Design and participants

This study was part of a large study which adopted a case study design to explore barriers and facilitators for performance of Caesarean section by Assistant Medical Officers in Tanzania. The descriptive phenomenology study design [19] was carried out in four districts in the Northern, Western, Eastern and Southern zones of Tanzania. The zones were purposely selected because they had caesarean section rates either above or below the national average. Further, the type of health facility (run by government or a faith-based organisation), and the location of the facility (rural or urban) were considered in order to tease out the factors that determined whether the referral protocols and system facilitated women’s access to comprehensive emergency obstetric services. Specifically, this study was conducted in 6 health facilities; including both Government and Faith-based organisations, and categorised as either rural or semi-urban (see Table 1). Although the healthcare facilities involved in the study were located in rural or semi-urban areas, they provided either basic emergency obstetric and new-born care (BEmONC) or comprehensive emergency obstetric and new-born care (CEmONC). However, due to various reasons they are compelled to refer patients to other healthcare facilities for appropriate obstetric care. The common reasons for referral were; inadequate number of CEmONC trained health care providers, limited (lack) of medical equipment and insufficient supply of essential obstetric medicines such as safe blood transfusion and magnesium sulphate [18, 20]. In Tanzania all healthcare facilities whether public, private or faith-based are implementing Government policies including maternal and neonatal referral guidelines.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Zones, region and type of healthcare facilities involved in the study.

https://doi.org/10.1371/journal.pone.0255475.t001

The study participants included Medical Officers, Assistant Medical Officers (AMOs), Nurse-Midwives and Health Secretaries working in the 6 healthcare facilities and District Medical Officers (DMOs) of the 4 study districts and were conveniently recruited. Because of the limited number of medical personnel in these study settings, convenience sampling strategy was used and therefore whoever was around during data collection period and willing to participate in the study was recruited [21]. Heads of the maternity services assisted with identification of participants based on set inclusion criteria. The participants had to work in the labour ward for 3 or more years to ensure that they had adequate experience managing women with obstetric complications and utilizing the referral system or had administrative roles in the maternity unit. The researchers met the labour ward staff, then provided information about the study purpose and issues of confidentiality. Thereafter, a convenient interview time was arranged with those who agreed to participate in the study.

Interviews

Twelve (12) individual, qualitative interviews were conducted using an in-depth interview guide. The guide was developed from a review of the literature that explored barriers to the implementation of a referral system, and focused on the experience of healthcare providers in rural and semi-urban district healthcare facilities for women. In order to avoid interference with the activities of the healthcare facilities, interviews were conducted when the participants were off duty. Interviews were conducted in Kiswahili in a quiet and comfortable room within the healthcare facility. At the end of each session, the recorded interviews were listened to allow new emerging issues to be included in the guide before the subsequent interviews [22]. After 10 interviews were conducted, it was noted that the knowledge saturation was reached, yet, 2 more interviews were done to adhere to the recommendations that 12 interviews will be required to achieve saturation [19, 23]. The sample size for qualitative studies depend on the complexity of the research questions, the purpose of the research study, the diversity of the sample, the nature of the analysis and the time and resources available [24–26]. On average, each interview took about 50 minutes. Before the interview began, participants signed the consent form to agree to participate and to allow recording of their conversations during the interview sessions. All participants were informed on the study aim, procedures and that their participation was purely voluntary so they were free to decline or withdraw at any time in the course of the study. The study was approved by the Senate Research and Ethics Committee of the Muhimbili University of Health and Allied Health Sciences (MUHAS) and the Ministry of Health Community Development Gender Elderly and Children (MOHCDGEC).

Analysis

Analysis of the data was guided by Braun and Clark thematic analysis, a flexible data analysis method that focuses in identifying themes and patterns [27, 28]. The analysis followed the six-stage in thematic analysis [20]. First, the recorded interviews were listened twice and thereafter transcribed verbatim and translated into English. The translated transcripts were then cross-checked against the Kiswahili transcripts to ensure the accuracy and completeness of translations before coding. When discrepancies occurred, these were corrected accordingly. Second, initial codes were extracted from the transcripts, whereby each researcher coded one transcript. After the team agreed on the codes, the NVivo10 software was used to organise the codes and categories. As the process continued, new, emerging codes were formulated and compared with the initial codes. Third, involves sorting the codes based on their similarities and differences into potential categories and themes, and collating all the relevant coded data extracts within the identified themes. Fourth, themes were reviewed by either splitting, combining, or discarded them during the process. Fifth, the themes and categories were named and were discussed and agreed upon among the team. Final stage, the report was produced in which a few quotes were included to demonstrate the essence of a point.

Participant characteristics

The 12 health care providers interviewed were between 34–59 years of age, 6 were medical practitioners (MD or AMO), 5 were nurse-midwives working in the labour ward and 1 was a health secretary. Seven 7 providers had worked in the wards where they had interviewed for 7 years or more.

Themes

Three (3) main and interconnected themes and subthemes were abstracted from the data relating to participants experiences referring women with obstetric complications for obstetric care: 1. Adhering to a rigid referral protocol; 2. Completing the referral of women to an adequate health facility and 3. Communicating the condition of the woman with obstetric complications (see Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Themes and subthemes from healthcare providers experience in implementing referral guidelines.

https://doi.org/10.1371/journal.pone.0255475.t002

Adhering to a rigid referral protocol

Completing the referral of women to an adequate health facility

Communicating the condition of the woman with obstetric complications

Referral and the referral protocol

Consistent with other studies [29, 30], the findings from this study indicate that the implementation of the referral guidelines is not followed. For example, the national referral regulations require that medical doctors sign the referral form in order for the patient to be transferred to the next level health facility for adequate or speciality care. However, because of the limited number of medical doctors in health facilities in the districts, many facilities are manned by the Assistant Medical Officers—a physician assistant-like professional who is an integral part of community healthcare [31]. Therefore, it is impractical to authorise only medical doctors to refer patients, in case of emergency other medical personnel should be allowed to sign referrals to save lives. Lack of specialists was reported as a major reason for lack of referral implementation in Iran [29]. The reason behind mandating only medical doctors to refer patients was to ensure that the transferred patient actually required specialised care, since medical doctors are believed to highly be clinical competent than other health care providers.

While the MOHCDGEC referral regulation does not authorize the AMOs who are the majority in health care facilities in the districts to sign referral forms, it was reassuring to learn that in the absence of medical doctors the AMO, clinical officer or a nurse could sign a referral form for the patient to be transferred to the next level health care facility. In a situation where there is only one medical doctor, then the guideline needs to be followed, often resulting in a delay in the transfer of patients who need urgent care. In the event of an obstetric referral, there is either a prediction that a woman may go on to develop obstetric complications or the complication has already developed and cannot be managed at a specific health care facility [32]. The decision to make a referral is dependent on; the clinical skills of the referring health care provider, the tools for diagnosis, and the availability of a health facility with the necessary specialty care. Other factors that need to be considered are the quality of care at the referral facility, the cost of care, the distance between facilities, the availability of reliable transportation, the availability of clear communication between providers, someone to accompany the patient as well as the feasibility of travel by the patient [33]. The right to the highest attainable standard of health is a fundamental human right [4]. In hierarchical health systems such as in Tanzania, a well-functioning referral system that allows for the continuity of care across different tiers of care is central to achieving this right [34].

Another barrier to safe and efficient referrals for complex patients is that; the national referral protocol requires the doctor to only refer patients to the next level referral health facility even when it does not have the required services. The referral system is designed to optimize the use of three levels of health services and to avoid unnecessary congestion and waste of human and material resources in the specialised levels [35]. The referring clinician should be given a leeway to refer patients to an adequate health facility if the next immediate level health facility has no services that would benefit the patient. This will ensure that women get appropriate management in a timely manner, contrary to rigid referral guidelines that mandate a linear referral procedure that does not meet the emergent clinical needs of the patient. This practice could be the reason for patients not following the referral system, as noted by Eskandari and colleagues [29].

Transport

The successful management of obstetric emergencies depends on how promptly the woman is able to get into a health facility for adequate obstetric care. Consistent with the findings from studies in Tanzania [18, 36] and elsewhere [37, 38], our study found that moving patients from one facility to another was a challenge. Frequently, women were not able to be moved to the next level health facility due to unavailability of the hospital vehicle or lack of fuel to run the ambulance. Patients were either compelled to travel using unsafe public transportation or buy fuel needed to run the hospital ambulance (if one was available), or at times the health care providers were compelled to use their own limited funds to transport patients to the next level facility. This commonly resulted in some practitioners being hesitant to refer women for adequate care leading to delays occurring in the healthcare facility [9]. Referral regulations also require the woman to be accompanied by a health care provider that makes unavailability of hospital ambulance more critical. The lack of reliable transportation to shuttle women to higher level healthcare facilities in the event of an obstetric emergency may contribute to the increased number of maternal deaths in these districts. The WHO describes the consequences of inadequate transportation for the delivery of health care to the most impoverished settings causing delays in both seeking and receiving care at health facilities. Also, it is a contributing factor to subsequent deaths and disabilities among women with obstetric complications during labour. The WHO estimated that 75% of maternal deaths can be prevented through timely access to childbirth-related care. Therefore, the transportation plays an important role in achieving sustainable development goal no 3 of reducing the global maternal mortality rate to less than 70 per 100,000 live births by 2030 [11]. For instance, India has significantly increased institutional deliveries and reduced maternal deaths through implementation of emergency transportation system a model that could be adopted to improve healthcare facilities referral [39].

Communication during referral

The communication of a patients’ clinical information at the time of a referral is essential to the delivery of high-quality health care. Health care providers in primary, secondary and tertiary care facilities value this information exchange for shared patients [40]. Our study found gaps in referral information on patients condition before transfer to the recipient hospital, and lack of feedback to the referring hospital following treatment. Health care facilities lack reliable communication mechanism to facilitate a swift transfer into a referral health care facility, as well as a mechanism to provide information on the patient’s management following referral consistent with studies reported elsewhere [29]. Some studies have proposed the use of software, phones, emails and texts for sharing patient information while maintaining confidentiality [41]. However, these tools are usually not available in most rural health facilities and some do not positively impact referral communication [42]. The number of referral patients, the limited number of medical practitioners and the cost of the technologies are likely to make these possible solutions impractical. However, having the appropriate clinical information about a woman’s condition prior to her transfer would help staff in the recipient facility prepare and enhance the quality of care. Evidence show that specialists who received timely patient referral information prior to referral were able to provide optimal care twice as often as specialists who did not [40].

Although the MOHCDGEC referral regulation requires feedback to the referring health facility after the management of the patient is complete, it does not specify nor suggest a mechanism to do so. In practice, the patient is usually discharged home with a discharge summary that provides little detail of what was done. In addition, many hospitals use health management systems that are not linked across health care facilities. All of this results in a lack of feedback to the referral health facility, consequently poor continuity of care. The referral process is a critical component of quality care and this breakdown in communication needs to be improved in order to achieve better maternal and newborn outcomes [43].

Methodological considerations

The rigor of data was increased through credibility, dependability, conformability and transferability [44]. Validation of the key themes through dialogue with all members of the research team, as well as continuous reflection and revisiting codes to ensure accurate fit [45]. For dependability, all transcripts were retained for review and served as an audit trail. In addition the researchers were receptive to the participants’ ideas to reduce bias in the findings. To ensure authenticity and credibility, the study used multiple data sources (i.e., Nurse-midwives, Assistant Medical Officers, Medical Officers and Health secretaries), drawn from diverse regions of Tanzania (i.e., Morogoro, Tanga, Kigoma and Mtwara), from a variety of settings (i.e., rural and semi-urban) and within a wide range of health facilities (i.e., public and faith-based). Nevertheless, there are some limitations to this current study. The analysis of the interviews was completed in English from translated transcripts; the quality of the participants’ accounts may be affected, as some Kiswahili words may not have a direct translation in English. However, the transcripts were verified by the research team fluent in Kiswahili and then back translated from English to Kiswahili to check the quality of translation and ensure that the translations were accurate. Further, all codes and themes were discussed among the researchers who were able to review the original transcripts. The direct quotes provide in-depth and rich data and could be used as transferable knowledge to similar contexts to that of Tanzania. Although in-depth interviews with a small sample size of healthcare professionals can be considered as strength, we also recognize it is a limitation as we involved only six healthcare facilities located in four different regions, and that we did not include the experiences of healthcare providers in the receiving referral healthcare facilities. Therefore, we recommend further studies be conducted in facilities receiving referrals to learn their perspectives and experiences of implementing referral guidelines.