7.5 ASHA Interaction and Overall Uptake of Biomedical Behavior

We have looked at a series of health related behaviors. Here, we want to see if mothers who interacted with ASHAs more often were more likely to engage in healthy behavior. To do this we look at the number of the biomedically recommended behaviors that each woman engaged in (meaning that their ‘yes’ and ‘no’ responses were consistent with biomedical recommendations for the 11 behaviors we looked at so far). We then take the count of the number of behaviors a woman engages in the recommended direction and examine the correlation between this count at the ASHA interaction score.

Here is the resulting figure (the points in the plot are jitter-ed and semi-translucent to make them more visible):

The relationship between mother health score and ASHA interaction score.

Figure 7.10: The relationship between mother health score and ASHA interaction score.

The slope of the above line is is 0.020 and is significant by conventional standards (generalized linear model with negative binomial because the response is count data and it is under-dispersed, P < 2e-16). The figure shows that increased ASHA interaction leads to higher uptake of health behavior.

7.5.1 Component parts of ASHA Interaction Score

How many women are having limited contact with the ASHA?

7.5.2 Factors Predicting Greater ASHA Interaction Scores

Increased interaction with ASHAs leads to greater uptake of recommended behaviors. Here we are interested in individual-level characteristics that may affect the frequency of these interactions.

What characteristics are linked to more ASHA interaction?

Table 7.8: Modeling factors that might correlate with ASHA Interaction Score: age, age of marriage, parity, and education
expEst Estimate Std. Error t value Pr(>|t|)
(Intercept) 8.179 2.102 0.073 28.928 0.000
ageclass21-24 1.045 0.044 0.059 0.736 0.462
ageclass25-28 1.141 0.132 0.075 1.757 0.079
ageclass29-33 1.165 0.153 0.093 1.639 0.101
ageclass34+ 1.167 0.155 0.123 1.253 0.210
ageclass_married15-17 1.113 0.107 0.063 1.684 0.092
ageclass_married18-20 1.141 0.132 0.068 1.952 0.051
ageclass_married21+ 1.029 0.028 0.114 0.248 0.804
nkidscat2 0.956 -0.046 0.059 -0.767 0.443
nkidscat3 0.988 -0.012 0.071 -0.171 0.864
nkidscat4 1.023 0.023 0.085 0.271 0.787
nkidscat5+ 0.896 -0.109 0.103 -1.057 0.291
m_educat1to7 1.066 0.064 0.061 1.044 0.297
m_educat8to10 1.031 0.031 0.052 0.600 0.548
m_educat11to13 1.057 0.055 0.076 0.720 0.472
m_educat14to17 1.184 0.169 0.089 1.905 0.057

In this first model (Table 7.8), we see that women with intermediate ages-at-marriage (15 - 17 and 18 - 20) had higher ASHA interaction scores than women with the earliest ages of marriage (10 - 14). Women in the highest education category (14 - 17 years) had more ASHA interaction than those in the lowest category (0 years).

Table 7.9: Modeling factors that might correlate with ASHA Interaction Score: age of marriage, education, and caste
expEst Estimate Std. Error t value Pr(>|t|)
(Intercept) 8.288 2.115 0.096 22.024 0.000
ageclass_married15-17 1.119 0.112 0.063 1.786 0.074
ageclass_married18-20 1.156 0.145 0.066 2.192 0.029
ageclass_married21+ 1.077 0.074 0.111 0.667 0.505
m_educat1to7 1.049 0.048 0.061 0.782 0.434
m_educat8to10 0.992 -0.008 0.050 -0.152 0.879
m_educat11to13 1.006 0.006 0.074 0.078 0.938
m_educat14to17 1.158 0.147 0.087 1.679 0.093
castefSCHEDULED CASTE 1.013 0.013 0.085 0.157 0.875
castefSCHEDULED TRIBE 1.335 0.289 0.155 1.861 0.063
castefOBC 1.058 0.057 0.080 0.708 0.479

In this second model we change the control variables in the model and include caste (Table 7.9). The same impression regarding age at marriage and education is present. There is also an indication of a somewhat unexpected result: that women in scheduled tribe may have slightly greater ASHA interaction scores than those in General.

Table 7.10: Modeling factors that might correlate with ASHA Interaction Score: age of marriage, education, and caste
expEst Estimate Std. Error t value Pr(>|t|)
(Intercept) 8.511 2.141 0.057 37.843 0.000
ageclass_married15-17 1.116 0.110 0.063 1.752 0.080
ageclass_married18-20 1.155 0.145 0.065 2.211 0.027
ageclass_married21+ 1.062 0.060 0.108 0.554 0.580
m_educat1to7 1.044 0.043 0.060 0.716 0.474
m_educat8to10 1.004 0.004 0.050 0.084 0.933
m_educat11to13 1.021 0.021 0.073 0.287 0.774
m_educat14to17 1.172 0.159 0.087 1.824 0.068
religfrecodeMUSLIM 1.115 0.109 0.056 1.933 0.053
religfrecodeOTHER 1.154 0.144 0.216 0.664 0.507

In the next model we replace caste with religion and find that Muslim women may have slightly higher ASHA interaction scores than Hindu women.

Table 7.11: Modeling factors that might correlate with ASHA Interaction Score: age, age at marriage, caste, religion, education, and wealth
expEst Estimate Std. Error t value Pr(>|t|)
(Intercept) 7.760 2.049 0.114 17.962 0.000
ageclass21-24 1.062 0.060 0.059 1.016 0.310
ageclass25-28 1.167 0.154 0.075 2.070 0.039
ageclass29-33 1.197 0.180 0.093 1.937 0.053
ageclass34+ 1.197 0.179 0.123 1.459 0.145
ageclass_married15-17 1.106 0.101 0.063 1.608 0.108
ageclass_married18-20 1.135 0.127 0.067 1.878 0.061
ageclass_married21+ 1.051 0.050 0.116 0.433 0.665
nkidscat2 0.953 -0.048 0.059 -0.825 0.410
nkidscat3 0.975 -0.026 0.070 -0.366 0.715
nkidscat4 1.003 0.003 0.084 0.041 0.967
nkidscat5+ 0.870 -0.140 0.103 -1.356 0.175
castefSCHEDULED CASTE 1.041 0.040 0.088 0.453 0.651
castefSCHEDULED TRIBE 1.377 0.320 0.157 2.041 0.041
castefOBC 1.067 0.064 0.080 0.809 0.419
religfrecodeMUSLIM 1.139 0.130 0.059 2.195 0.028
religfrecodeOTHER 1.127 0.119 0.216 0.551 0.582
m_educat1to7 1.078 0.075 0.061 1.223 0.221
m_educat8to10 1.059 0.057 0.054 1.061 0.289
m_educat11to13 1.112 0.107 0.080 1.331 0.184
m_educat14to17 1.247 0.221 0.092 2.404 0.016
wealthquintilesSecond 0.938 -0.064 0.061 -1.046 0.296
wealthquintilesMiddle 0.978 -0.022 0.063 -0.351 0.726
wealthquintilesFourth 1.029 0.029 0.065 0.440 0.660
wealthquintilesRichest 0.862 -0.148 0.073 -2.044 0.041

Lastly, a final model for this section includes wealth, caste, and relgion, along with the previous controls (Table 7.11). This model reproduces previous impressions for age at marriage, Scheduled Tribe, being Muslim, and education, and also suggests that the wealthiest category interacts with ASHAs less than the lowest wealth category.

Identifying precisely what predicts greater levels of ASHA interaction will need more investigation, but from this exploratory exercise we have a few general impressions. Intermediate levels of age and age at marriage lead to more ASHA interaction. There is some indication that being in the wealthiest quintile reduces ASHA interaction and that being in the most educated subgroup leads to more ASHA interaction. Muslim women have greater ASHA interaction scores than Hindu women (note, the ethnography suggested that some ASHAs may look down on majority Muslim villages or areas).

Take-home message

  • The ASHA interaction score is a sum that includes the number of visits reported in each trimester and post partum, and each time a woman references an ASHA as being responsible for a service in a followup question (Table ??).

  • The score is used as a measure of ASHA interaction to see if more interactions are correlated with adherence to recommended health behaviors as well as to see if some factors, such as differences between ASHA and beneficiary, limit interactions.

  • The more years of experience ASHAs have the more respected they are within the community. However, it is important to note that ASHAs begin at a high level of initial respect in their first years on the job, which potentially refers to the community’s generally high regard for and acceptance of the ASHA program (apart from an ASHA’s individual efforts throughout their years at the job)

  • Many mothers do not receive any postnatal visits by the ASHAs despite incentives for these visits. This seems like a potential area of concern, as a majority of newborn and maternal morbidity and mortality occur during this period, which requires a higher level of care and attention by the ASHAs. It is also a missed opportunity because it is an important life event marked with several rituals and traditional practices that ASHAs can augment with health advice for higher impact.