Table 2.
Predictors of per-protocol MDA coverage in DeWorm3 at the cluster level, by site
| Benin (n = 120 observations)a |
India (n = 116 observations)a |
Malawi (n = 120 observations)a |
||||
|---|---|---|---|---|---|---|
| Predictors | Adj. dy/dx (95% CI)b | P-value | Adj. dy/dx (95% CI)b | P-value | Adj. dy/dx (95% CI)b | P-value |
| MDA round | ||||||
| 2 | −0.058 (−0.0137, 0.021) | 0.15 | −0.022 (−0.052, 0.008) | 0.14 | −0.022 (−0.053, 0.009) | 0.16 |
| 3 | 0.104 (0.080, 0.128) | < 0.001 | 0.020 (−0.003, 0.043) | 0.08 | 0.095 (0.063, 0.127) | < 0.001 |
| 4 | 0.028 (−0.011, 0.068) | 0.16 | 0.005 (−0.023, 0.033) | 0.73 | 0.095 (0.065, 0.124) | < 0.001 |
| 5 | 0.021 (−0.028, 0.071) | 0.39 | −0.058 (−0.109, −0.008) | 0.02 | 0.131 (0.095, 0.166) | < 0.001 |
| 6 | 0.010 (−0.034, 0.055) | 0.66 | −0.007 (−0.057, 0.043) | 0.78 | 0.103 (0.074, 0.133) | < 0.001 |
| Proportion of population in cluster | ||||||
| Adult (over 15 years), per tertile | 0.004 (−0.016, 0.025) | 0.67 | −0.008 (−0.025, 0.009) | 0.37 | −0.003 (−0.020, 0.015) | 0.76 |
| Male | −0.001 (−0.011, 0.010) | 0.91 | −0.002 (−0.012, 0.008) | 0.66 | −0.007 (−0.015, 0.002) | 0.12 |
| Migratory, per quintile | −0.002 (−0.017, 0.012) | 0.77 | 0.001 (−0.009, 0.011) | 0.90 | 0.003 (−0.011, 0.016) | 0.71 |
| Speaking minority languagesc | −0.023 (−0.041, −0.005) | 0.01 | 0.012 (0.001, 0.023) | 0.04 | −0.010 (−0.027, 0.007) | 0.24 |
| Minority religiond | 0.012 (−0.016, 0.040) | 0.40 | −0.003 (−0.014, 0.009) | 0.66 | 0.003 (−0.015, 0.021) | 0.72 |
| Polygamouse | 0.002 (−0.025, 0.029) | 0.89 | NA | −0.007 (−0.021, 0.008) | 0.37 | |
| Cluster mean population densityf | 0.00002 (−0.001, 0.001) | 0.96 | −0.0002 (−0.001, 0.0003) | 0.46 | −0.001 (−0.003, 0.001) | 0.59 |
| Number of CDDs trainedg | 0.007 (−0.010, 0.024) | 0.41 | −0.001 (−0.004, 0.002) | 0.47 | NA | |
| Sensitization activities | ||||||
| Community meetings | −0.017 (−0.034, 0.001) | 0.06 | 0.001 (−0.001, 0.003) | 0.20 | 0.004 (−0.002, 0.009) | 0.20 |
| Public dialogue event | 0.003 (−0.001, 0.006) | 0.17 | Dropped | −0.0004 (−0.004, 0.003) | 0.81 | |
| Printed IEC materials, quartiles | −0.019 (−0.043, 0.006) | 0.14 | −0.002 (−0.007, 0.003) | 0.38 | 0.005 (−0.013, 0.024) | 0.59 |
| Door to door visits, quartiles | −0.003 (−0.012, 0.005) | 0.48 | −0.001 (−0.009, 0.008) | 0.90 | 0.006 (−0.002, 0.013) | 0.14 |
| Radio, quartiles | 0.033 (0.007, 0.058) | 0.01 | No activity | No activity | ||
| Other mass media | No activity | Dropped | No activity | |||
Models used generalized estimating equations with binomial distribution, autoregressive correlation structure and robust standard errors. Adjusted models were selected using the quasi-likelihood under the independence model criterion (QIC). Differences in per-protocol coverage are reported per one unit increase in each predictor (dy/dx). Factors found to be significant at the 0.05 level in the fully adjusted model are indicated in bold
aNumber of observations per site reflect 20 intervention clusters per site multiplied by six MDA rounds. Four clusters in MDA round 5 were omitted from the India data due to COVID-19 related government shutdown orders
bDifference in proportion treated per one unit increase
cThree categories in Benin and Malawi, four in India. Majority language: Pedah, Sahoue, Watchi, Mina, Adja, or Xwla in Benin; Chiyao in Malawi; Tamil in India. Minority language: Fon or other in Benin; Chichewa or other in Malawi; Telegu, Urdu, Hindi or other in India
dMajority religion: Christianity in Benin; Islam in Malawi; Hinduism in India. Minority religion: Islam, Voodoo, or traditional religion in Benin; Christianity or other in Malawi; Christianity, Islam, or other in India
eNot assessed in India. Variable is grouped as “greater than median” or “less than or equal to median”
fMean population density within 0.5 km of households, per 100 people/km2
gData not available for all MDA rounds in Malawi