Further, bivariate multinomial logistic regressions (MLR) were run between the dependent variable, BW, and each of the predictor variables. Whichever predictor variable that one of the categories (LBW and/or Macrosomia) did not make a p-value ≤ 0.1 was not to be included in the multivariate MLR, except where theoretical consideration permitted. Among those included without any of their categories being at the threshold of p = 0.1 were ANC attendance and smoking habit of a responding mother as indicated in Table 2.
Table 2
The bivariate multinomial logistic regression for the predictors of birth weight
Dependent Variable | Birth Weight |
| Low Birth weight | Macrosomia |
Independent variable | Coef | p-value | CI | Coef | p-value | CI |
Socio-Econ characteristics | | | | | | |
Religion | -0.011 | 0.928 | -0.252–0.230 | 0.073 | 0.378 | -0.089–0.234 |
Maternal age 12–19 | 0.414 | 0.000 | 0.217–0.610 | -0.180 | 0.023 | -0.335 - -0.025 |
Maternal age 20–29 | -0.126 | 0.149 | -0.297–0.045 | -0.050 | 0.405 | -0.167–0.067 |
Maternal age 30–39 | -0.146 | 0.167 | -0.353–0.061 | 0.117 | 0.085 | -0.016–0.250 |
Maternal age 40–49 | -0.649 | 0.049 | -1.294 - − 0.004 | 0.435 | 0.003 | 0.149 - 0.721 |
Partner age 15–19 | 0.133 | 0.830 | -1.075–1.339 | -1.191 | 0.107 | -2.637–0.256 |
Partner age 20–29 | 0.194 | 0.059 | -0.008–0.396 | -0.285 | 0.000 | -0.433 - -0.136 |
Partner age 30–39 | -0.158 | 0.114 | -0.355–0.038 | -0.077 | 0.253 | -0.210–0.055 |
Partner age 40–49 | 0.071 | 0.549 | -0.162–0.304 | 0.301 | 0.000 | 0.1496–0.453 |
Partner age 50–95 | -0.342 | 0.141 | -0.798–0.113 | 0.382 | 0.002 | 0.143–0.620 |
Married | -0.154 | 0.158 | -0.368–0.0600 | -0.136 | 0.072 | -0.283–0.012 |
Not employed | 0.0780 | 0.494 | -0.145–0.299 | -0.359 | 0.000 | -0.531–0.187 |
Manual labourer | 0.321 | 0.000 | 0.142–0.500 | 0.256 | 0.000 | 0.135–0.377 |
skilled employee | -0.536 | 0.000 | -0.764 - -0.308 | -0.086 | 0.219 | -0.223–0.051 |
No education | 0.124 | 0.365 | -0.144–0.393 | 0.289 | 0.001 | 0.113–0.465 |
Primary education | -0.200 | 0.001 | -0.315 - -0.086 | -0.268 | 0.000 | -0.347 - -0.189 |
Secondary education | -0.266 | 0.014 | -0.477 - -0.055 | -0.157 | 0.028 | -0.296 - -0.017 |
Higher education | -0.399 | 0.022 | -0.741 - -0.056 | -0.761 | 0.000 | -1.024 - -0.498 |
Poorer quintile | 0.363 | 0.000 | 0.174–0.552 | -0.073 | 0.312 | -0.214–0.069 |
Poor quintile | 0.244 | 0.022 | 0.035–0.453 | 0.237 | 0.001 | 0.093–0.381 |
Middle quintile | -0.030 | 0.798 | -0.261–0.201 | 0.175 | 0.022 | 0.025–0.325 |
Richer quintile | -0.298 | 0.016 | -0.541 - -0.055 | 0.016 | 0.835 | -0.135–0.167 |
Richest quintile | -0.390 | 0.001 | -0.610 - -0.169 | -0.315 | 0.000 | -0.461 - -0.169 |
Neighbourhood characteristics | | | | | | |
Drinking water source | 0.031 | 0.679 | -0.117–0.180 | -0.065 | 0.223 | -0.168–0.039 |
Place of cooking | -0.042 | 0.762 | -0.313–0.229 | 0.149 | 0.137 | -0.048–0.346 |
hygienic toilet | -0.332 | 0.000 | -0.516 - -0.147 | -0.130 | 0.037 | -0.252 - -0.008 |
Cooking fuel pollution | -0.190 | 0.718 | -1.220–0.841 | -0.696 | 0.112 | -1.555–0.163 |
Type of housing | -0.000 | 0.876 | -0.005–0.005 | 0.001 | 0.453 | -0.002–0.004 |
Urban resident | -0.275 | 0.008 | -0.479–0.070 | -0.414 | 0.000 | -0.557–0.271 |
Maternal & obstetric characteristics | | | | | |
ANC attendance | -0.554 | 0.061 | -1.135–0.027 | -0.404 | 0.067 | -0.837–0.029 |
1st Trimester | 0.127 | 0.169 | -0.054–0.308 | 0.047 | 0.460 | -0.078–0.172 |
2st Trimester | -0.081 | 0.356 | -0.254–0.092 | -0.0559 | 0.352 | -0.174–0.062 |
3rd Trimester | -0.050 | 0.667 | -0.275–0.176 | 0.0288 | 0.708 | -0.122–0.179 |
Low BMI | -0.353 | 0.275 | -0.987–0.281 | -0.352 | 0.086 | -0.754–0.050 |
Normal BMI | -0.002 | 0.755 | -0.018–0.013 | -0.000 | 0.877 | -0.001–0.001 |
High BMI | -0.205 | 0.291 | -0.585–0.175 | 0.054 | 0.646 | -0.177–0.285 |
smoking behaviour | 0.064 | 0.619 | -0.187–0.313 | 0.043 | 0.628 | -0.130–0.215 |
Parity 1 | 0.366 | 0.000 | 0.179–0.553 | -0.475 | 0.000 | -0.629 - -0.320 |
Parity 2 | -0.080 | 0.476 | -0.298–0.139 | -0.163 | 0.036 | -0.315 - -0.011 |
Parity 3 | -0.233 | 0.069 | -0.484–0.018 | -0.141 | 0.096 | -0.306–0.025 |
Parity 4 | -0.115 | 0.417 | -0.392–0.162 | 0.218 | 0.012 | 0.047 − 0.389 |
Parity 5+ | -0.098 | 0.320 | -0.291–0.095 | 0.421 | 0.000 | 0.299–0.543 |
Delivery by caesarean | -0.128 | 0.434 | -0.447–0.192 | -0.013 | 0.900 | -0.222–0.195 |
Premature birth | 0.861 | 0.000 | 0.666–1.056 | -0.378 | 0.000 | -0.562 - -0.193 |
Normal delivery period | -0.623 | 0.000 | -0.806–0.441 | -0.056 | 0.427 | -0.195–0.082 |
Late Birth date | -0.266 | 0.144 | -0.624–0.091 | 0.570 | 0.000 | 0.384–0.756 |
Short birth interval | 0.241 | 0.016 | 0.045–0.437 | 0.185 | 0.007 | 0.051–0.319 |
Unwanted pregnancy | 0.219 | 0.012 | 0.048–0.391 | 0.066 | 0.273 | -0.052–0.184 |
Facility delivery | -0.670 | 0.000 | -0.953–0.379 | -0.642 | 0.000 | -0.847–0.436 |
Child characteristics factors | | | | | | |
male fetus | -0.196 | 0.025 | -0.368–0.025 | 0.318 | 0.000 | 0.200–0.436 |
≤ 4 siblings | -0.029 | 0.752 | -0.206–0.149 | -0.088 | 0.160 | -0.210- -0.035 |
Source: Researchers |
Interpretation of the Results of the bivariate multinomial logistic regression for the predictors of birth weight
The bivariate MLR of the socio-economic category p-value results of the predictors was as follows; religion had a p-value = 0.928 for LBW and p-value = 0.378 for macrosomia. Maternal age in years having been categorized as; 12–19, 20–29.99, 30–39, and 40–49; corresponding with p-values; p = 0.000 LBW & p = 0.023 macrosomia, p = 0.311 LBW & p = 0.533, for macrosomia, p = 0.170 LBW & p = 0.192, for macrosomia, and p = 0.030 LBW & p = 0.174 macrosomia respectively. More so, partner age in years was categorized; 15–19, 20–29, 30–39, 40–49, and 50–95; corresponding with; 0.830 LBW & p = 0.107 macrosomia, p = 0.059 LBW & p = 0.000 macrosomia, p = 0.114 LBW & p = 0.253 macrosomia, p = 0.549 LBW & p = 0.000 macrosomia, p = 0.549 LBW & p = 0.000 macrosomia, and p = 0.141 LBW & p = 0.002 macrosomia respectively. Besides, marital status was p = 0.158 LBW & p = 0.072 macrosomia; employment status categorized as; not employed, manual laborer, and skilled salaried employee, corresponding with; p = 0.494 LBW & p = 0.000 macrosomia, p = 0.000 LBW & p = 0.000 macrosomia, and p = 0.000 LBW & p = 0.219 macrosomia, respectively. Moreover, maternal education was categorized as; no education, primary education, secondary education, and higher education, corresponding with; p = 0.365 LBW & 0.001 macrosomia, p = 0.001 LBW & 0.000 macrosomia, p = 0.014 LBW & 0.028 macrosomia, and p = 0.022 LBW & p = 0.000 macrosomia, respectively. Still under the socio-economic characteristic predictors, the household wealth index was given due consideration using the five quintiles used in the UDHS as; poorest, poorer, middle, richer, and richest quintile corresponding to the following p-values; p = 0.000 LBW & p = 0.312 macrosomia, p = 0.022 LBW & p = 0.001 macrosomia, p = 0.798 LBW & p = 0.022 macrosomia, p = 0.016 LBW & p = 0.835 macrosomia, and p = 0.001 LBW & p = 0.000 macrosomia, respectively.
The bivariate MLR for the neighborhood characteristics produced the following p-values as per of the predictors: drinking water source, p = 0.679 LBW & p = 0.223 macrosomia; place of cooking, p = 0.762 LBW & p = 0.137 macrosomia; hygienic toilet, p = 0.000 LBW & p = 0.037 macrosomia; cooking fuel pollution, p = 0.718 LBW & p = 0.112 macrosomia; type of housing, p = 0.876 LBW & p = 0.453 macrosomia; and place of residence p = 0.008 LBW & p = 0.000 macrosomia.
Subsequently, the bivariate MLR for maternal & obstetric characteristic predictors indicated the p-values accordingly: ANC attendance with p = 0.061 LBW & p = 0.067 macrosomia; timing of the ANC categorized as; first 3 months, fourth to 5th month, 6th to last month of birth, with corresponding p-values as p = 0.169 LBW & p = 0.460 macrosomia, p = 0.356 LBW & p = 0.352 macrosomia, and p = 0.667 LBW & p = 0.708 macrosomia, respectively. BMI categorizes as low BMI, normal BMI, and high BMI with the corresponding p-values; p = 0.275 LBW & p = 0.086 macrosomia, p = 0.755 LBW & p = 0.877 macrosomia, and p = 0.291 LBW & p = 0.646 macrosomia, respectively. The respondents’ smoking behavior was also tested at the bivariate MLR level indicating p-values of p = 0.619 LBW & p = 0.628 macrosomia. Further analysis of parity was done with categories of, 1st parity, 2nd parity, 3rd parity, 4th parity, and more than 5 parity giving the p-values as; p = 0.000 LBW & p = 0.000 macrosomia, p = 0.476 LBW & p = 0.036 macrosomia, p = 0.069 LBW & p = 0.096 macrosomia, p = 0.417 LBW & p = 0.012 macrosomia, and p = 0.320 LBW & p = 0.000 macrosomia, respectively. Delivery by caesarean was also considered with arising p-values as p = 0.434 LBW & p = 0.900 macrosomia. In addition, period of delivery was categorized into three categories including; pre-term delivery period, normal delivery period, and beyond delivery period; the corresponding p-values are; p = 0.000 LBW & p = 0.000 macrosomia, p = 0.000 LBW & p = 0.427 macrosomia, and p = 0.144 lbw & p = 0.000 macrosomia, respectively. The other predictor is the preceding birth interval before the previous birth whose p-values came out as p = 0.016 LBW & p = 0.007 macrosomia. Unwanted pregnancy is another predictor, with p-values p = 0.012 LBW & p = 0.273 macrosomia. Lastly for the maternal and obstetric predictors is maternal place of delivery with p-values of p = 0.000 LBW & p = 0.000 macrosomia.
Finally, for the bivariate MLR, are the child characteristics with child sex having p-values of p = 0.025 LBW & p = 0.000 macrosomia and number of siblings having p-values of 0.752 LBW & p = 0.160 macrosomia.
Results Of The Multivariate Multinomial Logistic Regression For The Predictors Of Birth Weight
Consequently, the predictors that were selected from the bivariate stage of MLR and included in the multivariate MLR model, produced the results presented in Table 3, drawn using a stata command asdoc formed by Shah [23] indicating their p-values, Relative Risk ratios, and confidence intervals. During the running of the multivariate MLR, the likelihood ratio test rejected the null hypothesis of the fitted model that, all predictor regression coefficients in the model are simultaneously equal to zero; that is, the likelihood ration (LR) chi-square test for LBW relative to normal birth weight (NBW) and macrosomia relative to NBW. This is justified by the LR chi-square statistic calculated has: -2*(L (null model) – L (fitted model)) = -2(-3602.6505) – (-3477.8695) = LR chi(50) = 249.56; and with Prob > chi2 = 0.0000.
The analysis as indicated in Table 3, from the LBW category, several predictors were in line with the expected signs, whereas, others, were contrary to what theory stipulates. The macrosomia category of the dependent variables, the predictor signs were in line with what the theory stipulates.
Table 3
Multinomial logistic regression results
Birth weight | RRR | St.Err. | t-value | p-value | [95% Conf | Interval] | Sig |
Low Birth weight | | | | | | | |
Maternal age at birth (years) | | | | | |
12–19 | 1.397 | 0.315 | 1.48 | 0.138 | 0.898 | 2.174 | |
20–29 | 1.026 | 0.171 | 0.16 | 0.876 | 0.741 | 1.422 | |
30–39 | … | … | … | … | … | … | |
40–49 | 0.676 | 0.238 | -1.11 | 0.267 | 0.339 | 1.349 | |
Partner’s age (years) | | | | | | |
15–19 | … | … | … | … | … | … | |
20–29 | … | … | … | … | … | … | |
30–39 | … | … | … | … | … | … | |
40–49 | 1.270 | 0.194 | 1.56 | 0.118 | 0.941 | 1.713 | |
50–95 | 0.878 | 0.233 | -0.49 | 0.624 | 0.521 | 1.478 | |
HH wealth index | | | | | | | |
Poorer quint | 1.422 | 0.371 | 1.35 | 0.177 | 0.853 | 2.371 | |
Poor quintile | 1.225 | 0.321 | 0.78 | 0.439 | 0.733 | 2.046 | |
Middle quintile | 1.121 | 0.286 | 0.45 | 0.655 | 0.679 | 1.849 | |
Richer quintile | 0.848 | 0.206 | -0.68 | 0.497 | 0.527 | 1.365 | |
Riches quintile | … | … | … | … | … | … | |
Hygienic toilet | 0.942 | 0.157 | -0.36 | 0.719 | 0.680 | 1.305 | |
Urban resident | 0.812 | 0.141 | -1.20 | 0.231 | 0.578 | 1.141 | |
Antenatal att. | 0.380 | 0.143 | -2.57 | 0.010 | 0.182 | 0.794 | ** |
Period of delivery | | | | | | |
Preterm birth | 3.456 | 0.939 | 4.57 | 0.000 | 2.029 | 5.886 | *** |
Normal period | 1.427 | 0.363 | 1.40 | 0.163 | 0.866 | 2.350 | |
Short interval | 1.134 | 0.128 | 1.11 | 0.265 | 0.909 | 1.415 | |
Facility deliver | 0.584 | 0.112 | -2.81 | 0.005 | 0.401 | 0.850 | *** |
Male neonate | 0.808 | 0.093 | -1.85 | 0.064 | 0.645 | 1.013 | * |
Smoking habit | 1.023 | 0.165 | 0.14 | 0.890 | 0.745 | 1.404 | |
No education | 1.099 | 0.252 | 0.41 | 0.682 | 0.700 | 1.724 | |
Primary educ | 1.075 | 0.148 | 0.53 | 0.599 | 0.821 | 1.407 | |
Secondary educ | 1.034 | 0.188 | 0.18 | 0.856 | 0.723 | 1.477 | |
Unwanted preg | 1.097 | 0.132 | 0.77 | 0.443 | 0.866 | 1.389 | |
Parity 1 | … | … | … | … | ... | … | |
Parity 2 | 1.083 | 0.209 | 0.41 | 0.680 | 0.742 | 1.579 | |
Parity 3 | 0.890 | 0.167 | -0.62 | 0.535 | 0.616 | 1.286 | |
Parity 4 | 1.042 | 0.190 | 0.22 | 0.823 | 0.728 | 1.490 | |
Constant | 0.212 | 0.125 | -2.64 | 0.008 | 0.067 | 0.671 | *** |
Normal Birth weight (Base category) | | | | | |
Macrosomia | | | | | | | |
Maternal age at birth (years) | | | | | | |
12–19 | 1.244 | 0.202 | 1.35 | 0.177 | 0.906 | 1.709 | |
20–29 | 0.964 | 0.106 | -0.33 | 0.741 | 0.777 | 1.196 | |
30–39 | … | … | … | … | … | … | |
40–49.00 | 1.304 | 0.224 | 1.55 | 0.122 | 0.931 | 1.827 | |
Partner’s age (years) | | | | | | |
15–19 | … | … | … | … | … | … | |
20–29 | … | … | … | … | … | … | |
30–39 | … | … | … | … | … | … | |
40–49 | 1.148 | 0.109 | 1.45 | 0.149 | 0.952 | 1.384 | |
50–95 | 1.177 | 0.171 | 1.12 | 0.262 | 0.885 | 1.563 | |
HH wealth index | | | | | | |
Poorest quintile | 0.881 | 0.144 | -0.78 | 0.437 | 0.640 | 1.212 | |
Poorer quintile | 1.187 | 0.186 | 1.10 | 0.273 | 0.873 | 1.615 | |
Middle quintile | 1.082 | 0.165 | 0.52 | 0.604 | 0.803 | 1.459 | |
Richer quintile | 1.023 | 0.142 | 0.16 | 0.871 | 0.779 | 1.343 | |
Hygienic toilet | 1.175 | 0.115 | 1.65 | 0.100 | 0.970 | 1.425 | |
Urban resident | 0.690 | 0.075 | -3.39 | 0.001 | 0.557 | 0.855 | *** |
Antenatal att. | 0.598 | 0.172 | -1.78 | 0.075 | 0.340 | 1.052 | * |
Period of delivery | | | | | | | |
Preterm birth | 0.472 | 0.075 | -4.71 | 0.000 | 0.345 | 0.645 | *** |
Normal period | 0.602 | 0.069 | -4.44 | 0.000 | 0.482 | 0.753 | *** |
Short interval | 1.224 | 0.093 | 2.65 | 0.008 | 1.054 | 1.421 | *** |
Facility Deliv | 0.594 | 0.074 | -4.18 | 0.000 | 0.465 | 0.758 | *** |
Male fetus | 1.297 | 0.095 | 3.54 | 0.000 | 1.123 | 1.497 | *** |
Smoking habit | 1.020 | 0.107 | 0.19 | 0.850 | 0.830 | 1.253 | |
No education | 0.901 | 0.136 | -0.69 | 0.490 | 0.670 | 1.212 | |
Primary educ | 0.800 | 0.076 | -2.36 | 0.018 | 0.664 | 0.963 | ** |
Secondary edu | 1.292 | 0.155 | 2.14 | 0.032 | 1.022 | 1.634 | ** |
Unwanted preg | 0.971 | 0.074 | -0.38 | 0.702 | 0.836 | 1.128 | |
Parity 2 | 0.780 | 0.098 | -1.98 | 0.048 | 0.609 | 0.998 | ** |
Parity 3 | 0.786 | 0.091 | -2.08 | 0.038 | 0.626 | 0.987 | ** |
Parity 4 | 1.054 | 0.117 | 0.48 | 0.633 | 0.848 | 1.311 | |
Constant | 1.407 | 0.545 | 0.88 | 0.378 | 0.659 | 3.007 | |
Mean dependent var | 2.145 | SD dependent var | 0.521 | |
Pseudo r-squared | 0.035 | Number of obs | 4687.000 | |
Chi-square | 249.562 | Prob > chi2 | 0.000 | |
Akaike crit. (AIC) | 7059.739 | Bayesian crit. (BIC) | 7395.271 | |
*** p < 0.01, ** p < 0.05, * p < 0.1 | |
Interpretation For The Mlr Multivariate Results Of The Predictors Of Birth Weight
According to the results from the multivariate multinomial logistic regression model, none of the variables of LBW comparison category belonging to the socio-economic, neighborhood and child characteristic factors where significant. The low birth weight comparison category only had three variables significant at p-value ≤ 0.05, of which all belong to the maternal and obstetric characteristic factors of the theory guiding this study. That is, attendance of antenatal care, premature birth delivery, and place of delivery. As shown in Table 3, only premature birth had a positive association with the LBW comparison category. This means that, a unit increase in the premature birth delivery variable, puts the relative risk ratio of belonging to a LBW category by 3.5 times more likely, than belonging to the base category, normal birth weight (NBW), at a p-value = 0.000 and CI (2.029,5.886); when other variables in the model are held constant. Conversely, antenatal care attendance of four times and more, has a negative association with LBW. This means that a unit increase in antenatal care attendance of more than four times, would reduce the relative risk of belonging to the LBW category by 0.38 times, and this implies belonging to the NBW base category at p-value = 0.010 and CI (0.182,0.795); when other variables in the model are held constant.
Then place of delivery, where a unit increase in delivery at a health facility, would reduce the relative risk ratio of belonging to comparison category, LBW, by 0.58 times, and hence, belonging to the NBW base category, at p-value = 0.005, with CI (0.401, 0.850), holding other factors in the model constant. Yet, from the child characteristics, at 0.1 level of significance, child sex is significant at 0.064, whose unit increase in giving birth to a male baby compared to a female baby, reduces the relative risk of belonging to a comparison group by 0.81 times, at CI (0.645, 1.013), and favors the base category group NBW, holding the other variables in the model constant.
Furthermore, for the macrosomia comparison category; the results from the MLR model, show that two variables were significant at p-value ≤ 0.05 from the Socio-economic characteristics category, that is, respondent belonging to primary education category and respondent belonging to secondary school category. From the neighborhood characteristics, only one variable was significant at p-value ≤ 0.05, that is, place of residence. Yet, the maternal and obstetric characteristics, six variables were significant at p-value ≤ 0.05, that is, premature birth, normal delivery period, preceding birth interval, place of delivery, second parity, and third parity. Lastly, from the child characteristics, only child sex was the only significant variable at p-value ≤ 0.05.
Interpretation of these results shows that; a one-unit increase in primary education variable, reduces the relative risk of belonging to the comparison category, macrosomia, relative to the base category NBW; reducing by 0.80 times, at p-value = 0.018, CI (0.664,0.963); placing primary school to the base category, while holding all the other variables in the model constant. Besides, a unit increase in secondary education increases the relative risk of belonging to the macrosomia comparison category, relative to normal birth weight by 1.29 times, at p-value = 0.032, CI(1.022,1.634); while holding other variables in the model constant. For the neighborhood characteristics, a unit increase in the place of residence of living in urban areas versus living in rural areas, reduces the relative risk of belonging to the macrosomia category relative to NBW the base category, by 0.69 times, with p-value = 0.001, CI (0.557,0.855); holding other variables in the model constant.
Additionally, for the six variables of the maternal and obstetric characteristics; a unit increase in premature birth delivery variable reduces the relative risk of belonging to the macrosomia comparison category relative to NBW base category, by 0.47 times at p-value = 0.000, CI (0.345,0.645); holding other variables in the model constant. A unit increase in normal birth delivery, reduces the relative risk of belong to the macrosomia comparison category relative to NBW base category, by 0.60 times at p-value = 0.000, CI (0.482, 0.753); holding all other variables in the model constant. A unit increase in a birth interval less than two years, increases the relative risk of belonging to the comparison group macrosomia, relative to NBW, the base category, by 1.22 times at p-value = 0.008, CI (1.054, 1.421); holding other variables in the model constant. A unit increase in delivery in a health facility, reduces the relative risk of belonging to the comparison category, macrosomia, relative to the base category NBW, by 0.59 times at a p-value = 0.000, CI (0.465, 0.758); holding other variables in the model constant.
A unit increase in the second parity variable, reduces the relative risk of belonging to macrosomia, the comparison group, relative to NBW, the base category by 0.78 times, at p-value = 0.048, CI (0.609, 0.998); holding other variables in the model constant. A unit increase in the in the third parity variable, reduces the relative risk of belonging to the comparison group macrosomia, relative to base category, NBW, by 0.79 times, at p-value = 0.038, CI (0.626, 0.987); holding all other variables in the model constant. Lastly for maternal and obstetric characteristics, significant at 0.1, is attendance of ANC. A unit increase in ANC attendance, reduces the relative risk of belonging to the comparison group, relative to the reference group, NBW, by 0.60 times, at p-value = 0.075, CI (0.340, 1.052); holding all other variables in the model constant.
Finally, the child characteristics, unit increase in giving birth to a male than a female in the macrosomia comparison category, relative to NBW, increases the relative risk of belonging to macrosomia category than NBW, by 1.30 times, at p-value = 0.000, CI (1.223, 1.498); holding all other variables in the model constant.