Home Based Life Saving Skills Training in a Rural Area in Tanzania

STATEMENT OF THE PROBLEM AND RATIONALE Despite several interventions to address the unacceptably high rates of maternal and neonatal mortality, the problem still persists. Maternal Mortality ratio in Tanzania has been persistently high without signs of declining. Currently it stands between 578 and 950 per 100,000 live births (TDHS 2005, WHO 2007). The neonatal mortality rate as well in the country is 32 per 1,000 live births (TDHS 2005). Almost 94% of pregnant women attend antenatal care in health facilities at least once, but only 47% deliver in these facilities and skilled attendance at birth is 46% (TDHS 2005, Mpembeni et al 2007). Births that occur outside the health facilities are assisted by relatives, TBA and some women do not have assistance at all (Kayombo 1995, Kayombo 2008, Comonsky 1983 ). Studies in Tanzania have shown that home birth with unskilled attendants is often the norm (TDHS 1999, TDHS 2005, Mpembeni 2007). This indicates that there are factors that impede these women delivering in the health facilities. Among the factors may include socio-cultural, economic activities, family, community factors. If these factors are compounded by low awareness of danger signs of obstetric complication then these women and neonates are more at risk of death when a complication occurs. Previous studies in Rufiji district showed that community has low acceptance of facility delivery and referral to hospitals when advised (Pembe 2008b). Furthermore, like in other parts of Tanzania women have low knowledge of danger signs of complications during pregnancy, delivery and after delivery despite attending antenatal care (Mpembeni 2007, Pembe 2008a). In another study on the use of essential obstetric care facilities, a quarter (60/232) of women with complications were referred to the hospital level but less than half (48%) arrived at their referral points. Out of 60 referred cases 31 (52%) did not arrive in any of the two hospitals of the district. Nineteen were traced for their outcome and it was found that three died on the way to the hospital (Urassa 2005). Consequently, there is a great need for an innovative and empowering community-based intervention whose characteristics address this problem. METHODOLOGY DESIGN This will be a cluster randomized trial. The cluster has been chosen because the intervention is done in the community in groups rather than individuals. The unit of randomization will be the health care facility. Randomly selected facility clusters will receive HBLSS in addition to the antenatal care received and the in the control group only routine ANC will be done. Study area This study will be conducted in Rufiji district. Rufiji district is one of the six districts of the Coastal Region in Tanzania. It is situated about 300 km in Southern of the capital city Dar es Salaam. The district is more rural with few suburban areas. The district has an estimated population of 203,102 according to census results 2002 with a male population being 98,398 and female 104,704 (BoS 2002). Geographically the Rufiji river divides the district into two halves Northern and Southern parts with the formation of flood plain, coastal-delta, and plateau zones. Since inauguration of the "Mkapa" bridge in 2003 across the Rufiji river, transport between northern and the southern part of the district as well as to the southern regions has improved. Only the main road from Dar es Salaam to the Southern regions is passable in all seasons, while other roads in the district are difficult to pass especially during the rainy season. The limited transport system makes people dependent more on the health services available within the district. According to the administrative structure in Tanzania, the district is divided into five divisions, which are further divided into 19 wards. There are eight wards in the flood plain, four in the coastal-delta, and seven in the plateau zone. There are 128 villages in the district. Each village has 2 to 4 hamlets (Vitongoji) under the leadership of a hamlet chairperson. The average population in a village is 1600, therefore with an annual population growth rate of 4% it is estimated that 64 women will become pregnant in a year. The district has two hospitals, Utete a government district hospital and Mchukwi a private not for profit hospital owned by the Pentecostal church. Both hospitals have doctors, nurses and midwives providing medical care and surgical care including comprehensive emergency obstetric care. There are four rural health centres (RHCs) and 52 dispensaries two of which are private not for profit and another two are private for profit. Antenatal care (ANC) is provided in hospitals, all RHCs and all dispensaries. The normal practice is for the pregnant woman to register for antenatal care at the nearest health facility. Dispensaries and RHCs have guidelines for referring women with risks or complications to hospitals. Participants Randomly selected facility clusters will receive HBLSS in addition to the antenatal care received or only routine antenatal care. The unit of randomization is health facility. 28 health facilities will be selected and randomised into two groups each with 14 facilities. One group will be provided with HBLSS educational programme through their VHW while another group will be monitored for the purpose of comparison Key participants will be pregnant women together with their immediate relatives such as aunt and the husband. They will receive the HBLSS training. Village health workers who will be trained by HBLSS trainers, will be used to provide the training in the community Intervention The selected clusters will receive HBLSS training. This training is being conducted in some other parts of Tanzania. The HBLSS procedure The HBLSS was conceived as a component of the Community Partnership for Safe Motherhood Model. It targets a homebirth team consisting of all of those who can be expected to be present at a birth, namely, the pregnant woman, her family caregivers, and the birth attendant. The education, motivation, cohesion and mobilization of pregnant women, families and communities are seen necessary to improve pregnancy outcomes. HBLSS training was designed by the American College of Nurse-midwives has been successfully implemented in other developing countries including India and Ethiopia (Sibley et al 2001). In Tanzania it was adopted and adapted for use and was launched by Honourable Mohamed Shein the Vice President of the United Republic of Tanzania on the 25th of March 2007 in Morogoro. The White Ribbon Alliance (WRA) adapted the HBLLS training curriculum taking into consideration the major causes of maternal mortality in the region and the Tanzanian context, and translated it into Swahili. The training and materials have been field tested and the first Training of trainers (TOTs) and the training of community was done in Manyara region. Other trainings are ongoing. Recently WRA conducted training in Rukwa rural district that involved district health providers, 20 primary health care providers and 500 community members from 20 villages. HIV awareness and prevention of mother-to-child HIV transmission are integrated into the training manual. The strategy is to build on existing practices and to negotiate acceptable and feasible safe practices with the homebirth team members to develop both a consensus on the appropriate action and a capacity to take that action when needed. The HBLSS training manual has a flexible, modular design comprised of 12 preventive and life saving skill topics. Drawing on best practices and current research, to maximize effective communication and learning among community members who may be unable to read, essential content is reinforced through pictorial Take Action Cards that are taken home for reference. The trainer uses various teaching methods to maximize effective communication and learning among participants including: * Stories and case histories * Discussion and demonstrations * Identification of complications using picture cards * Take Action Cards. * There will be one Take Action Card for each topic or problem. These cards will be used during training and a booklet made of these cards will be left at woman's home. HBLSS sensitization and training The HBLSS training in the district will start with ten (10) TOTs training at the district level. These will train the research assistants (RAs) preferably community health worker (VHW) or health assistants, one from each selected village. The RAs will train the community on HBLSS. All pregnant mothers will be identified by the RA. The RA will make appointment to the family according to their convenience. The RA will visit the family every month to discuss the danger signs, decision making, and birth preparedness and emergency readiness with the family. The RA will keep all the records of training each family, the number visits made and whether the woman attended antenatal care or develop any complication. When the woman delivers the RA will record the place of delivery and whether there was any referral given. Sample size: Baseline study Since the study will employ cluster-sampling, technique, a design effect was taken into consideration when calculating the sample size. We used a design effect of 2 a value which have been reported in the most recent Demographic and Health Surveys as a design effect of the variable "Mothers received medical assistance at delivery" in the Coastal Zone where this study will be conducted. Therefore using the above formula and taking into consideration the design effect, a minimum of 766 women is required for the baseline survey. Sample size for Cluster Randomized control trial The sample size for the intervention study which is a cluster randomized trial was calculated using the Acluster programme methods described by Donner and Klar, 2000. We estimated the mean cluster size of 100 pregnant women, an intracluster correlation coefficient of 0.05. We further assume that a power of 90% to detect an increase of 15% in the rate of hospital delivery. In this case we will need 28 clusters, 14 for intervention group and 14 as control. We will need to have 1347 women in intervention group and 1347 in the control group. A 20% in increase will be done so as to offset the loss of follow-up. Sequence generation This will be done by including all women who book for antenatal care during the time of the study, both in the control and intervention clusters. Allocation concealment Four health centers will be chosen randomly for intervention and control (2 each). Then random selection of 24 dispensaries will be done by using random number tables. Allocation to intervention and control will be done randomly by lottery method. Statistical methods Analysis will be done taking consideration on clustering. Primary analysis will be undertaken as intention to treat at cluster level and individual level. Pearson Chi-square test will used to compare variables. The result will be deemed significant if p value \< 0.05. Multiple logistical regression methods will also be used to see the predictor variables on the use of health facility for delivery. This will be done for both cluster and at individual level. DATA COLLECTION Baseline information The provision of emergency obstetric care services using United Nations (UN) signal functions and Safe motherhood Initiative (SMI) assessment tool in all the facilities will be done. Health workers will be interviewed to determine their capabilities to undertake emergency obstetric care in their facilities. Birth preparedness, emergency readiness including awareness of danger signs of obstetric and neonatal complications among women, their spouses, head of the households and other family members will be done. Information on male involvement will also be sought. A structured questionnaire will be used to gather the information. Information relating to traditionally preferred and/or recommended discouraged foods during pregnancy will also be collected. In addition, information on other traditional practices including rituals for the first or the last born child and or twins will be documented. Qualitative interviews including focus group discussions (FGDs) and in-depth interviews will be conducted. FGDs will be done in the community on perceptions, attitudes, male involvement and health seeking behaviour on the use of emergency obstetric care services. In-depth interview will be conducted with key informants such as religious leaders and other opinion people in the community to compliment the data on perceptions of place of delivery and of obstetric and neonatal danger signs. In-depth interviews will also be conducted to health care providers to supplements the documentation with focus on factors which make many women deliver outside health facilities and efforts made to make them deliver in health facilities. In addition review reports of ward health committee to see if maternal and newborn features out in the report as a sign of awareness. Information on traditional pregnancy management practices, including use of herbal plants and their preparation in facilitating child delivery and their perceived safety will be collected. Plants used will be collected and botanic identification will be done by a botanist for scientific names and for literature review for their safety. In addition where possible concoctions used for facilitating child delivery and for managing pregnancy and infant health problems will be collected for laboratory screening. Information from the qualitative studies will also help in preparing instruments for quantitative studies. Topic guides will be used during the FGDs and in-depth interviews. We will also conduct an economic valuation of the training so as to obtain a cost-effectiveness analysis of this type of educational program. DATA PROCESSING AND ANALYSIS Quantitative data: When data collection will be completed, serial numbers will be assigned to the questionnaires to facilitate identification of the different patient data. Thereafter responses for the open-ended questions will be reviewed, categorized and coded for the computer data entry. Data entry into the computer will be done by the principal investigator with the help of an assistant using the statistical package for social sciences (SPSS) and EPI INFO6 computer programmes. Data cleaning will be done. Chi- squared test will be used to determine association between variables. Where the chi-square test will not be valid the Yates corrected Chi square test will be used. The results will be deemed significant if p-value will be less than 0.05. Qualitative data: The audio-taped information will be transcribed and then back translated from Swahili to English to Swahili. A qualitative content analysis method as described by Graneheim and Lundman will be used for analysis (Graneheim 2004). All texts will be read several times to identify meaning units, that is, statements that relate to the topic of analysis. The meaning units will then be condensed. The condensed meaning units will be coded then the codes will be categorized according to similarities and differences in content.