The dataset was huge and all the predictors were significant in the univariate model except place of residence, therefore we performed the full model analysis. However we first decided to make the final model through the backward elimination method, but since the data was hardly changing with that, we included all the variables in the final model.
Our results highlight that most of the sociodemographic characteristics of a woman (age, literacy, marital status, wealth, ethnicity, region, access to FP information and women empowerment) were significantly associated with the uptake of short term contraceptive which is in line with our hypothesis. These results mostly follow the findings of several studies in the region and all over the world (16–20), however the place of residence did not make a significant difference which was not similar to what was already known in the region (15).
Similar to several studies in the region and other developing countries,(17, 18, 21) our study revealed that age of women was associated with the uptake of contraceptives. In our study, the prevalence of short term contraceptive uptake was higher among women aged between 30–40 years old while women younger than 20 years had the lowest rate. This could be due to several reasons: they are newly married and do not want to avoid pregnancy; these women are less aware of the methods; reproductive health is still a taboo and thus, young women feel ashamed to talk about such issues and to seek FP services.
Similar to other studies conducted in the region (22, 23) our study found a strong association between employment and use of short term contraception. We found that women who were working outside the home, regardless of the type of occupation, were more likely to use contraceptives.
Ethnicity and region also play a role in the usage of contraception. Women living in the western parts of country were more likely to use contraceptives compared to those who are living in the eastern and central parts. Similarly, the prevalence of short term contraceptives is very low among Nooristani women who are living in the eastern parts of the country. The basic conception for this variation could be the geography and security issues which could restricts the provision of health service in these regions.
Access to FP information is an important factor for contraceptive uptake. In the DHS survey women were already considered as knowledgeable only if they were able to name at least one method of FP. Having stated in the DHS survey to know about contraceptives does not necessarily mean that knowledge about its correct application is prevalent nor does it determine actual use. What we looked at in our study was their exposure to FP information which we assumed, increased their knowledge about the methods, the benefits, side effects and further aspects of FP. This finding is similar to other studies which have been conducted in the region and in other developing countries.(24–26) Moreover if we look towards other aspects there could be discrepancies between knowledge and empowerment. The strong cultural anchor of having big families plays an important role which puts women under the pressure to bear as many children as possible and not use contraceptives, even they have the knowledge.
In our study we only looked at media exposure, and due to a huge amount of missing values in the dataset, we did not include other channels (exposure at the health facility and information through community health workers). We found a significant association between media exposure and short term contraceptive uptake and we think that if we would have been able to include other channels into our analysis, the association might have become stronger. In the women empowerment category, we categorized women as empowered if they had answered positive to at least one of those four questions. This was a conservative approach and if we had used stricter criteria, then the effect would have been much stronger and the influence of real empowerment most probably much higher.
Barriers of contraception uptake:
Poor access to contraception, insufficient information about free provision of FP services, fear of side effects, insecurity, cultural and religious disagreements, and gender issues are considered potential barriers for contraception uptake in Afghanistan.
Since we only conducted a quantitative analysis, we identified some of the potential barriers, but there might be some underpinning factors and barriers which need investigation using qualitative research methods with women of reproductive age.
Recommendations
FP is a wide area which requires a multisectoral collaboration and approach to promote in the society. Community based provisions and the use of community health workers (CHWs) to educate both men and women and provide FP services is recommended. Mass media should be used to promote and advocate for FP. Since Afghanistan is a religious country, one of the effective approaches to advocate for the promotion of FP would be to use religious leaders in the society. Inclusion of FP methods as clinical attachments or part of the medical and nursing curriculum, and conducting regular trainings for service providers on proper counseling are also known to be effective approaches for contraceptive promotion. Improving logistics and supply chain management specifically in the eastern parts of the country would be an important action.
We also suggest to conduct qualitative research on non use and discontinuation of methods to find out why women do not use or stop using contraceptives. Conducting FGDs and In-depth interviews not only with women but also with husbands and in laws about different aspects of FP such as expectations of the society, decision making regarding FP, and level of knowledge on contraception, is highly recommended.
The above mentioned recommendations could serve the Afghan policymakers to develop and implement interventions based on empirical evidence which may contribute to the improvement of the health sector, especially the field of mother and child health.
Strengths and limitations
Using DHS data provides us with a trustworthy source and good data quality compared to many single studies. However, in the original study it was assumed that security issues might not have allowed to cover all the selected areas, therefore 101 reserve clusters were selected for replacement and later on 70 clusters, replaced the areas which were identified insecure. At the end the AfDHS was unable to collect data from Zabul province in the south eastern region because it was inaccessible for security reasons. Taking this into account, our results could have looked differently if the initially selected cluster had been used.
The large sample size of 22,974 participants provides a solid basis for statistically robust uni and multivariable analyses. Since the dataset was huge and all the variables were significant, conservatively we used a very strict confidence interval of 99%, which is considered a strength of the study. On the other hand, the cross sectional design of the study restricts the establishment of causal relationships between outcome and exposure. Moreover, some of the potential influencing factors such as number of children, distance from the health facility, number of ANC visits were not included in this study which might have introduced some residual confounding to this analysis. Also, men were not included in this analysis.
Additionally, the dataset had 44 (0.2%) missing values for ethnicity, 19 (0.1%) missing values for occupation, access to FP information had 85 (0,4%) missing values and the women empowerment variable had 866 (3.8%) missing values. Regarding the women empowerment variable, most probably the questions were too sensitive to be answered by all which resulted in information bias. Since these missing values were counting less than 5 % of our sample size, we did the complete case analysis.