Trend and predictors of change of unmet need for family planning among reproductive age women in Ethiopia, based on Ethiopian demographic and health survey from 2005-2016: Multivariate decomposition analysis

Background: Family planning is a key method for reducing population growth and improving maternal and child health by spacing births and preventing unwanted pregnancies. have an unmet need for family planning is dened as women believed to be sexually involved but are not using some form of contraception, either do not want to have more children (Limiting) or want to delay their next birth for at least two years, (Spacing). Methods: The data for this study arrived from the Ethiopia Demographic Health Surveys in 2005, 2011, and 2016 to investigate trends and Predictors of change of unmet need for family planning among reproductive age women in Ethiopia. A pooled weighted sample of 26,230 (7761 in 2005, 9136 in 2011 and 9,333 in 2016 Ethiopian demographic health surveys) reproductive-age women used for this study. For the overall trend (2005-2016) multivariate decomposition analysis for non-linear response outcome was calibrated to identify the factors contributed to the change of unmet need for family planning. The Logit based multivariate decomposition analysis utilizes the output from the logistic regression model to assign the observed change in in unmet need for family planning over time into two components. Stata version 16.0 was used to analysis the data. Result: among reproductive age women in Ethiopia the magnitude of unmet need for family planning decreased from 39.6% in 2005 to 23.6% in 2016. From the decomposition analysis change of unmet need for family planning was due to change in characteristics and coecients. About nine in ten changes in unmet need for family planning was attributable to the difference in coecients. Factors that associated with the change of unmet need for family planning over the last 11 years were educational status, birth order, and desired number of children. Conclusion: Remarkable change in unmet need for FP was observed between the period of 2005 and 2016. Both change in characteristics and coecient were the contributing to observed change. Majority of the change in unmet need for FP was due to difference in coecient over the study period. Mainly the change of unmet need for FP was due to change in women having birth order of ve and above, having secondary education and women who desired number of children below ve.


Introduction Background
Family planning is a key method for reducing population growth and improving maternal and child health by spacing births and preventing unwanted pregnancies (1). Both fecund reproductive-age women who are married and in consensual marriage have an unmet need for family planning and believed to be sexually involved but are not using some form of contraception, either do not want to have more children (Limiting) or want to delay their next birth for at least two years, (Spacing) (2,3).
In 2015, globally unmet need for family planning among married women were 12 percent(4). In a developing country approximately 225 million people had an unmet need for modern contraception. Of this number, 160 million were using no method and 65 million were using a traditional method and an estimated 74 million unintended pregnancies occur annually, 52 million of these premature births could be avoided, saving 70,000 women from pregnancy-related death in 2014. (5) Unmet need for family planning among married or in union women of reproductive age in Sub-Saharan Africa was 25%, posing a major public health concern (6, 7). In Burkina Faso, Malawi, Cameroon, and Ghana the prevalence of unmet need for family planning among reproductive age group women was 18.3%, 21.0%, 46.6%, and 38.9% respectively (8-11). In Ethiopia according to an Ethiopian demographic and health survey, unmet FP needs fell from 37% in 2000 to 22% in 2016 (12).
Respective researchers have examined at various factors that in uence women's unmet need for family planning, such as age, parity, and religion (13), Discussions with partners and wellness extension staff, as well as awareness of contraceptive methods (14), A visit to a health center, media exposure, a husband and wife's educational status, and residence (15), Due to contraception-related factors like availability, accessibility, affordability, and side effects (16), early marriage, wealth index (17), Number of children alive, use of contraceptive methods(18), partner's attitude toward the use of family planning, current menstrual status, healthcare providers visit and discussion about family planning issues (19).
Unmet need for family planning can have serious consequences for women and their families, including unsafe abortion, physical violence, and a high fertility rate linked to poverty and poor maternal and child health (20,21).Though, improving family planning (FP) access is fundamental for sustainable development goal (SDG) achievement. It is linked to human rights, gender equality and women's empowerment and has an impact on maternal, newborn, child and adolescent health.
Different researchers in Ethiopia have identi ed the prevalence and determinant factors of unmet family planning needs., as far as our deep literature reviews, studies concerning on the trend and identify the contributing factors for the change in unmet need for FP in Ethiopia over the studying period among reproductive age group women are limited.
Using multivariate decomposition analysis to identify what socio demographic predictors are strongly correlated with the change in unmet need for FP among reproductive age women(15-49 years) is important to target on factors that decrease/increase unmet need for family planning and to help policy and programs development that focus on reducing unmet need for family planning in Ethiopia. Therefore, this study is aimed to address the trends and contributing factors for change in unmet need for family planning over time by using multivariate decomposition analysis based on the 2005-2016 Ethiopian Demographic and Health Survey (EDHS).

Method And Materials
Study design and sampling procedures The data for this study was from 2005, 2011, 2016 Ethiopian Demographic Health Surveys (EDHS) to investigate trends and Predictors of change of unmet need for family planning among reproductive age women in Ethiopia. In each of the surveys, a two-stage cluster sampling was employed. In the rst stage, 540 Enumeration Areas (EAs) in EDHS 2005, 624 EAs for EDHS 2011, and 645 EAs in EDHS 2016 were randomly selected proportional to their EA size and, on average, 27 to 32 households per EAs were selected in the second stage. A pooled weighted sample of 26,230 (7,761 in EDHS 2005, 9,136 in EDHS 2011 and 9,333 in EDHS 2016) reproductive-age women was included for this study. The detailed information about sampling procedures was presented in the EDHS report (12,22).

Study Variables Outcome variable
The outcome variable was unmet need for FP, where it composed of unmet need for spacing and limiting.
It refers to the proportion of women who desire to either delay the next pregnancy or limit future pregnancies but are not using any method of modern method of contraception (11). The outcome variable was categorized as "unmet need" if women had unmet need for spacing and limiting were and coded as 1, while those using FP methods for spacing or limiting or with no unmet need were "met need" coded as 0.

Independent variables
The independent variables included in this study were: respondent's age, respondent's educational status, religion, husband's education status, marital status, place of residence, women working status, husband working status, wealth status, media exposure, termination of pregnancy, knowledge about family planning, visited health facility last 12 months, visited by eld worker in the last 12 months, perceived distance to health facility, age at rst marriage, birth order, sex of household head, region and desired number of children.

Statistical Analysis
Important variables were extracted from the Individual Record (IR) datasets. Data were weighted using "svyset" STATA command and it was applied for descriptive analysis. The variables required for the "svyset" is the weight variable (v005), primary sampling unit (v021), and strata (v023). Trend analysis of unmet need for family planning and decomposition of the change in the prevalence of unmet need for family planning over time was done .The trend analysis has been done by separating based on time For the overall trend (2005-2016) multivariate decomposition analysis for non-linear response outcome was calibrated to identify the factors contributed to the change of unmet need for family planning across the two surveys. For our study, Logit based decomposition analysis was employed. The Logit based multivariate decomposition analysis utilizes the output from the logistic regression model to assign the observed change in in unmet need for family planning over time into components.
For our study, the 2016 EDHS data was appended to the 2005 EDHS data using the "append" Stata command, and the Logit based multivariate decomposition analysis (using mvdcmp STATA command) was used to identify factors that contributed to the change in unmet need for family planning over the last 11 years. The change in unmet need for family planning can be explained by the compositional difference between surveys (i.e. differences in characteristics) and/or the difference in effects of explanatory variables (i.e. differences in the coe cients) between the surveys.
Hence, the observed decrease in unmet need over time is additively decomposed into a compositional difference of respondents of each survey (endowments) component and a coe cient (or effects of characteristics) component.
For logistic regression, the Logit or log-odd of unmet need for family planning is taken as: = X indicates independent variables (unmet need for FP in this study) β denotes that, regression coe cient of each selected explanatory variables The E component refers to the part of the differential owing to differences in endowments or characteristics. The C component refers to that part of the differential attributable to differences in coe cients or effects.

Ethical approval and consent
Authors have requested DHS Program through an online request by written letter of objective and signi cance of the study. Permission for data access was granted to download and use the data from http://www.dhsprogram.com . The EDHS programs permitted data access, and data were used for only the current study.

Result
Characteristics of the study population were the dominant percentage of women across the three successive surveys. Across the three successive surveys, there was a clear trend of decline in percentage of unmet need for spacing (from 25.5% to 14.34%) by 9.16 point percentage and for limiting (from 16.05% to 9.22%) by 6.83% point values.
Regarding educational status of the study participants, women with no education decreased by 17.2% in 11 years. However, the number of women with primary school and those with high school and above increased by 12.7 percent and 4.6 percent, respectively. Percentage of orthodox Christian declined by 1.   The characteristics with largest effect among the individual characteristics affecting change in unmet need for FP between 2005 and 2016 was perceived distance from health facility. This means that the decline in women who perceives distance from health facility as not big problem accounts for 6.4 percent rise in unmet need for FP. Second largest characteristics effect on the observed change in unmet need was due to decrease of women who have higher number of children ( ve and above) i.e. explains about 5% increase in unmet need. Similarly, 2.5 percent increment in unmet need for family planning was due to decrease in the composition of women who reside in rural areas. (Table 4)   with previous study done in Ethiopia and Pakistan (24,25). For the decrement of unmet need over the last 11 years in Ethiopia, the expansion of health extension programs to all over the country take its lion's share, over the last decade Ethiopia showed improvement in accessibility and availability of health facilities and There is also documented evidence of improvement of women's autonomy in making decision for their own health (26,27).
The results of this study suggest that rural residents (table2) have had superior decline in unmet need for family planning compared to urban residents. This may be due to government dedication to raising rural community awareness related to maternal and child health and provision of healthcare facilities over the last decades.
From the decomposition analysis decrease in composition of women who reported distance to health facility is not big problem, who had ve and more children, being rural dweller and women from rich wealth status households contributes to compositional rise of unmet need for family planning. Decrease in composition of women of who perceived distance from health facility is not big problem shows signi cant effect on rise of unmet need for FP over the entire study periods (2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016). This may be attributed to Ethiopia's improvement in health facilities physical accessibility and affordability over the recent decades. In addition, specially trained community health workers (health extension workers) and accessibility of health post to nearby community had their own contribution in enhancement of maternal health services such as family planning (28).
Similarly, decreased composition of women who had ve and above birth order results a signi cant impact on raise of unmet need for FP compared to women with only on birth order. Women with high number of children are more likely to face unmet need for FP Because of due to having too many children unlikely to fear of child death and they think of themselves as reached the planned level of fertility.
From 2005 to 2016 survey the compositional decrement of rural residence raised the unmet need for FP by 2.5% point percentage relative to urban resident. This can explained by speedy urbanization over the last decades (29). Urbanization was important in enhancing access to health facility and have a higher knowledge of maternal health service use than rural residents.
From change due to coe cient differences between the two surveys having secondary education, having many children (highest birth order) and ideal number of desired children were signi cantly associated with the change of unmet need for FP. About 28% decrease in unmet need for FP was attributable to women having 5 and above birth order. This nding is consistent with study done in Ethiopia (30).
Women who desired to have less than ve number of children positively contributed to the rise of unmet need for family planning by 20% compared to women who desired greater than ve number of children, a result that is consistent with other recent Ethiopian study (29). This can be explained by; women who desired to have less than ve number of children more likely to face challenges related to unmet need for FP to limit their number of children below the desired number.
Other nding of this study was having secondary education contributed to decrease unmet need for FP by 5.5% point percentage, similar to what had been documented in other Ethiopian and Kenyan studies (31,32). Women with a secondary education may have more access to knowledge about family planning, or formal education may have allowed them to better understanding of contraception (33).
These educated women were also more likely to become independent decision maker to use of family planning (34). The ndings of the study may inform maternal health programmers to strengthen home visit by health care workers to improve family planning uptake. Even though, the authors have compared three large data sets to show the trend over the time and associated differences in the women's characteristics, this analysis did not consider other signi cant predictors of unmet need for FP (cultural, clinical and other factors) which were not collected by EDHS program.

Conclusion
Remarkable change in unmet need for FP was observed between the period of 2005 and 2016. Both change in characteristics and coe cient were the contributing elements to the observed change in unmet need for FP. Majority of the change in unmet need for FP was due to difference in coe cient over the study period. Mainly the change of unmet need for FP was due to change in women having birth order of ve and above, having secondary education and women who desired number of children below ve.
Empowering uneducated women about maternal health services speci cally about family planning is required. The government and any concerned body could be better to focus on the enhancement of house hold economic status and health facility accessibility.
Declarations Figure 1 The trend in rate of unmet need for family planning among reproductive age women in Ethiopia in the past 11 years.