An exploratory case study design that adopted a qualitative approach was used for identifying challenges facing the Assistant Medical Officers training for the performance of Caesarean section delivery in Tanzania. A qualitative case study was necessary for undertaking this study as the training of AMOs is a real phenomenon that involves social processes [15,16].
Context of the study
Tanzania is divided into seven geopolitical zones, namely: Northern, Eastern, Central, Western, Lake, Southern highlands, and Southern zones. The south, west, and central zones are considered more rural compared to the rest. The country has seven AMO schools with three located in the Northern zone, two in the Eastern zone, one in the lake zone, and one in the Southern highland zone (fig 1). Each AMO school had a capacity of admitting up to 40 AMO trainees  Tanzania has five cities, two located in the northern zone, and the rest located in eastern, lake, and southern highland zones. Dar es Salaam, the largest business city that contains the largest number of the health workforce in the country is located in the eastern zone.
The organization of the health care services provision in Tanzania is in a pyramid of three levels. The primary level (comprising of District hospital/s, Health centres, dispensaries, health posts, and the communities), secondary level (comprising of Regional and Regional Referral Hospitals), and tertiary level (comprising of Zonal, specialized hospitals, consultant hospitals, and National hospitals). At all levels be in rural or urban areas, health care services are provided by both public and private health facilities.
This study was carried out in four rural districts (Handeni, Kasulu, Kilombero, and Masasi) located in the four zones (Northern, Western, Eastern and Southern in that order), two AMO schools (one in the northern zone and one in the eastern zone) and at the national level with officials from the ministry of health responsible for the health workforce development and training. The selected AMO schools involved one that was owned and managed by the ministry of health and one under the public-private partnership.
The four zones were purposefully selected to include both rural and urban zones and zones with AMO schools operating under public-private partnership and those operating under the ministry of health (public alone). In each zone, a random selection of rural districts was done whereby one rural district was included in the study.
This study involved participants from different levels of the health care system who are involved in training, supervision of AMOs after training, and those working with the AMOs. These included; Principals from AMOs training schools, AMOs tutors, AMO trainees, Regional Medical Officers, District Medical Officers, Medical Officers in charge of the district hospitals, Senior AMOs at the district hospitals, and one retired AMO (table 2).
The purposeful sampling strategy was used to enrol key informants for this study. The enrolment started by identifying key people who are dealing with AMOs training, supervision, and those who work with AMOs. The latter was done through consultation with officials from the directorate of Human resources development and training from the Ministry of Health at the section of allied health training and Regional and District Medical Officers of the selected study sites. From the Ministry of health, key informants were the officials dealing with overseeing the training of AMOs. These were those dealing with the selection of AMO trainees and overseeing of the AMO schools. From the regions and districts, this study involved the health managers. In this category, the Regional Medical Officers and the District Medical Officers were included as they are responsible for the work and work environment, permission for further studies, and incentives to the AMOs. At the selected health facilities, this study involved the immediate work supervisors; the Medical Officers in charge of the district hospitals and Senior AMOs at the district hospitals. These are responsible for supervising and overseeing the day to day practice of the AMOs to include the performance of Caesarean section. To get the perspective of changes that have taken place in the training and scope of practice of AMOs, one retired AMO who was trained and practiced as an AMO and later trained as a Medical Doctor was included in this study. The latter was identified through consultation with senior AMOs from study sites and senior gynaecologist who worked with this AMO.
For the focused group discussion, a convenience sampling strategy was used to obtain AMO trainees. Participants who were present during the data collection period and agreed to participate in the study were enrolled from the two AMO schools. In each AMO school, two focused group discussions were conducted one with male and one with female AMO trainees.
Data for this study were collected between September 2015 and February 2017. Semi-structured Interview and Focus Group Discussion guides developed in English and later translated to Kiswahili language were used for conducting the Key- Informant Interviews (KIIs) and Focused Group Discussions (FGDs). To ensure quality, experienced research assistants who are fluent in both English and Kiswahili languages were recruited and trained on the objectives of the study, the guides, the informed consent, and the full research process.
Before data collection, the selected informants were contacted by the lead researcher via phone call to set the appointment for the interview. For the AMO trainees, the principals of the training schools were contacted in advance to organize for the FGDs. During data collection, the researchers carried out most of the interviews and FGDs, and the research assistants were taking field notes. Audio records of the interviews were transferred into a computer by the Data Manager and kept in a PIN folder in a computer that was only accessible to him. The transcripts were all kept by the Data Manager but only shared with the research team for analysis.
Key Informant interviews
We used different semi-structured interview guides containing questions specific to each group of informants to carry out 29 KIIs. (table 2). The interview guides were prepared based on experiences on the training of AMOs and task sharing in the country as documented from the available literature [10,17,18]. The questions in the guides solicited information on the challenges at the AMO schools, at the districts, and the national level about the challenges facing the Assistant Medical Officers training for the performance of Caesarean section delivery in Tanzania. The interviews were carried out at a designated office of the informant and it was recorded using a digital audio recorder. Each interview lasted between 60 and 100 minutes.
Focused Group Discussions
We used a semi-structured FGD guide developed based on the competencies detailed in the AMOs' training curriculum and available literature on task sharing and Caesarean section delivery [9,10] to carry out four FGDs with AMO trainees from the two AMO schools involved in this study. In each school, we carried two FGDs, one with the female and the other with the male AMO trainees. The number of participants in each FGD ranged from 7-12. In total 35 AMO trainees participated in the four FGDs. From the FGDs, we explored challenges related to the training of the AMOs for acquiring knowledge and skills for the performance of the Caesarean section as stated in their curriculum. The FGDs lasted between 55 and 120 minutes. A researcher moderated all FGDs.
All interviews and FGDs transcripts were transcribed verbatim. The Kiswahili transcripts were then translated into English before the analysis. A team of four researchers with vast experience in qualitative research, health systems, medical education, and maternal health cross-checked the accuracy and completeness of translations against the original notes before coding. Any gaps identified or clarifications needed were discussed and corrections made accordingly.
Qualitative content analysis as described by Graneheim and Lundman was used to guide the analysis . Codes were extracted from the reduced meaningful unit. In the beginning, the research team read and re-read the transcripts to familiarize themselves with the data before the coding process. The first author developed the initial codebook, based on our study objective and the conceptual understanding of the training of the AMOs in Tanzania. The codebook was discussed by all authors, further developed, and a final codebook was imported into NVivo 10 qualitative data analysis computer software. The agreed codebook was tested by independently coding the first two interview transcripts by three authors. Their coding was almost similar and, hence, the codebook was not modified at this time. The team then distributed the transcripts among each other for the coding process.
We coded the meaningful units of text to the codes (nodes) that were found to represent that unit. Some of the meaningful units were coded more than once. At this stage, although the data analysis was guided, it was not confined to the primary codes. Inductive coding was assigned to text segments which represented a new code that was not pre-determined. The new codes were assigned as separate codes or an expansion of the codes available in the initial codebook. All the coded transcripts were then organized by using NVIVO 10 qualitative data analysis software.
Similar codes were grouped together and through abstraction, sub-categories were formed. Through comparison and checking and rechecking of similarities and differences between the sub-categories, the sub-categories were sorted to form categories to reflect the manifest content of the text that were supported with suitable quotes from the transcripts. Further interpretation of the categories was then used to ensure the latent meaning is also brought into focus. The whole process although described as a linear process, it was iterative at all points to ensure that both the manifest and latent meaning of the data is not lost.
Ethical approval was obtained from the Muhimbili University of Health and Allied Sciences Research and Ethical Review Committee. Permission to conduct the study in the four study settings was granted by the Ministry of Health. Written informed consent was obtained from each participant after receiving explanations about the study aim and they were informed that their participation was voluntary and they were free to decline or withdraw at any time in the course of the study. All participants were informed that there was no financial compensation for participating in the study and only water was provided during the interview or discussion. Participants’ privacy was assured by not using their names or facility identity during the data collection and dissemination process through written reports and peer referred publications. The latter aimed to ensure that no one out of the research team could identify the place where data was collected. Permission was requested for the use of an audio recorder during interviews and discussions.