This section contains the presentation, analysis and the discussion of the findings of the study. These are presented according to the main research questions raised to guide the study conducted by NISR on RDHS 2014.2015. First part of this chapter included the background information of the respondents. Second part focused on the main findings of the study about the contribution of community health workers in the prevention of the diseases caused by malnutrition.
6.1 Respondents Characteristics
This section dealt with the information collected on female population respondents in Rwanda. The characteristics of respondents discussed mainly include sex of respondents, female age group at first birth, place of residence, women marital status, place of residence and their highest education attainment.
6.2 Univariate analysis
This part presents the description of child characteristics: health workers care, wealth index, mother education level, place of residence, province, antenatal care, age, sex, breast feeding, for the purpose of ascertaining the percentage and frequency distributions.
Diseases caused by malnutrition and child characteristics
Stunting | Freq. Percent
|
------------+------------------------
|
Non-stunted | 473 41.60
|
stunted | 664 58.40
|
------------+------------------------
|
Total | 1,137 100.00
|
wasting
|
|
|
------------+------------------------
|
Non-wasted | 547 48.11
|
wasted | 590 51.89
|
------------+------------------------
|
Total | 1,137 100.00
|
weight
|
|
|
----------------+--------------------
|
non_
|
|
|
|
|
|
underweight | 725 63.76
|
underweight | 412 36.24
|
----------------+--------------------
|
Total | 1,137 100.00
|
|
|
|
|
Health workers
|
|
|
------------+-------------------------
|
no | 338 29.73
|
yes | 799 70.27
|
------------+-------------------------
|
Total | 1,137 100.00
|
|
|
|
|
Education level
|
|
|
--------------------------------------
|
Noeducation| 114 10.03
|
primary | 798 70.18
|
secondary | 170 14.95
|
higher | 55 4.84
|
--------------------------------------
|
Total | 1,137 100.00
|
|
|
|
|
wealth index
|
|
|
------------+-------------------------
|
poor | 428 37.64
|
middle | 200 17.59
|
rich | 509 44.77
|
------------+-------------------------
|
Total | 1,137 100.00
|
Breastfeeding
|
|
|
------------+-------------------------
|
no | 142 12.49
|
yes | 995 87.51
|
------------+-------------------------
|
Total | 1,137 100.00
|
|
|
|
|
|
|
Marital |
|
|
|
|
status |
|
|
|
------------+-------------------------
|
single | 38 3.34
|
married | 1,078 94.81
|
widowed | 1 0.09
|
divorced | 20 1.76
|
------------+-------------------------
|
Total | 1,137 100.00
|
sex of |
|
|
|
|
child |
|
|
|
------------+-------------------------
|
female | 591 51.98
|
male | 546 48.02
|
------------+-------------------------
|
Total | 1,137 100.00
|
|
|
|
|
|
|
Antenatal |
|
|
|
|
care |
|
|
|
------------+-------------------------
|
no | 1,026 90.24
|
yes | 111 9.76
|
------------+-------------------------
|
Total | 1,137 100.00
|
Source: DHS2014-15
According to all total respondent 58% had the problem stunting 51.89% were wasted and 36. 24% of the total population were under-weight. The majority met the health worker (70.27%) while other 29.63% did not. And most of them had primary level education with 70.18% followed by with those in secondary level with 14.95%, 44.77% are rich and 37.64% are poor. Here high number of children have had a breastfeeding while low number did not, 87.51% and 1.49% respectively. According to respondent’s marital status, most of them were married with 94.81% of the whole surveyed population and 51.98% were female and 49.02% were male. Above 90% did not get antenatal care
6.3 Bivariate
This section presents the results of bivariate analysis on diseases caused by malnutrition. As earlier mentioned, it establishes the strength of association between diseases caused by malnutrition and each explanatory variables by using cross tabulations and chi-square statistics. Tables below show the details of death patterns by each of the selected independent variables.
Factors that influence stunting
| stunting status
province total non-stunt (%) stunted (%) p value
----------------------------------------------------------------------
kigali city|168 104(61.90) 64(38.10)
south |301 108(35.88) 193(64.12)
west |239 84(35.15) 155(64.85)
north |164 54(32.93) 110(67.07)
east |265 123(46.42) 142(53.58) 0.000
----------------------------------------------------------------------
Health |
workers |
----------------------------------------------------------------------
no |338 79(23.37) 259(76.63)
yes |799 394(49.31) 405(50.69) 0.000
----------------------------------------------------------------------
type of |
place of |
residence |
----------------------------------------------------------------------
rural |836 295(35.29) 541(64.71)
urban |301 178(59.14) 123(40.86) | 0.000
----------------------------------------------------------------------
education |
level |
----------------------------------------------------------------------
Noeducation|114 30(26.32) 84(73.68)
primary |798 308(38.60) 490(61.40) |
secondary |170 96(56.47) 74(43.53) |
higher |55 39(70.91) 16(29.09) | 0.000
------------+----------------------+-------------------------------
wealth |
index |
----------------------------------------------------------------------
poor |428 134(31.31) 294(68.69) |
middle |200 62(31.00) 138(69.00) |
rich | 509 277(54.42) 232(45.58) | 0.000
----------------------------------------------------------------------
breastfeeding |
----------------------------------------------------------------------
no |142 104(73.24) 38(26.76) |
yes |995 369(37.09) 626(62.91) | 0.000
----------------------------------------------------------------------
Marital |
status |
----------------------------------------------------------------------
single | 38 16(42.11) 22(57.89) |
married |1,078 448(41.56) 630(58.44) |
widowed |1 1(100.00) 0(0.00) |
divorced | 20 8(40.00) 12(60.00) | 0.699
--------------------------------------------------------------------
sex of |
child |
----------------------------------------------------------------------
female |591 262(44.33) 329(55.67) |
male |546 211(38.64) 335(61.36) | 0.052
-------------------------------------------------------------------
Antenatal |
care |
----------------------------------------------------------------------
no |1026 381(37.13) 645(62.87) |
yes |111 92 (82.88) 19(17.12) | 0.000
----------------------------------------------------------------------
Source: RDHS2014-15
The results of bivariate analysis are shown in the table above. And it shows that, province, health worker, place of residence, breastfeeding, wealth index and antenatal care are statistically significant because their p-value are less than 0.05 (p value <0.05). Means that the all mentioned variables have an influence on stunting status.
Basing on findings indicated in table above, it shows that the problem of stunting among Rwandan children is still high. According the surveyed household distributed by province north, west and south provinces are at high level of stunting with 67.07%, 64.85% and 64.07% respectively and the province which has the small number of stunting is Kigali with 38.1% of the total population. Among the total surveyed children 76.63% of the total children who didn’t get help from health workers were stunted while 50.69% of the total children who met with health worker were also stunted.
From our findings it is indicated that large number of children who were stunted, their mother had low educational level, for those with no education 73.68% were stunted while with those of primary level 61.4% also were stunted. As the level of education increase number of children with stunting decreases. And large number stunting was from rural distributed children with 64.71% while urban is 40.86%. And male children are more likely to meet stunting problem compared to female. Also 62.87% of the total children who didn’t get antenatal care were stunted while for those who did, only 17.12% were stunted.
Factors that influence wasting
wasting status
province | Total non-waste (%) wasted (%) P value
------------+----------------------+---------------------------------
kigali city | 168 100(59.52) 68(40.48)
south | 301 146(48.50) 155(51.50)
west | 239 90(37.66) 149(62.34)
north| 164 76(46.34) 88(53.66)
east | 265 135(50.94) 130(49.06) 0.000
------------+----------------------+-------------------------------
Health |
workers |
-----------+---------------------------------------------------------
no | 338 143(42.31) 195(57.69)
yes | 799 404(50.56) 395(49.44) 0.000
------------+----------------------+----------------------------------
place of |
residence |
-----------+---------------------------------------------------------
rural | 836 360(43.06) 476(56.94)
urban | 301 187(62.13) 114(37.87) 0.000
------------+----------------------+---------------------------------
education |
level |
-----------+---------------------------------------------------------
no education | 114 40(35.09) 74(64.91)
primary | 798 370(46.37) 428(53.63)
secondary | 170 99(58.24) 71(41.76)
higher | 55 38(69.09) 17(30.91) 0.000
-----------+----------------------+-----------------------------------
wealth |
index |
-----------+---------------------------------------------------------
poor | 428 162(37.85) 266(62.15)
middle | 200 88(44.00) 112(56.00)
rich | 509 297(58.35) 212(41.65) 0.000
-----------+--------------------------------------------------------- breastfeeding
-----------+---------------------------------------------------------
no | 142 85(59.86) 57(40.14)
yes | 995 462(46.43) 533(53.57) 0.003
-----------+----------------------+-----------------------------------
Marital |
status |
-----------+---------------------------------------------------------
single | 38 17(44.74) 21(55.26)
married| 1,078 523(48.52) 555(51.48)
widowed | 1 0 (0.00) 1(100.00)
divorced | 20 7(35.00) 13(65.00) 0.467
-----------+----------------------+-----------------------------------
sex of |
child |
-----------+--------------------------------------------------------- female | 591 307(51.95) 284(48.05)
Male | 546 240(43.96) 306(56.04) 0.007
-----------+----------------------+----------------------------------- Antenatal care
-----------+---------------------------------------------------------
no | 1,026 470(45.81) 556(54.19)
yes | 111 77(69.37) 34(30.63) 0.000
source: RDHS 2014-15
The results of bivariate analysis are shown in the table above. And it shows that, province, health worker, place of residence, breastfeeding, wealth index and antenatal care are statistically significant because their p-value are less than 0.05 (p value <0.05) except maternal marital status. Means that the all mentioned variables have an influence on wasting status.
Among the total children who didn’t get facility from the health workers 57.69% were wasted while 49.44% only were wasted for those who got facility from health workers which means that health workers have a role in reducing the number of wasted children. The result shows that western province is heading other province in having high number of children leaving with the problem of wasting, with 62.34% of the total surveyed children are wasted and is followed by north province which has 62.34% and the province which has low number of wasted children is Kigali city with 40.48%.
For maternal educational level, the results indicated that children whose mothers has high level of education are more likely not to have this problem of wasting. For those with no education wasting was 64.91% while for those in higher education wasting was at 30.91 % as shown in the table. And higher number of wasting children was in rural areas, with 56.94% of the total surveyed children.
Factors that influence underweight
| Underweight status
Province |Total non-under(%) underweight(%) P-value
------------+----------------------+----------------------------------
kigali city |168 105(62.50) 63(37.50)
south |301 207(68.77) 94(31.23)
west |239 135(56.49) 104(43.51)
north |164 113(68.90) 51(31.10)
east |265 165(62.26) 100(37.74) 0.027
---------------------------------------------------------------------
Health |
workers |
------------+----------------------+----------------------------------
no |338 219(64.79) 119(35.21)
yes |799 506(63.33) 293(36.67)
------------+----------------------+----------------------------------
type of |
place of |
residence |
------------+----------------------+----------------------------------
rural |836 522(62.44) 314(37.56)
urban |301 203(67.44) 98(32.56) 0.122
--------------------------------------------------------------------
education|
level|
------------+----------------------+----------------------------------
no education|114 78(68.42) 36(31.58)
primary|798 507(63.53) 291(36.47)
secondary|170 102(60.00) 68(40.00)
higher|55 38(69.09) 17(30.91) 0.422
------------+----------------------+----------------------------------
wealth |
index |
------------+----------------------+----------------------------------
poor |428 272(63.55) 156(36.45)
middle |200 123(61.50) 77(38.50)
rich |509 330(64.83) 179(35.17) 0.703
----------------------------------------------------------------------
breastfeeding
------------+----------------------+----------------------------------
no |142 84(59.15) 58(40.85)
yes |995 641(64.42) 354(35.58) 0.222
---------------------------------------------------------------------
Marital |
status |
------------+----------------------+----------------------------------
single |38 23(60.53) 15(39.47)
married |1078 688(63.82) 390(36.18)
widowed |1 0(0.00) 1(100.00)
divorced |20 14(70.00) 6(30.00) 0.518
----------------------------------------------------------------------
sex of |
child |
---------------------------------------------------------------------
female |591 374(63.28) 217(36.72)
male |546 351(64.29) 195(35.71) 0.725
----------------------------------------------------------------------
Antenatal |
care |
---------------------------------------------------------------------
no |1026 656(63.94) 370(36.06)
yes |111 69(62.16) 42(37.84) 0.712
----------------------------------------------------------------------
SOURCE: RDHS 2014-15
The results of bivariate analysis are shown in the table above. And it shows that, all factors are statistically insignificant except province because their p-values are less than 0.05(p value <0.05). Means that the all mentioned variables have no influence on underweight status.
Here the province which is at higher number of underweight, western province and followed by east, and Kigali city with 43.51%, 37.74% and 37.50%. and the majority were male (56.04%).
6.4 Multivariate analysis
Binary logit parameter estimates of model 1
--------------------------------------------------------------------------------------
Stunting status| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+------------------------------------------------------------------------
Health worker | -1.217217 .1636756 -7.44 0.000 -1.538016 -.8964191
province | -.0568698 .0511028 -1.11 0.266 -.1570294 .0432898
place residence| -.571168 .1838945 -3.11 0.002 -.9315946 -.2107414
education level| -.3417421 .1208649 -2.83 0.005 -.5786329 -.1048513
wealth index | -.277401 .0861079 -3.22 0.001 -.4461695 -.1086326
breast feeding | 1.178479 .219007 5.38 0.000 .7492327 1.607724
antenatal care | -1.59828 .2778857 -5.75 0.000 -2.142926 -1.053634
_cons | 1.626605 .3673184 4.43 0.000 .9066738 2.346536
--------------------------------------------------------------------------------------
LR chi2(9) = 250.08 Prob > chi2 =0.0000 Pseudo R2= 0.1620Log likelihood = -646.94952
Source: DHS2015-15
The model is:
log(p)=1.626605-1.217217 health worker-.0568698 province-.571168 place residence -.3417421 education level-.277401 wealth index +1.178479 breast feeding -1.59828 antenatal care.
The likelihood of the model is the probability that you would observe the dichotomous (actually multichotomous) outcomes of the sample, given the coefficient estimates. The logit algorithms maximize the logarithm of this likelihood, and since the probabilities is between 0 and 1, then the log likelihood is always negative. Our maximum log likelihood of obtaining outcomes of health insurance coverage from the sample is -646.94952 and the test chi-square is 250.08.
As the calculated chi-square statistics is highly significant at 5%, then the null hypothesis that the constrained model is correct is rejected. This means that the probability in the upper tail beyond the calculated statistic is smaller than the significance level chosen for test. Hence our explanatory variables are important covariates to the model.
According to the finding shown in the table above all factors are statistically significant except province which is not because the p-value (0.266) is greater than 0.05.
Holding other factors constant, factors constant health worker reduces the odds of stunting status by 1.21 percent which indicate that health workers have a large contribution in reducing this disease caused by malnutrition after antenatal care which has 1.6.
Other factors like distribution by province also reduce the stunting status by 0.056 ceteric purbus, and place of residence has a contribution in reducing stunting status by 0.57. When holding other factor constant level of education reduce the odds of stunting status by 0.34 while breast feeding increase the odds by 1.17 here is also another factor called wealth index which reduces the odds of stunting status by 0.277 when other factors are hold constant.
Binary logit parameter estimates of model 2
------------------------------------------------------------------------------------
wasting | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------------
Health worker | -.3689097 .1396012 -2.64 0.008 -.642523 -.0952964
province| -.0440565 .046641 -0.94 0.345 -.1354712 .0473582
place residence| -.4262017 .1686871 -2.53 0.012 -.7568224 -.0955811
education level| -.1954709 .108764 -1.80 0.072 -.4086444 .0177026
wealth index | -.2587137 .0792108 -3.27 0.001 -.4139641 -.1034633
breast feeding| .2289394 .1946632 1.18 0.240 -.1525935 .6104723
child sex | .3073024 .1234634 2.49 0.013 .0653185 .5492863
antenatal care| -.6047335 .229367 -2.64 0.008 -1.054285 -.1551824
_cons | .9052192 .3919809 2.31 0.021 .1369506 1.673488
-------------+----------------------------------------------------------------------
LR chi2(8)= 82.22 Prob > chi2=0.0000 Pseudo R2=0.0522 Log likelihood = -746.18373
Source DHS2014-15
The model is:
log(p)= .9052192-.3689097Health worker -.0440565 province-.4262017place residence-.1954709education level-.2587137wealth index+.2289394breast feeding+.3073024child sex-.6047335antenatal care
The likelihood of the model is the probability that you would observe the dichotomous (actually multichotomous) outcomes of the sample, given the coefficient estimates. The logit algorithms maximize the logarithm of this likelihood, and since the probabilities is between 0 and 1, then the log likelihood is always negative. Our maximum log likelihood of obtaining outcomes of health insurance coverage from the sample is -746.18373 and the test chi-square is 82.22.
As the calculated chi-square statistics is highly significant at 5%, then the null hypothesis that the constrained model is correct is rejected. This means that the probability in the upper tail beyond the calculated statistic is smaller than the significance level chosen for test. Hence our explanatory variables are important covariates to the model.
According to the finding shown in the table above all factors are statistically significant except province, level of education and breast feeding which are not because the p-value is greater than 0.05.
Holding other factors constant, factors constant health worker reduces the odds of stunting status by 0.36 which indicate that health workers have a large contribution in reducing this disease caused by malnutrition after antenatal care which has 0.90 when other factors remains constant.
Holding other factors constant, place of residence has a contribution in reducing the odds of wasting status by 0.42, here is also another factor called wealth index which reduces the odds of wasting status by 0.26 when other factors are hold constant. Ceteric paribus also sex of child has a contribution 0.30 in increasing odds of wasting status.
Binary logit parameter estimates of model 3
------------------------------------------------------------------------------------
Underweight Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+--------------------------------------------------------province| .0204013 .0444034 0.46 0.646 -.0666278 .1074305
_cons | -.6274855 .1492666 -4.20 0.000 -.9200427 -.3349283
----------------------------------------------------------------------
Source DHS2014-15
Model is log(p)= -.6274855+.0204013 province
And because the p value of province is greater than 0.05 hence it is not statistically significant.