Study period and setting
The study was conducted from May 01-30/2016 in Ambo town, Oromia Regional State, located at 110 km distance on the west of Addis Ababa which is the capital city of the country. The town has six administrative kebeles. The total population of the town for the year 2016 was estimated to be 79, 059 (18).
The proportion of women in childbearing age was 22.2 %( 17,551) of the population. The town has one General Hospital, two public health centers, two higher and three medium private clinics that provide LARC methods service as an integral component of other health care services (Fig 1).
All these facilities in the study area are providing LARC methods services for free and the national government is making these methods available.
Study design
A community based unmatched case-control study design was conducted to identify determinants of long-acting reversible contraceptive methods utilization among married women of the reproductive age group in ambo town, Oromia region, west Ethiopia.
Population
Users were married women who were using of one of long-acting reversible contraceptive methods and non-users were married women who were not using any of modern contraceptive methods. Users and non-users who were lived at least for six months in the study area were included in the study.
Sampling: The sample size was determined by using EpiInfo software version 7.1 with an assumption of 95% confidence interval, power 80%, Case to control ratio 1:2; with inter-spousal discussion as exposure variable, 93.6% of users and 83% of non-users with exposure, Odds ratio of 3.0 (24) and 10% non-response compensation in both groups. Accordingly, 140 users and 280 non-users were included in the study.
The entire six kebeles (the lowest administrative units in Ambo town) were included in the study. Prior to carry out the actual study, pilot survey was conducted by the trained UHEW to identify & know the number of existing users and non-users in the six kebeles of Ambo town and the response rate was 100%. Users were first identified by self-report that they are using the LARC method and then they were asked to bring a service identification card/appointment card to cross-check the service card against the method they were using. Non-users were first identified by self-report and then cross-checked from the UHEW registration book to confirm that they really non-users. A total of 1,302 users and 1,806 non-users were identified during the pilot survey. During the pilot survey, unique identification numbers were given for each user and non-user which was also written on their residential homes to facilitate sampling process. A separate sampling frame was prepared for users and non-users. Then computer-generated random number method was used to select users and non-users.
Data collection procedure and measurement
Data was collected by face to face interview method (interviewer-administered method). Six twelve grade complete females who can speak the local language (Afan Oromo) were selected, trained and recruited as data collectors. Two Diploma nurses were employed as supervisors. The interview was delivered in Afan Oromo language. The data collectors were trained on the contents of the questionnaire, interview approach and confidentiality. The supervisors were trained on the contents of the questionnaire, interview approach and how to support data collectors. One data collector was assigned per each kebele and collected the data by moving from home to home. Supervisors supported data collectors by providing logistics required for data collection, collected the filled questionnaire from each data collectors, checked completeness and consistency of the collected data and submitted to Principal investigator daily.
Variables included in this study were; Socio-demographic and economic variables which comprises age, education, occupation, income; reproductive health-related variables which comprises number of parity, number of living children, sex composition of living children, history of stillbirths, history of induced abortion, intention to give birth in the future; individual-related variables which comprises knowledge of LARC methods, attitude toward LARC methods utilization, inter-spousal discussion, responsible person to limit number of children and source of information and Health facility related variables which comprises distance, expectation of availability of method mix and expectation of restriction to method use.
Knowledge of the respondents on LARC methods was measured by the total number of correct answers to 11 items on knowledge questions, with a minimum score of 0 and a maximum of 11. Those who scored 80% and above were declared as having good knowledge, those who scored 50-79% were declared as having moderate knowledge and those who scored less than 50% were declared as having poor knowledge (24).
The attitude of the respondents toward LARC methods was measured by Likert scale type questions. This was measured by the total number of correct answers from fifteen attitude questions toward LARC methods and the mean score of these answers was computed. Then respondents were declared as having a positive attitude and a negative attitude. Those who scored above mean to the correct answers from attitude measuring LARC methods questions were considered as having a positive attitude and those who scored mean and below mean to the correct answers from attitude measuring LARC methods questions were considered as having a negative attitude (26).
Wealth index was computed as a composite indicator of living standard based on variables related to ownership of selected household assets, presence of livestock and materials used in the house. Variables that used to measure wealth index at household level were piped water source, flush toilet piped to sewer system, electricity, separated room for sleeping, separated room for cooking, refrigerator, mobile phone, fixed phone line, radio, electrically working griddle, own home, cement roof type, vehicle, and livestock. The computation was made using principal component analysis (PCA) and composite variables were extracted by summing up the principal components into three components. The adequacy of the model for PCA was checked by the value of Kaiser-Meyer-Olkin measure of sample adequacy (KMOSA) and it became 0.78 and the sample was adequate. Eigen values were used to decide the number of PCs to be retained. Only PCs with Eigen values greater than 1.0 were retained. Three components explained the wealth index with the overall cumulative variance percentage of 69.4%. Detection of outliers & inter-item consistency was performed. To check inter-item consistency, Chronbach alpha of factor lodgings were computed and the value was 0.82. Also, Quintiles of the wealth index were computed.
Both interviewers and supervisors were trained for three days. The content of the training were interview approach, ways to maintain confidentiality and how to keep the privacy of the study participants. Pre-testing of the questionnaire was also done on 21 married women of reproductive age group (7 users and 14 non-users) in Guder town, which is located at 12 km distance to the west of Ambo town. Finally, the data collection tool was refined based on the findings obtained from the pre-test.
Data processing and Analysis
Data were coded manually and checked for its completeness and consistency. Then data were entered and cleaned by EpiData version 3.1 and exported to SPSS version 21.0; where recoding, computing, and other statistical analysis were performed. Descriptive statistics were computed to explore frequency distribution, central tendency, variability (dispersion) and distribution of outcome and explanatory variables. Bivariable analysis was performed to identify candidate variables (P-value less than 0.25) for multivariable logistic regression. Finally, multivariable logistic regression was fitted using a standard enter method to identify independent predictors of LARC methods utilization. Hosmer and Lemeshow goodness of fit test was used to assess model fitness (P-value = 0.02).Adjusted odds ratios together with their corresponding 95% CI were calculated to assess the strength of association and statistical significance.
Ethical Assurance
Ethical clearance and supportive letter to undertake the study was obtained from the Ethical Review Board of the College of Health Science of Jimma University. Permission letter to conduct the research was obtained from the Oromia Regional Health Bureau.
Prior to data collection, the participants were informed about the purpose of the study, their right to refuse participation, discontinue the interview or measurement and their full right to say "no" (opt-out), and it was clearly stated that their decision of "no" will not affect any of their right to health provisions intended for women. Written consent was obtained from study participants of age greater than 16 years. Written informed consent was obtained from a parent or guardian for participants under 16 years old.
Confidentiality and privacy of the study participants were assured and protected by using a unique questionnaire identification number during an interview. Two women, one for IUCD and one for Implant, those in a need of using LARC methods utilization were linked to health facilities where the services are available.