This is the old United Nations University website. Visit the new site at http://unu.edu
A lowcost tool for traditional birth attendants to identify lowbirthweight infants
M. M. A. Magzoub, T. E. Ali, and A. B. Ali
The research was a crosssectional study of newborns delivered in the Wad Medani teaching hospital, the largest hospital in Central State. It covered all deliveries from April through June 1991 and included 435 newborns. As our aim was not to ascertain the prevalence of low birth weight or its risk factors, all births were included, even preterm and twins.
Weight and chest, arm, and head circumferences were measured within the first 24 hours after birth, by three medical students who were trained in the techniques. The training addressed the difficulty of standardizing the chestcircumference measurements of newborns because they vary depending on the phases of respiration. The difference between repeated measurements should not be greater than 1 cm [6]. Weight was measured to the nearest 10 g, and the scale was checked every day for accuracy. Chest circumference was measured to the nearest 1 mm at a point crossing both nipples. Arm circumference was measured to the nearest 1 mm at a point equidistant from the olecranon and the acromion. Head occipitofrontal circumference was measured to the nearest 1 mm.
Analysis of the data included calculation of the mean values for each sex. Multiple regression analysis was used for prediction. Correlation coefficients were determined for all the measurements compared with birth weight and with each other. Regression equations were calculated for all the measurements, and predicted birthweight values were computed for the best predictor, chest circumference. The sensitivity, specificity, and positive predictive value (see table 1) of various chest circumferences as a predictor of low birth weight were also computed.
TABLE 1. Definitions of sensitivity, specificity, and positive predictive value
Predictor (other body measurement) 
Birth weight 

Positive 
Negative 

Positive  A, true positives  B. false positives 
Negative  C, false negatives  D, true negatives 
Sensitivity = A/(A + C)  
Specificity = D/(B + D)  
Positive predictive value = A/(A + B) 
Adapted from ref. 7.
On the basis of the findings, a measuring tape was developed relating chest circumferences to estimated birth weights, with sections coloured to indicate low, marginal, and normal weights, as described in the "Discussion" section below. Six village midwives, instructors, and ten students from the Wad Medani midwifery school were interviewed for comments on the new method, including its acceptability, feasibility, and practicability and the possibility of incorporating it into the curriculum.
Of the 435 infants, 226 (52%) were boys and 209 (48%) girls. Eightyfour per cent were fullterm and 6% preterm, and 8.3% were twins. Table 2 compares the measurements classified by sex. The boys were significantly heavier than the girls (p=.03).
TABLE 2. Analysis of variance for the effect of sex on body measurements
Males 
Females 
Significance 

N 
Mean 
SD 
N 
Mean 
SD 

Birth weight (kg)  224 
2.97 
0.54 
208 
2.85 
0.51 
0.03* 
Chest circumference (cm)  226 
31.65 
2 67 
209 
31.38 
2.45 
0.17 NS 
Arm circumference (cm)  226 
10.2 
1.18 
208 
10.08 
1.03 
0.07 NS 
Head circumference (cm)  224 
34.51 
2.05 
208 
33.88 
1.83 
0 14 NS 
*p < .05. NS = not significant.
The results of the regression analysis shown in table 3 suggest that chest circumference is the best predictor of birth weight (R2=82%). There was only a 5% improvement on the model when arm and head circumferences were added. Table 4 shows the correlation’s between birth weight and the various body measurements; again chest circumference shows the highest correlation with birth weight (r=.907). Figure 2 (see Figure. 2. Birth weights plotted against chest circumferences) plots birth weight against chest circumference.
TABLE 3. Stepwise regression analysis of the effect of birth weight on different body measurements
Regression estimate 
Partial R2 
Significance 

Intercept   3.9622 
< .001 

Chest circumference  0.1082 
.82 
< .001 
Arm circumference  0.1371 
.03 
< .001 
Head circumference  0.065 
.02 
< .001 
TABLE 4. Correlation coefficients among selected body measurements
BW 
CC 
AC 
HC 

BW  1.00 

CC  0.907* 
1.00 

AC  0.861* 
0.851* 
1.00 

HC  0.760* 
0.694* 
0.645* 
1.00 
BW = birth weight, CC = chest circumference, AC = arm circumference, HC = head circumference.
*n < .001
Table 5 compares the predicted birthweight values (BOO) for some chestcircumference measurements (CC) as computed from the regression equation with the actual mean birth weights. A chest circumference of 29.34 cm was the cutoff point for predieting low birth weight. (The regression equations for arm circumference [AC] and head circumference [HC] were BW=  1.36+0.42 AC and BW=  4.1] + 0.21 HC respectively.)
TABLE 5. Predicted and actual mean birth weights for some chestcircumference measurements
Chest circumference (cm) 
Predicted weight (kg) 
Actual value 

N 
Mean weight (kg) 

26  1.85 
3 
1.74 
28  2.22 
7 
2.15 
29  2.42 
11 
2.45 
30  2.62 
7 
2.63 
31  2.80 
13 
2.84 
32  2.99 
12 
3.04 
33  3.19 
10 
3.21 
34  3.38 
4 
3.46 
BW =  3.11 + 0.19 CC
Table 6 shows the sensitivity, specificity, and positive predictive value of some chest measurement' compared with the 2.5kg cutoff point for low birth weight.
TABLE 6. Sensitivity, specificity, and positive predictive value of various chest circumferences taken as the cutoff point for predicting low birth weight
Cutoff point (cm) 
Sensitivity 
Specificity 
Positive predictive value 
Low birth weight (%) 

Actual^{a} 
By chest measurement^{b} 

29  67.4 
96.5 
83.8 
21.3 
17.1 
30  91.3 
89.1 
69.4 
21.3 
28.0 
31  97.8 
72.6 
49.2 
21.3 
42.4 
32  98.9 
54.1 
36.8 
21.3 
57.2 
a. Percentage of newborns in the study group whose actual birth weight was below 2.5 kg.
b. Percentage who would be identified as of low birth weight if the respective chest measurement were used as the cutoff point.
Assessment of instructors' and students' feedback in the Wad Medani midwifery school showed the proposed new method is easy to understand learn, and apply. It look only one hour for the students to understand it.