Study setting
The study was conducted in Adjumani refugee affected district in west Nile region, Uganda. Adjumani district is located on the eastern bank of the Albert Nile, and shares common border with Moyo district in the north, Amuru district in the south and east, Arua and Yumbe districts in the west. Adjumani is bordered by the Republic of South Sudan in the northeast. The district has a total land area of 3128 square kilometers, with 46.8 square kilometers covered by water, 37.44 square kilometers is occupied by forest and 1455 square kilometers is arable land.
Adjumani district hosts more refugees 244,374 (50.7%) than the indigenous population of 237,400 (49.3%) inhabitants. The current refugee population come mainly from South Sudan and have been settled in Adjumani since 2013. According to the Office of the Prime Minister (OPM), in 2022 a total of 244,374 refugees and asylum seekers were living in nineteen settlements in Adjumani district. Among the refugees, 85% (182,977) are women and children. This study was conducted in three settlements namely; Nyumanzi, Pagirinya and Mirieyi and the surrounding host communities.
Study design
This was a comparative cross- sectional study. The study population was women of reproductive age (15 - 49 years), both users and non-users of modern FP methods and men whose partners were either or not using any FP method.
Sample size determination
A total of 1307 respondents were sampled and interviewed. This comprised 646 of host and 661 refugees. The sample sizes were computed using comparative study sampling procedure. See table 1
Of all the respondents interviewed, about half comprised the host population, [49%, n= 646] with [33%, n = 214 from Mirieyi, 34%, n = 219 from Nyumanzi and [33%, n = 213 from Pagirinya]. While more than half of the respondents interviewed [51%, n= 661] were refugees [9%, n = 57 from Mirieyi]; 46.4%, n = 307 from Nyumanzi and 44.9%, n = 297 from Pagirinya]. See table 2
Sampling procedure
Multi stage cluster sampling technique was used to select both refugee and host populations. Three refugee settlements were purposively selected because of having the highest number of refugees and having existed in the district for long over 23 years i.e (Nyumanzi, 40,877; Pagirinya, 36,784 and Mirieyi, 7,067) being the newest settlement in Adjumani district. Simple random sampling was used to select the respondents from three clusters or zones, three villages, nine blocks and 661 households. For the host population, three counties were purposively selected. The four sub counties, eight parishes, sixteen villages and 646 households were selected using simple random sampling accordingly.
The number of participants per group required to detect a difference in P1 and P2 in the proportions with significant level α and power 1-β was estimated using the formula below;
Although sample size for each population was estimated at 248 respondents, the acceptance rate to participate was segmented at 90%. Thus, the researcher accounted for the remaining 10%. Accounting for design effect, the sample size was doubled giving a total of 496 for refugee and a total of 496 for host populations, totaling to 992 respondents. To increase power of the study, a total of 1307 respondents were then interviewed.
Data collection procedures
Data were collected through surveys from both refugee and host populations using translated pre-tested structured questionnaires. The research assistants were trained for 4 days prior to data collection. Of every 2 women interviewed, the 3rd one was a man in both communities. A total of 661 respondents were interviewed from the refugee cluster/zone and 646 respondents were interviewed from the host communities. Of these a total of 445 women and 216 men were interviewed from the refugee population and a total of 388 women and 258 men from the host population respectively.
Ethical clearance was obtained from the Makerere University School of Public Health Higher Degrees Research and Ethic Committee (HDREC) # 188 and Uganda National Council of Science and Technology. Written permission was also obtained from the Office of the Prime Minister in Kampala and Adjumani district.
Data management and analysis
Data were collected using the Open Data Kit (ODK) and uploaded into Epidata 3.1. The data were reviewed for completeness, consistency and accuracy. The quantitative data were analyzed using STATA software Version 14 C. Univariate, bivariate and multivariate analysis were carried out to establish factors associated with decision making for FP use among refugee and host populations. Statistical significance was set at p-value < 0.05.
Findings
See table 3
The majority of respondents [64%, n = 833] were females and [36%, n = 474 were males. Two thirds [67%, n = 879] of the respondents were in monogamy marriage/cohabiting. More than a third of the respondents [39%, n = 510] were in the age group of 25-35 years. The mean age for respondents in the host communities was 32.34, SD 10.30. Whereas for the refugees, the mean age was 30.67, SD 10.07.
More than a third of the refugee respondents [38%, n = 492] were Catholics compared to [17%, n = 228] of the host population. More than half [55%, n = 720] of all respondents had no formal education. A higher proportion of the refugee [8%, n = 100] compared to host respondents [0.5%, n = 6] had attained tertiary education.
Contraceptive prevalence amongst host and refugee populations
See table 4
The study found that a similar proportion of refugees [40%, n = 266] and host [42%, n = 269] have ever used a method of FP. However, the majority of hosts 58%, n=377] and [60%, n=772] of refugees have never used FP method (Table 4).
See figure 1
The study shows that more than a quarter [29%, n= 77] of the host and [28%, n = 76] of the refugee population are currently using modern family planning methods. See table 5
The study revealed that a significantly higher proportion of refugees [81%, n = 535] compared to [76%, n = 488] of the host population [p=0.02] would not like to have another child. And a significantly higher proportion of refugees [63%, n = 337] compared to host population [39%, n = 190] preferred to have more than three children. The study also showed that a higher proportion of host population [45%, n = 289] compared to refugees [34%, n = 225] prefer birth interval of more than two years. The study further revealed that more than half of the host population [52%, n = 334] compared to a quarter of the refugee population [26%, n = 171] plan to use modern contraception to prevent pregnancy before the desired time [p=0.00]. However, more than a fifth of respondents [22%, n = 283], with [24%, n = 157] of refugees compared to [20%, n = 126] of host population who had plans to prevent pregnancy before the desired time were not using any modern methods of FP.
See table 6
The study revealed that the main reasons for not using contraception among the host population were fear of side effects [36%, n = 135], don’t want to use modern FP method [28%, n = 107], husband/partner disapproval [6%, n = 22], infrequent sex [6%, n = 21], want to become pregnant [5%, n = 19] and inconvenient to use [5%, n = 19].
However, for the refugee population, the main reasons for not using modern FP methods included; don’t want to use FP method [46%, n = 181], infrequent sex /husband away [18%, n = 69], fear of side effects [6%, n = 23], wanted to become pregnant [7%, n = 27], husband/partner disapproved [6%, n = 22].
See table 7
The study found that the main FP methods being used by the host population were male condoms [34%, n = 65], injectable [21%, n = 40] and implants [19%, n = 36].
Similarly, the main modern FP methods being used by the refugees included; male condoms [23%, n = 43], injectable [13%, n = 25] and implants [20%, n = 37]. However, a significant proportion of host population [21, n=40] compared to refugee population [13%, n=25] were using injectable.
See table 8
The study found that the main FP methods used by the host population were male condoms [34%, n = 65], injectable [21%, n = 40] and implants [19%, n = 36].
Similarly, the main modern FP methods used by the refugees included; male condoms [23%, n = 43], injectable [13%, n = 25] and implants [20%, n = 37]. However, a significant proportion of host population [21, n=40] compared to refugee population [13%, n=25] were using injectable
See table 9
This study showed that amongst the host population, the main factor significantly associated with decision making is religion. Anglicans are 6 times more likely to make decision to use FP compared to the Catholics AOR = 6.21, 95% CI [1.11-34.57]. The study revealed that other factors significantly associated with decision making to use modern FP include joint decision by couples, AOR = 0.04, 95% CI [0.89-138.39] and p-value 0.04; having no prior plans to use FP methods AOR = 0.05, 95% CI [0.01-0.15] and having family interest for large families AOR 0.02, 95% CI [0.00-0.40]. The study further revealed that not consulting key actors in the family prior to FP use is associated with; harassment and separation/divorce. Women who reported fear for being harassed and separated/divorced because of no approval by the key actors prior to FP use at family level were less likely to use modern FP methods, AOR 0.07, 95% CI [0.01-0.79] and AOR 0.03, 95% CI [0.00-0.35] respectively.
See table 10
The study showed that among the refugee population, factors significantly associated with decision making for FP use included religion, joint decision and level of education. Anglicans were more likely to use modern FP methods compared to the Catholics in the settlements; AOR= 5, 95% CI [4.75-132.23]. women who had joint decision with their husbands and women who had attained secondary education were more likely to make decision to use modern FP; AOR 6.43, 95% CI [1.78-16.08] and AOR = 12.15, 95% CI [1.65-89.02] respectively.