Effectiveness of maternal referral system in a rural setting: A case study from Rufiji district, Tanzania

Background: The functional referral system is important in backing-up antenatal, labour and delivery, and postnatal services in the primary level of care facilities. The aim of this study was to evaluate the effectiveness of the maternal referral system through determining proportion of women reaching the hospitals after referral advice, appropriateness of the referral indications, reasons for non-compliance and to find out if compliance to referrals makes a difference in the perinatal outcome. Methods. A follow-up study was conducted in Rufiji rural district in Tanzania. A total of 1538 women referred from 18 primary level of care facilities during a 13 months period were registered and then identified at hospitals. Those not reaching the hospitals were traced and interviewed. Results. Out of 1538 women referred 70% were referred for demographic risks, 12% for obstetric historical risks, 12% for prenatal complications and 5.5% for natal and immediate postnatal complications. Five or more pregnancies as well as age <20 years were the most common referral indications. The compliance rate was 37% for women referred due to demographic risks and more than 50% among women referred in the other groups. Among women who did not comply with referral advice, almost half of them mentioned financial constraints as the major factor. Lack of compliance with the referral did not significantly increase the risk for a perinatal death. Conclusion. Majority of the maternal referrals were due to demographic risks, where few women complied. To improve compliance to maternal referrals there is need to review the referral indications and strengthen counseling on birth preparedness and complication readiness. © 2010 Pembe et al; licensee BioMed Central Ltd. According to the projection from the 2002 population and housing census the population in Rufiji district in 2007 was estimated at 240,000 with a yearly growth rate of 2.5% [13]. Geographically the Rufiji river intersects the district from West to East dividing it into flood plain, coastal-delta, and plateau zones. The district experiences a heavy rainy season from February to May and a less intense one from October to December. The majority of the population are peasants growing cassava, rice, sorghum and maize as food crops. Cashew nuts and coconuts are the main cash crops available. Commonly their farms are located some distance from the family home and residents shift to temporary dwellings at the farms during the heavy rain season. Transport in the district includes canoes, boats, motor vehicles and bicycles. Most of the roads in the district are unpaved and difficult to pass especially during the rainy season. The district has two hospitals; the government owned district hospital located in the district town Utete south of the river and the non profit mission hospital Mchukwi located in the northern part of the river. The hospitals serve as a first line centre for women living close by and first referral for all primary level of care facilities. Caesarean section and blood transfusion can only be provided at the hospitals, while vacuum extraction rarely can be provided at primary level of care or hospitals. There are 56 primary level of care facilities of which four are rural health centres (RHCs) owned by the government, 47 government dispensaries and five private dispensaries. The primary levels of care facilities are the main access point for maternal and child health services in rural areas. They are staffed primarily by clinical officers who are prescribers and most common the in-charge of the facilities, nurses and/or maternal and child health (MCH) aides as trained maternal service providers. In some dispensaries nurse auxiliaries with one year training or on job training on provision of maternal services are providing maternal care. The rural health centers are larger than the dispensaries and have in addition beds for admitting patients. There is no user fee for antenatal and delivery care services except at Mchukwi mission hospital where investigations and operative procedures are charged. Women book for antenatal care at the health facility nearest to them. All the primary level of care facilities provide antenatal and delivery care for low risk women and are supposed to refer women according to referral indications stipulated in the Ministry of Health and Social Welfare RCHC-4. Only the district hospital has an ambulance which serves all the primary level of care facilities with emergency referrals to the district hospital. Another vehicle owned by the Mchukwi hospital shuttle three times per day between the hospital and the nearby towns. In emergency referrals health workers assist the woman and family to arrange for transport by calling for the ambulance from the district hospital, but they have to pay for fuel. The RCHC-4 is divided into three sections; the pregnancy care, delivery and immediately after delivery care, and follow up after delivery care. In the first part, the prenatal care, there are three categories of referral indications named A, B and C. A woman with one of the indications in category A should be referred for further investigations while one with indication in category B, should be referred for delivery. A woman requiring referral in category C is supposed to be referred immediately to hospital. The referral indications in category A, B and C are presented in the table below (Table ​(Table11). Referral indications during pregnancy care according to RCHC-4, Tanzania aCaesarean section is in category A and B. bDanger signs are not stated In the natal and immediate postnatal period women with referral indication should as well be referred immediately to the hospital. Indications during this period include spontaneous rupture of membranes without labour, labour before 34 weeks, labour for >12 hours/obstructed, abnormal lie or presentation of the baby, vaginal bleeding, variability of foetal heart beats ( < 120 or >160 beats per minute), elevated body temperature of >38 Centigrade, eclampsia or blood pressure ≥140/90 mmHg, haemoglobin <60% (8.5 gm/dl), small pelvis or big baby, meconeum, retained placenta, severe perineal tear and blood loss ≥500 mls. During the first visit women are screened for referral indications in category A and B. For those identified as in need of referral a tick is made on a specific box on the antenatal card and the woman is informed of the need to go to hospital for further assessment or for delivery. During the subsequent antenatal care visits women given referral advice should be emphasized on the referral advice given. Those referred for delivery are advised to stay near the hospitals at the late months of pregnancy despite there are no maternity waiting homes near the hospitals. Women and their families have to arrange for a place to stay either in guesthouses or by their relatives. If a woman develops any of the indications for referral in category C during her prenatal care, she is referred immediately to hospital. Some women given a referral to hospital for further assessment or delivery actually come back to the primary level of care in labour for delivery. It is up to the discretion of the health worker to decide to re-refer the women or conduct the delivery. According to Kielmann et al 1995 and UNICEF 1997, a random sample of 25% to 30% of the health facilities in a district of an average size is usually adequate and feasible to represent a district health service situation [14,15]. All four RHCs and 14 randomly selected dispensaries in the flood plains and plateau zones among those with five or more deliveries per month were included in the study. The delta zone was not included due to difficulties in accessing the area. The primary levels of care facilities included covered 54% of the population in the district. The sample size of referred women was calculated using the soft ware Epi Info 6. Based on a study in Gutu, Zimbabwe [16] with an antenatal and delivery referral rate of 36%, a desired precision of 5%, 95% confidence interval, and a power of 90% a sample size of 364 referred women could have been sufficient. However to be able to compare with other studies on the use of obstetric care and captured variation of maternal referrals during rain and dry seasons, all maternal referrals to the hospitals from 1 June 2007 to 30 June 2008 were recorded. A parallel data collection system was established since the routine data collection indicated the risks but no information whether the women were referred to hospital or not. Health workers at the primary level of care facilities received refresh training on the RCHC-4 with an emphasis on the referral indications. Accurate recording of all women referred to hospital during pregnancy, delivery and after delivery was emphasized. During the training, it was emphasized that all health workers should stick to the national guidelines on referral indications and the health workers should repeat advising the women on referral in the subsequent visits. Information on women's socio-demographic characteristics and indications for referral were collected. If a woman had more than one indication for referral, the one associated with worse outcome or needing urgent attention based on obstetrician assessment was taken. A research identification number tag was stapled on the woman's antenatal card. Women that were referred were identified by the trained health workers at the hospitals, who recorded the treatments and outcome of deliveries if delivery takes place at the hospital. The data collection forms were reviewed by the first author in each primary level of care and the health workers completed missing information. The woman was regarded complied with referral advice when she reached the hospital after being referred from the primary level of care facility due to any of the indications. Women who did not reach the hospitals were identified by comparing the register books in the primary level of care and the hospitals. These women were traced by the primary level of care providers in their respective catchment's areas. Those women contacted were asked about reasons for not going to the hospital and their pregnancy outcome. If the mother had deceased, a relative or anybody who was with the mother during the incident was interviewed. The structural quality of the hospitals was assessed. Information on qualified staff for provision of emergency obstetric care, functioning operative theatre, blood transfusion facilities, evacuation facilities, functioning vacuum extractor and availability of antibiotics, oxytocics and anticonvulsants. The referral indications in category A and B in the RCHC-4 card is a mixture of demographic and obstetrical risks, we re-grouped the referral indications in these two categories to make the group containing only the demographic risks and a second group with risks from the obstetrical history and risks related to delivery. The final groups were the demographic risk factors, obstetric historical risks, prenatal complications, natal complications and immediately postnatal complications. Factors which may have higher risks of perinatal mortality were calculated. All factors in the demographic risk and obstetric historical risks groups were used in the calculation of risk of perinatal death. The prenatal and natal complications groups were calculated together and the risk factors included were haemoglobin <60%, blood pressure ≥140/90 mmHg, abnormal lie/presentation, vaginal bleeding, large fundal height for gestation age, labour for >12 hours/obstructed and eclampsia. The risk of perinatal death was not calculated in the postnatal complications group as complications in this group do not affect perinatal outcome. The software SPSS was used for statistical analysis. The risk for perinatal death for women not complying with the referral in relation to those complying was calculated using odds ratio (OR) and 95% confidence intervals (CI). The protocol was reviewed and approved by the Muhimbili University of Health and Allied Sciences, the Senate research and Publication committee as part of the on going studies on quality assessment and monitoring of maternal referrals in the district. Permission to conduct the study was obtained from the Rufiji District Medical Officer and District Executive Director’s office. All participants agreed voluntarily to participate in the study after being informed of the aim and the consent sought from them.