Timely referral saves the lives of mothers and newborns: Midwifery led continuum of care in marginalized teagarden communities – A qualitative case study in Bangladesh

Background: Prompt and efficient identification, referral of pregnancy related complications and emergencies are key factors to the reduction of maternal and newborn morbidity and mortality. As a response to this critical need, a midwifery led continuum of reproductive health care was introduced in five teagardens in the Sylhet division, Bangladesh during 2016. Within this intervention, professional midwives provided reproductive healthcare to pregnant teagarden women in the community. This study evaluates the effect of the referral of pregnancy related complications. Methods: A qualitative case study design by reviewing records retrospectively was used to explore the effect of deploying midwives on referrals of pregnancy related complications from the selected teagardens to the referral health facilities in Moulvibazar district of the Sylhet division during 2016. In depth analyses was also performed on 15 randomly selected cases to understand the facts behind the referral. Results: Out of a total population of 450 pregnant women identified by the midwives, 72 complicated mothers were referred from the five teagardens to the facilities. 76.4% of mothers were referred to conduct delivery at facilities, and 31.1% of them were referred with the complication of prolonged labour. Other major complications were pre-eclampsia (17.8%), retention of the placenta with post-partum hemorrhage (11.1%) and premature rupture of the membrane (8.9%). About 60% of complicated mothers were referred to the primary health care centre, and among them 14% of mothers were delivered by caesarean section. 94% deliveries resulted in livebirths and only 6% were stillbirths. Conclusions: This study reveals that early detection of pregnancy complications by skilled professionals and timely referral to a facility is beneficial in saving the majority of baby’s as well as mother’s lives in resource-poor teagardens with a considerable access barrier to health facilities.


Introduction
Globally 830 maternal deaths occur every day, 99% of which occur in developing countries 1,2 . According to the World Health Organization, roughly 303,000 maternal deaths are caused as a result of pregnancy and childbirth related complications 3,4 . Globally, about 3.7 million neonatal deaths occurred within the first 28 days, with 75% in the first week of life 5 . Only 19 out of 186 countries have achieved the Millennium Development Goal-5, related to reduction in maternal mortality 6 ; unfortunately, Bangladesh is not one of them. Estimations suggest that about 87% of maternal deaths occurred in South Asian and Sub-Saharan African regions 7 . According to the Demographic and Health Survey, neonatal mortality rates range from 28 to 54 per 1000 live births in Bangladesh, India and Pakistan 8 . In 2010 the Bangladesh Maternal Mortality and Health Care Survey (BMMS) claimed that the lifetime risk of maternal death is 1 in 500 due to pregnancy and delivery related complication, and two third of these deaths occurred in the postpartum period.
A study in Bangladesh found that 38% of the maternal deaths occurred by haemorrhage, which is the most common cause, 20% occurred by eclampsia, and 8.1% occurred by sepsis 9 .
Another study in the teagarden area of Bangladesh revealed that maternal death in teagarden areas is higher due to lack of knowledge on maternal complication. Ignorance, traditional myths, family restriction on seeking better care, and dependency on traditional birth attendants and village doctors also influence these maternal deaths in teagarden communities 10 .
Referral is the process of coordinated movement of health care seeker to reach a high-level care within a small window of time 11 . The goal of timely referral is to minimize or prevent the delay for transportation (called second delay), and ensure pre-hospital care while transporting a patient to the referral facility 12,13 . In 2014, Directorate General of Health Services (DGHS) reported that out of 120 maternal deaths 47 deaths occurred in the teagarden area of Moulvibazar district of Bangladesh. Estimations suggest that about 46.4% of maternaldeaths occurred at home, and 7.1% while the women were on route towards a facility; this indicated the delay occurred as a result of delay in decision making of which facility to take the mothers for management, and arranging transport to go to the facility 14 . Another study stated that 22.2% of maternal deaths occurred with more than 6 hours delay in decision-making and 12.9% of deaths occurred with 1-2 hours transportation delay 9 . In light of this, it can be assumed that ensuring emergency obstetric care services, and quick referral during the perinatal period can help reduce maternal deaths.
To safeguard the reproductive age (15-49 years) of a woman, continuous care from family and community, along with support in getting easy access to referral healthcare facilities, is needed 15 . Transportation support, timeliness of referral, and inter-facility transfer are major contributing factors found to reduce the rate of maternal deaths 12,16 . A social autopsy study of maternal deaths found that very few mothers sought facility based care during complications, and that ensuring timely referral through transportation saved the lives of many of them 14 . It is recommended that five Emergency Obstetric and Newborn Care (EmONC) services, including four basic EmONC (BEmONC) and one Comprehensive EmONC (CEmONC), should be available and geographically distributed for each 500,000 individuals of a population 17 . The component of care (consisting of antenatal care, identification of high risk mothers, safe delivery conduction by skilled birth attendant, timely referral of complicated mothers and postnatal care including essential newborn care.) with high quality services can be ensured by the good referral system at all levels, both in facilities as well as in communities by the trained health care providers 9 . A shifting process is developed after the identification of high-risk pregnancies from a risk based approach to provide skilled care during delivery, and emergency obstetric care when complications occur 18,19 . This approach is not adequate to reduce maternal and neonatal mortality as the capacity is limited at the primary level of care, and is difficult to access in the referral facilities remaining in most of the low-income countries 20 . Professionally, a referral transport system must be managed for providing some basic intervention to the patient before reaching the referral facility 21 .
An intervention named "Bagan Mayer Jonno" has been implemented in the selected teagardens in the Moulvibazar district. The project ran through counseling and courtyard meetings of pregnant mothers, as well as an advocacy meeting with their guardians regarding quick referral of complicated mother. This project also supported the communities in detecting high-risk mothers by the active participation of volunteer and professional midwives. It also managed the provision of transportation and assistance of volunteers to ensure a quick and safe referral procedure. The present qualitative study describes the referral system using the midwifery led service delivery in five selected teagardens of Moulvibazar district in Bangladesh.

Methods
Qualitative method was used to collect information in this study. The referral records of 2016 in selected five teagardens were reviewed retrospectively and qualitative information of selected 15 referral cases were collected though in-depth interviews at the community.

Context of the tea gardens
The average distance between a teagarden and Upazila Health Complex (UHC) varies between 12-20 kilometers. Approximately, a population of 25,000 people with around 300 pregnant mothers at any point in time live in these gardens.

Bagan Mayer Jonno
As part of the intervention, community volunteers called 'Bagan Sebika' were placed in the community, and professional midwives were situated in teagarden health facilities. Bagan Sebika (Paid Volunteer) perform community based activities including home based counseling, courtyard meetings, and advocacy meetings with the pregnant mothers and family members. Bagan Sebika also facilitated the mothers recieving antenatal care (ANC) at the facility. They also accompanying the referred mothers to the referral centre. Midwives' role at teagarden facility includes ANC, counseling, delivery, referral, and postnatal care (PNC). Midwives also conduct delivery, referral, and PNC at a community level. Midwives also supervised the activities of Bagan Sebika at the community level.
A total of 25 Bagan Sebikas worked in the five selected teagardens. They were assigned to conduct regular home visits to households and met to pregnant mothers. The Bagan Sebikas raised awareness on various issues such as birth preparedness, pregnancy complications, danger signs, and the importance of referral. If the Bagan Sebika identified any complicated or high-risk pregnancy case, they immediately communicated with the professional midwife over mobile phone. Professional midwives are usually experienced in identifying high-risk pregnancies through ANC checkups and previous medical history of the patient. Based on severity of the complication, the professional midwife along with the Bagan Sebika motivate family members of the high-risk pregnant mother to quickly refer to the higher referral centers including the UHC, district hospital and Teagarden central hospital. This counseling assists family members in being aware of the situation and the risk involved, provides them with information on where to seek care, and motivates them to make quick decisions. The Bagan Sebika also assists family members to organize transport, and assist them throughout the referral process. All these steps combined together helps decrease instances of delay in decision making [first delay] and transportation delay or second delay in the target population. In cases of severe complications, the midwives themselves might also help in organizing transportation.
The present study. The present study was conducted by a facility-based retrospective record review of all referral cases occurring at the referral hospital from the selected five teagardens from January to December 2016. According to 2016 records, a total of 72 high risk pregnant mothers were referred from these five teagardens to the referral centres (Upazila health complex, district hospital and teagarden central hospital). Each teagarden has both permanent workers (registered) and causal workers (unregistered). The teagarden authority provides referral support for the registered mothers (workers), whereas, for the unregistered mothers, the referral support is very low or absent.
The professional midwives used a structured tool to document the referral history and treatment at the teagarden facilities and did follow up all referral mothers until outcome at the referral facility though Bagan Sebika. [ Table 1].
To conduct the retrospective record review of referral centers, a structured tool was developed by the research team. The tool contained data on mother's particulars, current pregnancy history, antenatal care, complications, treatment history, referral details, preparedness of the facilities to manage emergency obstetric complications and delivery outcome. This review was carried out by the professional midwives working in the teagarden facilities. The record review included sociodemographics of the mother, medical condition of the referred mother, causes of referral, and view of the feedback of the referred mother and their family members [ Table 2].

Data collection.
A total of 72 referral case data were entered into SPSS software (version 24.0). After entry, all data was checked for missing data and consistency. Once checking was complete, the data was cleaned, and all analysis was done using software SPSS. For case scenario description, a total of 20% of cases (n=15) were purposively selected from the five teagardens (three cases from each garden). Midwives went to the household and organized a meeting for each of the cases. The Midwife invited the family members, relative and neighbours to the meeting to gather on responses from the family and community, as well as understand the referral linkage and service delivery in the facility. The Bagan Sebika in the community organized the meeting based on suitable date and time given by the community. Descriptive statistics were computed for all variables of interest. Frequencies were established to examine the demography of referred mothers, condition of mothers during referral, and documented causes of referral. The project support and remarkable findings of the cases were analyzed through review of the case stories collected from the teagarden facilities. Themes were identified after reading and re-reading of the case stories 22,23 and finally thematic analysis was performed.

Ethics and consent
This study under "Bagan Mayer Jonno" intervention has been approved by the national ethical review committee of CIPRB (memo-CIPRB/ERC/2016/010). Verbal and written consent were received from each of the referred mothers before collecting the information for the study.

Results
From the review of records from 2016, Bagan Sebika identified the complicated mothers and immediately informed the project midwife. Then the project midwife decided whether the case needed to be referred. The project midwife also identified mothers as high risk during their routine ANC for referral. A total number of 72 complicated pregnancies (16%) were identified from a total of 450 pregnant mothers. These complicated mothers were identified at different stages during their antenatal visit, or during delivery, or immediately after delivery. Mothers informed the Bagan Sebika if any complication arose. Bagan Sebika also identified complicated mothers during their regular household visit. Then Bagan Sebika immediately informed to the project midwife. Professional midwives ensured immediate referral to the higher center after consultation, and coordinated with garden midwives, doctors and Bagan authorities. Unregistered workers in all cases directly referred to the Upazila or District facility, whereas registered workers were taken immediately to the garden's existing referral system. In about 85% of cases, the transportation support was provided for referral of the complicated mothers, and of them in 75% of cases the Bagan Sebika (Volunteer) participated during referral of the mothers [ Figure 1].

Age and occupation of the mothers
The referred mothers were mostly young. About 44% of mothers referred were in the age group 17-20 years, whereas 18% and 38% of mothers were from the age group of 21-25 years and 26-35 years, respectively. About 16.7% of referred mothers were housewives and the remaining were from other professions. Highest percentage (51.4%) of referral was among the unregistered teagarden workers (mothers), whereas only over 11% was registered teagarden workers. [ Table 3].
Gravida and stage of the referred mother 39%, 54% and 7% of referred mothers were identified as 1 st gravida, 2 nd to 3 rd gravida and 4 th gravida. Most of the mothers referred were in the labour stage (76%), whereas 12.5% were referred during the pregnancy period, and 11.1% after the delivery conduction [ Table 3].

Place and time of referral
With project support, about 60% mothers were referred to Upazila Health Complex and 28% referred to Sadar district hospital. Only 13% of registered mothers or dependent workers of the teagardens were referred to teagarden referral center [ Figure 2]. The time range at which most of the mothers (about 42%) were referred was between 10 a.m. to 2 p.m., where usually doctors, nurses and midwives are available in the government facilities. The remaining referrals occurred at times when only nurses and midwives are available in the facilities. But about 28% and 30% of mothers were referred within the period of 6 a.m. to before 10 a.m., and after 2 p.m. to 8:30 p.m. which is the vital period when doctors or service providers may not be found at government facilities [ Table 3].
Mode of delivery and outcome of referred mother About 14% of referred mothers needed Caesarian section for complications and 86% were normal vaginal delivery conducted by a nurse or midwife in the referral center. 94% of mothers delivered livebirths and 6% delivered stillbirths (2) and intrauterine deaths (2) at referral facilities with the assistance of skilled health care providers [ Table 3].
Delay to start treatment at referral center after complication arises The delay includes first (decision), second (transportation) and third (treatment) delays, which started from the complication arising, up to receiving treatment. In about 46% of cases family members needed more than 4 hours to make a decision  as whether to seek care at a facility or not. Whereas about 60% cases reached from teagarden dispensary to the referral center (UHC) within one hour, and 74% cases women received treatment within one hour after arriving at the facility. Midwifery counseling as well as transportation support from the project influenced much in reducing the community delays mainly first and second delay [ Figure 4].
Case scenario description A total number of 15 cases were selected randomly out of 72 cases for in-depth analysis and case scenario description. These description includes the socio-demography of the referred mothers, condition of the mothers for referral, responses of the family members and society, referral linkage and services delivery at referral centre [ Table 4].

Discussions
The study revealed that among the referred mothers around 51% were unregistered workers who referred with the support of the project Bagan Mayer Jonno as they were not entitled to get any referral support from teagarden authorities. About 76% of mothers were referred during the period of delivery and 31% referred with the complication of prolonged labour. Most of the mothers (about 60%) were referred to the Upazila Health Complex and after referral about 14% mothers delivered by Caesarian-section at the facilities. A  study conducted in rural Tanzania showed that about 28% of pregnant women were referred from primary level of care to tertiary level to ensure their better pregnancy outcome.
The same study also concluded that the most common referral complications found were multiparity (35%), young age of mother (30%), obstetric complications mostly due to prior history of caesarean section (12%), and previous existed prenatal risks like high blood pressure, severe anaemia etc. (12%) 20 . On the other hand, our study found that 31% of mothers referred with prolonged labour, 18% with pre-eclampsia, 11% with post-partum haemorrhage (PPH) due to retained placenta, 9% with premature rupture of membrane (PROM), 18% with severe anaemia, breech presentation & twin pregnancy, and remaining 13% with other complications.
Proper transportation with cost support along with a good communication technology is the prime concerns in establishing an effective referral 17 . Our study is also consistent with the findings that almost all referral occurs with transportation support, along with extra assistance from a midwife or volunteer, ensure the lives of many vulnerable mothers. The counseling of the midwife about the severe condition of the mother, as well as its dreadful consequences, and assistance of volunteers during referral motivated the family to quickly make their decision on referral.
Our study showed that about 50% of referred mothers received treatment within 6 hours of referral and 10.6% within 2 hours. Addressing of second delay, or transportation delay, has a significant role in reducing maternal mortalities. Many studies showed that referral transportation should be available within 30 minutes of worsening condition of a mother, so that the complicated mother can be taken to a referral center as early as possible to initiate her treatment 24 . A mechanism needs to be established for the proper utilization of easily accessible functional transport services, which could be either from government, or from a private referral transport services 24,25 . Our study found that availability of transport support and assistance of volunteers from that teagarden enhanced quick referral, which consequently reduced the first and second delay.
This study found that quality ANC support by a midwife from respective gardens not only helped to identify high-risk mothers, but also further assisted the family to make a decision and prepare to delivery at a facility. Projections show that the Government of Bangladesh (GoB) already started midwifery education in all nursing institutes from 2012 and the GoB have the mandate to continue this midwifery led service delivery system until 2021, with the vision to serve hard to reach communities of the country. Another study revealed that to ensure basic and life-saving intervention to the patient, consistent support of a skilled staff should be  Case-10 19 years old unregistered worker pregnant woman at 3 rd gravida. Her husband was also an un-registered worker Midwife conducted the delivery at home but placenta was not removed. The mother was referred for Retained placenta The family member wanted to conduct delivery at home due to their family tradition "This mother didn't come at facility for ANC during pregnancy. Her husband was also ignorant on MNH care at facility"-Midwife mentioned Guardians informed volunteer after 12 hrs of labour pain & she informed to midwife. Midwife conducted delivery at home but when placenta was not removed she carried the mother immediately to district hospital.
After admission in district hospital the placenta was removed by doctor & nurses.
The mother then safely returned home after one day observation. Panchayat president was accompanied with mother during referral. The family members were concern to carry the mother at UHC immediately after referral. The nurse & doctor provided special care of the mother at UHC. After 6 hrs after admission the mother delivered normally a live birth with the assistance of nurse. available until the patient reaches referral facilities. However, several studies stated that it is difficult to pre-determine complication occurrence during pregnancy or childbirth 17,24 . The government mandate to continue the midwifery led service delivery until 2021, it is therefore necessary to regularly review the referral indicators and counsel on complication readiness, as well as birth planning by a health attendant to improve compliance on maternal referral 20 .

Conclusions
Early detection of complicated mothers and quick transfer to the referral center can ensure the survival of many mothers and neonates. The GoB has plans to scale up the unique midwife led service delivery (both basic and emergency health care services) system to support high-risk mothers of under privileged communities including the teagardens. The teagarden board, owners of the teagardens and local government, including policy makers of every level, must come forward to work together in finding out the best possible way to support the mothers of teagarden. At the community level, professional midwives play a key role in timely referral of a complicated mother to the facility. An integrated approach based on existing government health care delivery system with support from garden health facilities for timely referral of complicated mothers can be beneficial in reducing maternal and neonatal mortality in Bangladesh which in turn will be effective in reaching sustainable developmental goal on time.

Data availability
Data is stored at the CIPRB. Due to sensitivity of the data (contains identifying information), permission is required from the Ethical Review Committee (ERC) of CIPRB, Dhaka, Bangladesh for sharing data with a third party. Data can be requested from the CIPRB, who will contact the Ethical Committee to gain approval to share the data. The conditions for gaining data access are a formal request with a clear objective and formal permission from the Ethical Committee. Please contact the corresponding author in order to request the data though email at info@ciprb.org. The article is very well articulated and the findings are supporting the discussion and conclusion very strongly. Additionally, it gives deep insight and shares positive experience of a community based referral intervention in the area of maternal health to a pro-poor, marginalized and isolated community. However, minor language editing will make the article more scientific, lucid, authentic and reader friendly. For example, the author could use the word 'explore' instead of 'evaluate' in the last line of the introduction part in the abstract. Methodology in abstract part needs reorganizing and rephrasing where qualitative method is described. The introduction of the methodology section could commence like thiso "This case study is designed as a mixed-methods retrospective assessment to explore the………" The body of the method section in the abstract can be described in this way, o "In-depth interviews and retrospective document were carried out to …… Thematic analysis was performed to analyze the qualitative data."

Open Peer Review
The Bangla phrase, " " should be mentioned in English "garden for mothers") Bagan Mayer Jonno for the non-Bengali speakers e.g. Method section in the main article needs to be elaborated and re-organized to make it reader friendly and self-explanatory. This section needs revision to maintain cohesion and coherence. Other than midwife is there any provision of additional service providers in the center inside the tea garden? Do midwives conduct home delivery? How do they maintain referral record? How is it documented? What does mean by the term 'professional midwife'? Data processing procedure is described under 'data collection' sub-heading. It should be renamed as 'data analysis' or 'data processing'. This section needs to be revised to address the cohesion and coherence also. Using the term, 'case study' instead of 'case story' will shape it more scientific. Need elaboration in the data analysis of qualitative methods. Need to mention, especially what types of qualitative method are used here. Sometimes, 'case story' and sometimes 'IDI's were mentioned, but the reader may feel difficulty to understand. Case scenario description' part may go to the method section. Otherwise, you need to rephrase the sub-heading such as 'findings from the case studies' or rewrite the body of the paragraph in line with the previous sub-heading, so that it could be understood that you are describing result, not the process or method. The study reveals many opportunities to reduce the maternal mortality and morbidity of tea garden mothers. However, does this study uncover any challenges or obstacles that need further attention? Additionally, one of the recommendations should be 'the scale-up of this intervention as mothers. However, does this study uncover any challenges or obstacles that need further attention? Additionally, one of the recommendations should be 'the scale-up of this intervention as a model for other marginalized communities in remote areas, who are experiencing 'poor' maternal health services.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Yes

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate? Not applicable Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
Referee Expertise: Qualitative Research in the area of maternal, neonatal and child health, nutrition and family planning I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. This is an interesting article on the workings of midwifery, especially referral by midwives, in a district in Bangladesh. The case study approach is appropriate but the description of the Methods is slightly odd. This is a typical case study based on secondary analysis. But the authors do not mention 'secondary analysis', let alone give a reference to a methods paper /textbook on the topic. Most record studies are retrospective, i.e. researchers using the record data after it was written. "A qualitative case study design by reviewing records retrospectively..." Also it is possible that authors used a Content Analysis (Krippendorff 2004) rather than a 'general Greater clarification on how the interviews were analysed: were they audio-recorded and transcribed? how did they come up with the 5 headings in the table -are these the emerging themes from the qualitative analysis? While Table 4 is interesting and gives a good insight into the cases, the paper would be greatly strengthened if this descriptive presentation could be synthesises and reported in the results. This synthesis of the key issues emerging from the interviews should also be included in the abstract. The quotations provided in Table 4 come from various people, not just the women e.g. TBAs, nurse, panchayat etc. Does this mean these individuals were interviewed? if so details of the methods used for these qualitative interviews also need to be given.
Further details on the socio-economic situation of the tea gardens would help readers understand the context.
Acronyms and Bangladeshi-specific words should be spelt out for an international audience.
This has the potential to be an interesting and valuable paper, but greater clarity on the qualitative methods and analysis is required before being indexed.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? No

Are the conclusions drawn adequately supported by the results? Partly
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.