When assessing targeting for vitamin A programs, usually it will include monitoring the coverage of high dose vitamin A capsule distribution (supplementation) to children 6 months to 5 years of age and women post-partum. It was the conculsion in a review by Mason and Gillespie in 1994 that vitamin A control through supplementation only reaches effectiveness when coverage is at least 65%, although this is rarely achieved, nor feasible. Through targeting the supplementation and securing other means of vitamin A intake and maintenance (through fortification and public health measures) the likelihood of reaching a low risk level for the population is greater.  In the following exercises, both program coverage and re-targeting will be addressed using national level data sets from Nepal (nepdr.sav)

To take a look at the association of the coverage of children with vitamin A capsules by the level of nightblindness in children, use the Crosstabs routine under Statistics, Summarize. (First, the variables will have to be categorized by choosing cut-points for high and low).

These results were obtained through a series of steps including re-categorization of the outcome and independent variables into high/low. The decisions for cut-points were made somewhat arbitrarily, as they often are but usually it is best to consider current recommendations for coverage and distribution of the data with which you are working. Here, in the Nepal data the high and low nightblindness levels were chosen using current WHO recommendations of 1% in preschool children (representing a moderate public health problem). The coverage of vitamin A capsules was dichotomized using a cut-off lower than the recommended level by Mason and Gillespie (50% was used vs. 65%).   As seen above, there are 4 of the 13 districts with high levels of nightblindness in children <5 years. In all four of these development regions, the service delivery of vitamin A capsules is very low (<50%, ranging from 0% to 41%).  Of all of the development regions, only one has vitamin A cap distribution measured for even close to the 65% recommended cut-off (yet still only 61%).  Using the information above, it would be wise to re-target some of the services to those four districts that have low vitamin A capsule distribution and high levels of nightblindness.  These areas are in immediate need of intervention.

Here is a graphical representation of the targeting of vitamin A distribution based on levels of nightblindness.

The blue bar represents the WARNING group, those that need re-targeted services.   The children represented by this blue area have alarming levels of vitamin A services and little or no intervention (supplementation).  It might be possible to also re-evaluate what the current risk is to the 2 districts that have low levels (actually 0% nightblindness) and try to re-distribute some of the "unneeded" intervention to those areas with greater immediate need.  To identify which areas are where on the scale of deficiency and intervention, it is possible to run a scatterplot using a labeling technique for the development regions.  Try running the nightblindness against vitamin A capsule distrubution using this routine and identify those four "needy" development regions.

As you can see using this technique, the four districts in greatest need are Terai Farwestern, Terai Midwestern, Terai Eastern, and Hill Western.  The two areas of lesser need with higher distribution of vitamin A capsules are Hill Midwestern and Terai Western.  This same result can be achieved using the ranking, which we will show now incorporating other indicators of risk, such as measles vaccination coverage and general malnutrition.

Another approach to targeting might be to look at the "potential" to reach the targeted recipients for a supported health measure.   For example, with vitamin A supplementation it is now suggested that mothers receive high dose vitamin A after birth (get reference and exact time period recommended).   If this policy is not currently supported but nightblindness in mothers is high (which it is in this data from Nepal), then you might want to see how easy it would be to access mothers post-partum to distrubute vitamin A so that the new born infant (through breastmilk) and mother (through the supplement) are replenished with vitamin A. (The problem with supplementing during pregnancy the risk of teratogenic effects of high doses of vitamin A on a fetus).   Run the cross-tabulation to show what percent of mothers do already access post-natal care against the variable for nightblindness in pregnancy.  Look first at a scatterplot of the variables of interest to get a sense of what the association is, what the ranges are, and what areas fall where (use the labeling option).  Use Graphs, Scatterplot with the nightblind in pregnancy (nbpreg) and postnatal care (postnata) as well as nightblind in pregnancy (nbpreg) against antenatal care (anc).  The scatterplot will be best viewed using the development region level data (nepdr.sav).

It seems that there is not high (in general terms) access to post-natal care for any area.  The highest areas are Terai Midwestern and Central, but they still have <30% using postnatal care.  When considering antenatal care, the usage is higher therefore indicating the mothers do at least have access to and use care during the time of pregnancy.

Look at the scatterplot of antenatal care and nightblindness in pregnancy here.   Try running this routine also.

Here again the results show that there is a negative association between nightblindness in pregnancy and use of health care (antenatal care in this case).  There is an overall higher percent that use the antenatal care (41%) than those who use postnatal care (13%).  For now, it appears that if you were trying to improve VAD in moms and children by providing vitamin A post-partum through health services, then it might be difficult because those women who are the most vitamin A deficient are not using the services.  Before making this conclusion though, take a look at a the crosstabulation of nightblindness in mothers by antenatal care and postnatal care using the nepal individual level data.

Looking at these results you can also find the population prevalence (for pregnant women) of nightblindness using the formula {(a+c)/ (a+b+c+d)}
postnatal care:  103/583= 17.7%
antenatal care:  103/579= 17.8%

You can also find the population coverage of postnatal and antenatal care for women who are pregnant using the formula {(a+b)/ (a+b+c+d)}
postnatal care: 170/583= 29.2%
antenatal care: 241/579 = 41.6%

Now you can find the coverage of "malnourished" moms (those with VAD) using the formula {a/ (a+c)}
postnatal care:  28/103= 27.2%
antenatal care:  44/103= 42.8%

To look at how well the antenatal and prenatal care are focused on the mothers with VAD, use this formula: {a/ (a+b)}
postnatal care:  28/170= 16.47%
antenatal care:  44/241= 18.26%

To see if the programs are well targeted, use the formula  F (focusing)/ PP (population prevalence):
postnatal care:  0.9305 (could use some re-targeting suggestions)
antenatal care:  1.026 (good targeting)