Pregnancy Intentions and Its Associated Factors Among Married Women in Resunga, Gulmi, Nepal

Pregnancy Pregnancy intentions can be categorized as and It is an important public health in every corner of the in developing countries. The study was carried out with an objective of determining the pregnancy intentions among married women in Resunga municipality, Gulmi, Nepal and test the association of pregnancy intentions with various socio-demographic, and reproductive health-related variables. The community based cross-sectional study design was used to achieve the objective of the study. Married women of reproductive age of the Resunga municipality, Gulmi were selected and a multistage random sampling technique was adopted. Total 406 participants were involved in this study. Pretested semi-structured interview schedule adapted from Demographic and Health Survey of Nepal was used for data collection. All the statistical tests were performed in SPSS, version 25.0. The p -value was set at 5% level of signicance. Descriptive and data exploratory analysis were performed. Bivariate and multivariate logistic regression analysis technique were conducted to explore the association between pregnancy intentions and predictor variables.

causing a socioeconomic burden to individuals and society, ultimately decreasing the quality of life and workforce e ciency. Therefore, it must be addressed adequately on time [11]. Unintended pregnancy is a contributing factor for maternal and infant mortality, which is still high in Nepal. To reduce maternal and infant morbidity, mortality, and the need for abortion, different types of programs, strategies, plans, and policies should be planned with a focus on the high-risk group. Therefore, it is crucial to identify and understand the factors in uencing unintended pregnancy, which is essential to maintain and improve the overall health and wellbeing of women, family, society, and nation [16]. This study validated this information. Similarly, very few studies have been carried out in Nepal regarding pregnancy intentions, with studies studying factors associated with pregnancy intentions taking primary data almost nonexistent. No study has been found on this topic in a similar area of the county. Considering the high unintended pregnancy rate, the proposed study has been forwarded to study the factors associated with it. It is expected that the ndings of this study would be useful for academic as well as health management purposes at a different level.
The main aims of this study were to determine pregnancy intentions and assess associated factors among married women in Resunga Municipality, Gulmi, Nepal. We estimated the prevalence of intended and unintended pregnancy, analyzed factors associated with unintended pregnancy, identi ed spousal communication on contraception between partners for pregnancy planning and identi ed autonomy to choose contraception.

Methods
This is a community-based, cross-sectional study. The study employed a quantitative paradigm with the aim of estimating pregnancy intentions and their association with different factors. The sociodemographic and reproductive health variables are shown in Fig. 1.
Participants DHS tries to assess the level of unwanted fertility rates by asking a chain of questions among the sample of current pregnancies and births in the ve years preceding the survey to women aged 15-49. This study included married women who became pregnant within the last year, including current pregnancy, as the sampling unit to reduce recall bias to some extent. Women of reproductive age (15-49 years) of the Resunga Municipality, Gulmi, were the study population. The sampling frame was created by preparing the list of all households having women who became pregnant within last year, including current pregnancy, with the help of the Female Community Health Volunteer (FCHV) register, Vitamin A distribution register of respective wards, Antenatal Care (ANC), Immunization register, and Maternal and Child Health (MCH) register.
A multistage random sampling technique (probability sampling) was used to select samples of preferred size, as shown in Fig. 2. The Resunga Municipality of Gulmi district was chosen purposively at rst, and then 11 wards out of 14 wards were selected using a simple random sampling technique (lottery method). After selecting wards, the required number of women from each ward was calculated using the population proportionate to size technique. A total of 410 married women who became pregnant within the last year, including current pregnancy, were selected from all selected wards by the simple random technique (Lottery).

Data Collection
The study was conducted in Resunga Municipality, Gulmi. Gulmi is a hill district of the Western development region of Nepal and one among the twelve districts of province ve. It has been politically divided into two municipalities (Resunga and Musikot) and ten rural municipalities. The districts cover 1,149 square kilometers (444 sq. mi). The Census of 2011 identi ed 64,921 households with a total population of 280,160 and a female population of 159,165 in the district. The district headquarters of Gulmi is Tamghas, which is located 208 kilometers (129.8 miles) west of Kathmandu. A populationbased study using national survey data concluded that women from the hill region were more likely to report mistimed pregnancies than women from other regions, and Gulmi is also in the hill region. The data was collected from ward numbers 1,2,4,6,7,8,9,10,11,12, and 13 wards of Resunga Municipality.
We collected data over a period of 1.5 months (October 2-November 18, 2018 AD) to achieve a sample size of 410 using the pretested semistructured interview schedule. We prepared a consent form and semistructured interview schedule. The questionnaire was pretested in a similar community that was different from the study site, taking 10% of the sample size to remove ambiguity in answering with the purpose of easy administration. The questionnaire was translated into Nepali and then back translated to English to ensure accuracy. The English version of the questionnaire is attached as supplementary le.
The study enquires about sociodemographic characteristics, household economic status, reproductive health and autonomy of women, which was developed based on a standard questionnaire of the Nepal Demographic Health Survey 2011 (NDHS) with some modi cation to the local context. The schedule was prepared in the Nepali language for the actual eld study. Pregnancy intentions can be measured by using a measure of the National Survey of Family Growth (NSFG), which is also known as a conventional measure of pregnancy intentions and is most used in the United States. Recognizing the need for a more re ned measure of NSFG has expanded its conventional measure, which is known as an alternative measure of pregnancy intentions. Another measure is the London Measure of unplanned pregnancy (LMUP) tool, which is validated and used in multiple settings, including South India. This LMUP tool had been modi ed to make it contextual as per study design and was used taking consent from the authority.

Data Analysis
The collected data were edited, organized, recorded and inserted into the latest data entry program prepared by the researcher in EpiData following coding. The compiled data were exported and analyzed using the latest Statistical Package for Social Science (SPSS) software, version 25, IBM, Chicago, United States. Then, the data were analyzed using both descriptive and inferential statistics.
Data analysis was performed in three stages. In the rst stage, descriptive analysis was carried out, and frequency and percentage were presented, and the mean, median and standard deviation were calculated for continuous variables as per the need. Principal component analysis was used to generate the wealth index of families using household assets.
The socioeconomic status of the surveyed families was measured using the wealth index. In determining the wealth index of the families covered in the study, the presence or absence of a list of principle elements were taken into consideration and were determined by the method commonly known as principal component analysis (factor analysis) and divided into ve quintiles of wealth. The rst quintile represented the poorest segment of the population, and the fth quintile represented the least poor segment.
The second stage of analysis involved testing the association between various independent variables and pregnancy intentions. Chi-square statistics and p-values at the 95% level of con dence was conducted. To maintain the minimum sample size in every cell for bivariate analysis, necessary consideration was made while categorizing the response of the participants. Cross-tabulation was performed for each variable of interest with pregnancy intentions.
The third stage involved logistic regression analysis between dependent and independent variables. Both bivariate and multivariate analyses were performed to determine the existence of statistically signi cant associations and strength of association between the dependent and independent variables of the study.
First, bivariate logistic regression was performed followed by multivariate logistic regression based on a binary logistic regression model to adjust for the effects of other variables within the model, control for possible confounders, and test the strength of any association noticed in the bivariate analysis. Those independent variables that were found to be signi cant in the bivariate analysis and those variables that were important but not found to be associated in this study were also included in the multivariate analysis based on previous studies. For this study, a p-value of 0.05 was considered signi cant. Hosmer and Lemeshow's goodness of t test was carried out to ensure that the model was t. The model was t for our data as shown by the Hosmer and Lemeshow test, and the test statistic was 0.929 (p 0.05).
Pregnancy intentions was the main outcome variable, which was measured using the London Measure of Unplanned Pregnancy. LMUP is a psychometrically validated measure of the degree of intentions of current and recent pregnancy that is designed to measure both pregnancy planning and intentions. It is validated in a diverse setting and is in increasing use as a research tool. LMUP includes six questions, each scored 0, 1, or 2. Then, the score is summed to create an ordinal variable on a scale of 0-12. The > score can be interpreted as 0-3, unplanned; 4-9, ambivalent; 10-12, planned re ecting increasing pregnancy intention with increasing each score [25]. Independent variables were compiled in the heading of socio-demographic characteristics and reproductive health-related variables. Sociodemographic variables included maternal age, age at marriage, ethnicity, household size, education, occupation, and socioeconomic status. Reproductive health-related variables include gravida, parity, history of abortion, age of youngest child, age at rst child, ideal number of children, use of contraception, spousal communication, and autonomy of woman.

Ethical Considerations
This research was approved by the Institutional Review Committee (IRC) of Institute of Medicine, Tribhuvan University (Ref: 108(6-11-E)/075/076). Formal permission was taken from the concerned authority (DPHO and Municipality) of Gulmi district. The interview was conducted only after obtaining informed verbal and written consent from the participants after explaining the study's aim, objectives, bene ts, and con dentiality. In cases of illiterate participants, consent was obtained by a thumbprint on the consent form. Study participants had the right to withdraw from the study at any time. They were told that they could decline the questions if they thought were sensitive and distressing to them. Con dentiality of each of the participants was maintained, and no personal identi er was included in the analysis. Coding and aggregate reporting were used to eliminate participants' identi cation and ensure anonymity. Respondents willing to know more about family planning were provided with some information at the end of the interview.

Results
A total of 410 married women who became pregnant within the last year, including current pregnancy, were enrolled in this study. The descriptions of sociodemographic and reproductive health-related variables are presented in Table 1. The scores of London Measure of Unplanned Pregnancy (LMUP) is demonstrated in Table 2.   Table 3. The highest proportion of women (87.6%) were from the age group 20-34 years of age. The youngest (< 20 years) respondents accounted for 10.5 percent of the total respondents. More than 77 percent of the women were married before and at the age of 20 years. Furthermore, most of the respondents were multigravida (54.9%; n = 410), and para 2 (51.3%; n = 232) aborted pregnancy at least once (80.4%; n = 51). The mean age of the women at her rst child was 20.74 (SD = 2.714 years). The majority (53.4%; n = 232) of the women's rst childbearing age was 20 years and less, followed by 20 years above (46.6%; n = 232). The study revealed that the majority (94.6%; n = 410) of the women had two and fewer children as an ideal number of children, and the remaining (5.4%; n = 410) expressed more than two children as the preferred number of children. It was interesting to see that approximately 32.0 percent (n = 410) of the women's husbands oppose family planning use. According to this study, 12.2 percent (n = 410) of the women never discussed family planning with their husbands before using contraception. Approximately 31.0 percent (n = 410) of the respondents reported that they were using any kind of contraceptive method to prevent pregnancy, with the majority of modern methods users (74.0%; n = 127). One in four women adopted natural methods (26.0%; n = 127) for the prevention of pregnancy before the current pregnancy.

Bivariate analysis
The bivariate analysis is shown in Table 4. The chi-square analysis showed that female age, socioeconomic status, husband opposition, methods used before pregnancy, and intention to use contraception in the future were signi cantly associated with pregnancy intentions. The analysis showed that marriage age, ethnicity, family size, women's education, spousal education, women's occupation, spousal occupation, and other reproductive health-related characteristics and level of autonomy were not signi cantly associated with the pregnancy intention of women of reproductive age group despite the fact that their signi cance has been reported in other studies. The proportion of unintended pregnancies was highest among women younger than 20 years (79.1%) compared to women aged 20 and older (58.3%). However, intended pregnancy was highest among women aged 20 years and older (41.7%). These differences in pregnancy intentions by age group were statistically signi cant with a p-value of 0.008 in the chi-square analysis. However, the association became statistically nonsigni cant when adjusted for other variables in multivariate analysis. The proportion of unintended pregnancies was highest among the women who were married before 20 years (62.1%) of age compared to women who were married after 20 years (54.8%). In contrast, the proportion of intended pregnancies was highest among women who were married after 20 years of age (45.2%).
Most of the unintended pregnancies were from a relatively disadvantaged (64.9%) ethnic group, followed by a disadvantaged (64.1%) and relatively advantageous (56.7%) ethnic group. Most of the unintended pregnancies were among women who belonged to the rst wealth quintile (74.4%), and intended pregnancies were more common among women who belonged to the third wealth quintile (56.5%). It is also statistically signi cant in chi-square analysis with p-value 0.001.
Unintended pregnancies were high among women who were Primi (62.9%), had more than two para (73.3%), had done abortion (60.8%), and had done an abortion at least one time (61.0%) nonspontaneously (65.0%). More than 60 percent of unintended pregnancies were among women whose youngest child was more than one year and who delivered her rst child before or at 20 years of age (60.5%). Most of the unintended pregnancies were among women who preferred less than or equal to 2 children (60.6%) as an ideal number of the child. Approximately 58.0 percent of unintended pregnancies were among women who did not discuss with her husband before using contraception. There were 69.5 percent of unintended pregnancies among women whose husbands opposed using contraception. Other factors signi cantly associated with pregnancy intentions, as shown by bivariate analysis, were husband opposition in using contraception (p = 0.011), method of family planning used before pregnancy (p = 0.002) and women's intention to use contraception in the future (p = 0.011).

Multivariate binary logistic regression
The multivariate binary logistic regression is shown in Table 5. In the multivariate logistic regression model, after possible effects of confounders were adjusted, the method of family planning used before pregnancy and wealth quintiles as the independent variable were signi cantly associated with pregnancy intentions at 95% CI. Women from the second (AOR = 0.048, CI = 0.007-0.329) and third (AOR = 0.134, CI = 0.026-0.702) wealth quintiles had a lower chance of having an unintended pregnancy in comparison to women from the fth wealth quintile. Women who used the natural family planning method before pregnancy had a lower chance of having an unintended pregnancy than women who used the modern family planning method before pregnancy (AOR = 0.078, CI: 0.016-0.390). However, husband opposition to use contraception, parity, the ideal number of children, spousal communication, level of autonomy, and future intention of using contraception were not depicted to be associated with pregnancy intentions as per this study. The analysis was t in a logistic regression model: y = β 0 +β 1 X 1 +β 2 X 2 + β 3 X 3 ………. + β i X i where y is the log odds of the dependent variable, β 0 is a constant, β i is the regression coe cient and Xi is an independent variable. The logistic regression revealed the following equation for pregnancy intentions: pregnancies were considered intended. This nding is consistent with the estimated national prevalence of unintended pregnancy by NDHS 2016 [24]. A hospital-based cross-sectional survey conducted in Pakistan in 2015 adopting the LMUP tool revealed 38.3% unintended pregnancies, of which 13.9 percent were ambivalent and 24.3 percent were unplanned [26]. Similarly, a study carried out on the basis of the NDHS 2011 dataset in 2015 showed that the unintended pregnancy rate was 24.59% [6]. In 2012, 40 percent of pregnancies were unintended globally [9]. A community-based cross-sectional study carried out in Ethiopia in 2013 showed 36.5 percent unintended pregnancy [13]. The estimate of this study is higher than that of previously reported data, which might be because previous studies used a dichotomous scale, whereas this study employed six-item LMUP. The prevalence of unintended pregnancy in this study is higher than that of the studies from Ethiopia (36.5%) [13], Pakistan (38.3%) [17], Bangladesh (40%) [27] and Brazil (55.4%) [14].
In our study, age was signi cantly associated with unintended pregnancy. Young women below the age of 20 years were more likely to report unintended pregnancy in comparison to women from the above age group, with a similar pattern from other studies of Nepal [6], Ethiopia [28], Pakistan [17], Malawi [21], Nairobi, Kenya [29], and Congo [27]. These effects of age on unintended pregnancy can be explained by the fact that young women may have sexual relations for reasons other than childbearing, and they may have inadequate knowledge and skills regarding birth control, therefore increasing the likelihood of pregnancies being unintended [30].
Likewise, the number of family members was not found to be associated with unintended pregnancy in this study, which is consistent with the ndings of Ethiopia [28]. This study revealed a higher percentage of unintended pregnancies among women with less than or equal to six family members than among women with more than six family members. The reason might be that women with a small number of family members may perceive their pregnancy as unintended due to the lack of care takers in house and being unable to allocate time for rearing and caring of children.
Regarding the educational status of respondents' husbands, it has been found that couples with higher education have better knowledge of the uses and bene ts of family planning, making every pregnancy planned. In contrast, it was found not to be associated with pregnancy intentions. This is consistent with the ndings made from another study conducted in Nepal [6] but contrasts with the ndings of a study by Malawi [21] and Congo [27], where there was an association between partner's education and pregnancy intention. This study has shown a higher percentage of unintended pregnancies among women whose husbands had taken formal education. The reason might be simply because they might have more con dence regarding the ability to control the timing of their pregnancies [6] and could also be because of negligence.
In the current study, the occupation of women was not found to be associated with unintended pregnancy, which is similar to the nding of a study in Egypt [31]. However, this nding is contrasting with the ndings of a study of Nepal [6], Iran [32], and Brazil [14], which could be because of differences in circumstances and varying natures in the measurement of study variables. This study has resulted higher percentage of unintended pregnancies among women who were engaged in non-agricultural paid work.
This could be due to being more career oriented and due to lack of leave as per their requirements, making them perceive their pregnancy unintended.
A signi cant association was seen between pregnancy intentions and socioeconomic status after adjusting for other factors, which is similar to the ndings of Ethiopia [28] and Nigeria [33]. This nding contradicts the ndings made in Nepal [6] and Pakistan [17]. Women from the lower wealth quintile that means the low-income group (the second (AOR = 0.048, CI = 0.007-0.329) and third (AOR = 0.134, CI = 0.026-0.702) wealth quintiles) had a lower chance of having an unintended pregnancy in comparison to the women of a uent ones ( fth wealth quintile). This could be because this study has not included educational status in the construction of the wealth index [28]. Therefore, education may play a vital role in decision making.
Gravidity was not found to be associated with pregnancy intention in this study, which is similar to the nding of a study by Ethiopia [34], but it contradicts the nding of a study by Tanzania [30] and Canada [35], where there was a signi cant association between the number of pregnancies and unintended pregnancies. In this study, there was a higher percentage of unintended pregnancies among women who had been pregnant for the rst time. The reason behind this nding might be that rst pregnancy might not necessarily be planned, making that unintended.
The frequency of abortion was not found to be associated with unintended pregnancy in this study, which is consistent with the ndings of a study in Egypt [31]. This study revealed a higher percentage of unintended pregnancies among women who had terminated their pregnancies at least once compared to those who had terminated their pregnancies more than or equal to twice. This result might be because women who had terminated their pregnancy more than once may become more cautious toward pregnancy planning due to the experience of complications and consequences of abortion.
The age of the youngest child was not found to be statistically signi cant with pregnancy intentions in this study, which contradicts the ndings of a study of Malawi in which time since last birth was associated with pregnancy intentions [21]. There was a lower percentage of unintended pregnancies among women whose young child was less than and equal to twelve months of age. This could be because couples might feel comfortable raising both children with a short gap between them so that they grow together.
Communication between husbands and wives regarding contraception might help to involve husbands in contraception decisions, leading to increased acceptance of contraception utilization and decreased failures. In contrast to this concept, it was not found to be associated with pregnancy intentions in this study, which is similar to the ndings of Congo [27] and Ethiopia [34]. This nding is inconsistent with the ndings of Tanzania [30], Ethiopia [19], and Egypt [31]. In this study, there was a 2.8 percent higher prevalence of unintended pregnancy among women who had communication with their husband. The reason might be that although they have communication between partners, women might not have autonomy to decide on fertility [34].
Husbands play a major role in family planning utilization. Husband opposition in using contraception was found to be signi cantly associated with pregnancy intention in the bivariate analysis of this study.
This nding is inconsistent with the results of the study of Ethiopia [36]. There were 69.5 percent of unintended pregnancies among women whose husband opposed using contraception, which is higher than that of Ethiopia, where there was 18.5 percent unintended pregnancy due to husband refusal to use contraception [37]. The reason might be due to the lack of autonomy regarding household decision making and deciding their health needs, including lack of knowledge regarding consequences of unintended pregnancy [36].
A signi cant association was seen between pregnancy intentions and method of family planning used before pregnancy to control pregnancy in bivariate and multivariate analyses in this study. This nding is similar to the outcome of a study by Congo [27] and Bangladesh [38]. Similarly, behind the higher probability of unintended pregnancy among women who had used modern contraception before pregnancy, there might be a possibility of problems with contraceptive use effectiveness, including contraceptive discontinuation and failure [38]. Another view might be that the users of modern methods might have high expectations towards limiting and spacing their pregnancies and misconceptions regarding lower fertility while using contraception. Therefore, they may consider their pregnancy as unintended [39].
The intention of women to use contraception in the future was found to be signi cantly associated with pregnancy intentions in the bivariate analysis in this study. There was a lower risk of having unintended pregnancy among women who had not intended to use contraception in the future (OR = 0.547, CI = 0.341-0.876) in comparison to those who intended to use contraception in the future. This nding is inconsistent with the outcome of the study of Nepal [6], where there were lower odds of unwanted pregnancy among women who were intended to use contraception in the future. The reason behind this nding might be that this study assessed intention to use contraception in the future as a proxy measure of behavior. The possible gap between intention and actual practice of women could be the presence of some factors that were considered to be constraints to their intentions to use contraception in the future, such as husbands' opposition and fear side effects. making differences between intentions and actual behavior [31].
Empowered women with strong decision-making power in major household decision making are more likely to plan their pregnancy. However, in contrast to previous literature, the level of autonomy was not found to be associated with pregnancy intentions in this study, which is similar to the nding of a study of Nepal that was conducted in 2009 [16] and contradicts the nding made from a study based on the NDHS dataset of 2011 in Nepal [6], Ethiopia [19], India [18], and Bangladesh. The study of Bangladesh has shown that a one-unit increase in the autonomy scale decreases the odds of unintended pregnancy by 16% [40]. In this study, most unintended pregnancies were among women with some autonomy. The mechanism behind this might be due to taking previous or current pregnancy unintended because of lower fertility aspiration by empowered women or limited decision-making opportunities to women being in the patriarchal society [6,16].
Although this study was strengthened by being community-based research and having used the validated pregnancy intentions scale (LMUP tool) to measure the pregnancy intentions of women, it had some notable limitations. This was a cross-sectional study, so it does not allow causal inference for pregnancy intentions and other independent variables, such as sociodemographic and reproductive health-related variables. This study included only married women, whereas the majority of unintended pregnancies resulted from illegitimate sexual intercourse, which is more common among sexually active teenagers and unmarried women. In addition, another limitation might be social desirability bias, leading to reporting intended pregnancy, although it was not. This might lead to underestimating the burden of unintended pregnancies. Similarly, this study does not determine the intentions and associated factors of women who terminate their pregnancy by abortion and miscarriage. Lastly, the results may not be generalized to the whole country, as the study was conducted in one municipality of a district.

Conclusion
This study estimated the prevalence of unintended and intended pregnancy among married women of reproductive age and identi ed the factors associated with it. The study showed that unintended pregnancy is still high, and more than fty percent of pregnancies are unintended. The factors that were found to be associated with an unintended pregnancy were socioeconomic status and method of family planning used before pregnancy. Women who were from the second and third wealth quintiles were less likely to have an unintended pregnancy than those from the fth wealth quintile. Likewise, women who used the natural family planning method before the most recent pregnancy were less likely to experience an unintended pregnancy than those who used a modern method of contraception before pregnancy.
Therefore, the relationship between the use of modern methods and increased risk of unintended pregnancy suggests an urgent need for further research.
However, other study factors, such as marriage age of women, ethnicity, family members, women's education, husbands' education, women's occupation, husbands' occupation, gravidity, parity, history, frequency, reason of abortion, age of youngest child, age of women at her rst child, ideal number of children, and contraception use before pregnancy, were not found to be associated with pregnancy intentions. Spousal communication and autonomy of women were also not associated with pregnancy intentions.
Unintended pregnancy was high in this study, so awareness programs should be raised at the community level. Women using natural family planning methods before pregnancy are less likely to have an unintended pregnancy. The use of natural family planning methods is not always accurate. Therefore, both types of methods need to be equally valued and promoted. Efforts to increase the knowledge and availability of family planning, especially the different methods, should be strengthened. The use of modern methods and increased risk of unintended pregnancy suggests an urgent need for further research. A qualitative study is recommended to explore the relationship between unintended pregnancy and the method of contraception used before pregnancy and socioeconomic status of the family. At the same time, it is a cross-sectional study that might not have shown a causal relation between various independent variables and pregnancy intentions. Therefore, further studies are required to evaluate the causal relationship between independent variables and pregnancy intentions. Conceptual Framework of the study Figure 2 Sampling Procedure

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Questionnaire.docx