Study area and period
The study was conducted at public health facilities that provide long-acting family planning services in the Toke Kutaye district of West Shoa Zone, Oromia Region, Ethiopia. Guder is the capital town of the district, which is located 126 km to the west of Addis Ababa, the capital city of Ethiopia. The district has 27 kebeles (the lowest administrative unit in Ethiopia), one hospital, four health centers and 24 health posts. Based on the current data, the population of the district is projected to be approximately 128,259, of which 63,873 are males and 64,386 are females, and the reproductive age of women from 15 to 49 is 23,895 [15]. The study was conducted from 20 May 2021 to 28 July 2021.
Study design
A mixed research design was employed.
Study population
Reproductive-age women (15–49 years) who had used the long-acting family planning method at least once, currently used who revisited the study settings during the data collection period were enrolled for quantitative data, and family planning methods providers (Nurses and Midwives) who work in the study health facilities were included for the qualitative study.
Eligibility criteria
Inclusion criteria
Long-acting family planning method users current or previous, age between 15 and 49 years who revisited the selected health facilities during the data collection period and providers were included in the study.
Exclusion criteria
Women who did not use any long-acting family planning methods at least once and age below 15 and above 49 years were excluded from this study.
Sample size determination
For quantitative study
The sample size was calculated using the single population proportion formula [(n = (Zά/2)2p (1—p)/d2)]. The following assumptions were made to calculate the required sample size, i.e., P = 40.4% [10], level of significance (α = 0.05), Z α/2 = 1.96 with 95% CI, 5% margin of error (d) tolerated for the p value n = sample size \(n=((Z \alpha )/2)2*P(1-p)\)/d2.
n = (1.96)2*0.404(1–0.404)/0.052 n = 370, considering a 5% non-respondent rate, the final sample size was = 389.
Sampling technique and procedure
The required sample size (n = 389) was selected from all five public health facilities in the district which includes Guder hospital, Guder health center, Goro Sole health center, Maruf health center and Toke health center. Two months prior to data collection, a total of 1,024 revisit long-acting family planning clients had attended those health facilities. Of these, 32 clients were at Guder hospital, 370 were at Guder health center, 300 at Goro Sole health center, 260 were at Maruf health center and 62 were at Toke health center. Accordingly, by using proportional to size allocation, twelve clients from Guder hospital, 140 clients from Guder health center, 114 clients from Goro Sole health center, 99 clients from Maruf health center and 24 clients from Toke health center were included in this study. Health facilities family planning client registration consisting list of repeat users of LAFP were used as a sampling frame. Then, systematic random sampling method was used to select participants every 3rd (1024/389) interval from those repeat clients who came after using LAFP methods.
For qualitative study
A purposive sampling method was used to select the study participants. An in-depth interview was conducted among twelve family planning providers. They were recruited purposively from the selected health facilities.
Study variables
Dependent variable
Long-acting family planning method switching.
Independent variables
Socio-demographic factors: age, ethnicity, religion, educational status, marital status, educational status of women and their spouse, occupation, place of residence. Reproductive history: parity, number of children, future fertility plan, and age at first birth. Perceived needs: family planning demands (limiting or spacing), obtain information about family planning methods and user’s family planning perception.
Operational definition
Long-acting family planning switching: the pattern of switching the contraceptive method from one long-acting (IUCD and implant) to another short-acting method [16].
Data collection tool and procedures
Data were collected from 20 May 2021 to 28 July 2021 using a semi-structured questionnaire and in-depth interviews. The questionnaire was first prepared in English and then translated to the local language (Afaan Oromo). The second version of the tool was retranslated to the original version by a language expert to evaluate its consistency. The data were collected by recruiting five midwives. The principal investigator supervised the data collection procedures at all study settings. Training was given for the data collectors for two days about the aim of the study, data collection techniques and procedures. For the qualitative study, the key informants (family planning providers) were interviewed in a quiet and confidential room. The in-depth interviews were audio recorded and notes were taken during the field. The interview continued until data saturation was achieved. The audio-recorded interview and field notes were transcribed verbatim immediately after the interview was completed. All the data were placed in a locked cabinet and secured.
Data quality control
To assure the quality of the data, the questionnaires were pretested on 5% of the sample size other than the study area among women who switched LAFP methods in Ambo General Hospital. Training was given to the data collectors to ensure the accuracy and consistency of the data. The principal investigator spot-checked and reviewed all the completed questionnaires to ensure the completeness and consistency of the collected data. Data entry was performed by the principal investigator to maintain the accuracy of the data. The trustworthiness of the qualitative data, such as the credibility, transferability, dependability and reliability of the data, was rigorously ensured [17,18,19].
Data management and analysis
All the questionnaires were checked, coded and entered into Epinfo version 7 and exported to SPSS version 21 software for data analysis. Binary logistic regression analysis was conducted to examine the statistical association between dependent and independent variables. Both bivariate and multivariable logistic regressions were used to identify the associated factors. Variables that had a statistical association in the bivariate logistic regression at p-value < 0.25 at 95% CI were entered into a multivariable logistic regression to control the confounding variables [20]. Hence, an adjusted odds ratio (AOR) with a 95% CI at a p value < 0.05 was considered to indicate the statistical significance. Finally, the results were presented in the form of tables, figures and text using frequencies and summary statistics such as the mean, standard deviation and percentage to describe the study population in relation to relevant variables.
The audio-recorded interview and field notes were transcribed verbatim immediately after the interview was completed. The audio-recorded interviews and field notes were transcribed verbatim carefully by a local language (Afan Oromo) word by word and arranged with the written notes taken at the time of discussion and interview. Then, the transcribed verbatim was translated to English. The researcher read the whole transcription several times. Then, the transcribed verbatim were imported and organized using NVivo version 12. Then, the data were coded, sorted and analysed thematically. The grouping codes that emerged from the data analysis were presented in themes and then triangulated with the quantitative findings to gain insight into a problem. Finally, the researcher shows how the qualitative data explain the quantitative results.
