjusted for community contextual factors. Model IV: contained a random-intercept fixed-slope and controlled for both individual and community-level factors. The Bold texts indicate a statistically significant association at a p-value less than 0.05. percent of women with fertility rate of 5 children and above; †† percent of households categorized above 60% of wealth index; § percent of women with primary school education and above; ¥ percent of households perceived distance to the nearest health facility as a big problem; percent of household with access to clean and safe drinking water sources specified by WHO/Unicef [34]; HF, health facility. ðMedian Odds Ratio for null Model; ¶Interclass Correlation for null Model; community-level variance for null Model; NA variable had p-value greater than 0.25 at chi-square