Introduction

Childbirth care remains suboptimal in many low-resource settings, causing unacceptable maternal and perinatal mortality and morbidity. Realistic, context-tailored clinical support is called for to assist birth attendants in providing best possible evidence-based and respectful care. The PartoMa pilot study from Zanzibar suggested that co-created clinical practice guidelines and low-dose, high-frequency training were associated with care improvements and perinatal survival. In the present study we will modify, implement and evaluate this intervention in five urban, high-volume maternity units in Tanzania.

Methods and Analysis

The study design is based on a theory of change, and includes three main steps: I. A mixed-methods situational analysis will explore factors affecting care. Step II. Based on step I., the PartoMa guidelines and training will be contextually modified through discussions with birth attendants and postpartum women. III. The modified intervention will be implemented through a stepped-wedge cluster trial, with embedded qualitative and economic analyses. Women in active labour and their offspring will be followed until discharge to assess intra-hospital stillbirths, intra-facility neonatal deaths and caesarean sections without medical indications, and the incremental cost-effectiveness ratio will be measured. Central intermediate outputs include health providers' knowledge, barriers and facilitators to intervention use, and clinical performance.

Background

Globally, 300,000 women and 5 million children die each year in relation to childbirth. Most would have survived had they been given proper care.

Multiple clinical guidelines have been developed to improve maternity care in low-resource settings. However, they are often incompatible with the contextual realities. Maternity units in low-resource settings often deal with limited capacity, high turnover of staff and an unworkable ratio of labouring women per birth attendant. This often makes it impossible to adhere to international clinical guidelines and calls for research on effective development, implementation and upscaling of achievable clinical guidelines for such settings.

The PartoMa project's focus on generating evidence-based and respectful childbirth care is a key priority to reach the Sustainable Development Goals - particularly on birth-related survival and gender equality, but also in regards to societal development and poverty reduction. Also, the PartoMa project might serve as an example of context-tailored development and implementation of guidelines for other areas of health care.

The PartoMa Dar es Salaam intervention study will be carried out in the five highest volume maternity units in the fast growing urban center of Dar es Salaam, Tanzania: Amana Hospital, Temeke Hospital, Mwanyamala Hospital, Sinza Health Centre, and Mbagala R. Health Centre. A total of 50,000 births occur in these five facilities each year, providing a typical example of overburdened urban maternity units in low-income countries. In these contexts, quality of care is negatively impacted by lack of staff, persistent increases in births due to urbanization, and increased demand for facility birth. These facilities primarily serve women of lower socioeconomic background. Each birth attendant typically takes care of at least 3-6 labouring women simultaneously.

Study objectives

To analyze barriers, facilitators, effects and costs of implementing context-modified clinical guidelines and low-dose, high-frequency training to improve the quality of care and survival during birth in five urban, low-resource and high-volume maternity units in Tanzania.

The specific objectives include:

I. To carry out a mixed-methods intrapartum care assessment in five maternity units, focusing on the present and past structures within which care is given, processes of care provision and birth outcomes.

II. To explore and develop necessary context-modifications for the PartoMa intervention of clinical guidelines and training to reflect birth attendants and labouring women's needs and circumstances in the five maternity units, and to assess whether and how the resources and experiences that hospital stakeholders bring to the process are informing and steering the process.

III. To assess the impact of the context-modified PartoMa intervention on perceptions, knowledge and skills among birth attendants, quality of care and birth outcomes in the five maternity units, and its cost-effectiveness as well as opportunities and barriers in the process.

IV. To develop a toolkit for co-creating and implementing clinical guidelines and associated training that may be of relevance within and beyond maternal health.

Methodology

We will assess the structure within which care is provided, processes of care delivery and birth outcome indicators will be analysed. Structural indicators will be assessed through structured questionnaires for birth attendants and a checklist for each facility. To assess the process of intrapartum clinical decision making, a case-control study will be conducted of intrahospital stillbirths (positive foetal heart rate on admission) and neonatal deaths, weighing at least 1000g, compared to randomly selected women with healthy perinatal outcomes (Apgar score 7-10). Furthermore, to assess quality of routine clinical data collection and use, the translations from clinical assessments to case file recordings (including partograph use) and from case files to hospital registers will be analysed, by direct, structured observations and reviews of case files and registers. Finally, a broader assessment of outcomes and modes of birth will be conducted for all births during six months.

Co-creation process

The PartoMa intervention from Zanzibar, including guidelines and training, will be modified in accordance with birth attendants' and pregnant women's needs and circumstances in the Dar es Salaam context. Based on the results of the situational analysis, a first draft of the modified intervention will be developed by the research team and shared with co-creators, including skilled birth attendants, hospital management and women who have given birth at the facilities. Through focus group discussions, co-creators will share their perspectives, which will inform further modifications of the guidelines and training content, and possibly changes to the training implementation strategy.

The first implementation month in the first maternity unit to receive the intervention will function as pilot testing and might lead to further modifications.

Intervention implementation and evaluation

The overall implementation design will be a pragmatic, stepped-wedge cluster-randomized trial in which each of the five maternity units (hospitals) will be randomized into three clusters (1+2+2 hospitals). In a random order, the clusters will receive the intervention with three-month intervals. Informed by the logic model, the evaluation levels follow the Kirkpatrick model and include birth attendants' perceptions and reactions, changes in their learning and clinical care practices, and the resulting changes in birth outcomes.

Quantitative evaluation

During the training seminars, participants' perceptions, return rate and learning over time will be assessed by providing all attendants with a unique research number. As no per diems or allowances are paid and the seminars happen outside working hours, attendance will be considered an indicator of interest and motivation to learn about the intervention. At each quarterly seminar that health providers choose to attend, they will be asked to fill in a Likert-scale questionnaire-based evaluation on their experiences of the intervention and a questionnaire-based knowledge and partograph skills test (before and after the seminar).

The primary and secondary birth outcome variables will be prospectively measured from baseline and until 12 months after the last cluster receives the intervention, including stillbirths weighing at least 1000g with positive foetal heart rate on admission, Apgar score 1-6, and mode of birth. The outcome data will be gathered daily from hospital registers and medical records, and cross-checked with the facilities' routinely kept birth registers. Women referred to other facilities during labour will be followed up.

During baseline and the 5th and 10th implementation months, the structure for care delivery, usage of the PartoMa guidelines and elements of (dis)respectful care will be assessed by structured observations. To assess potential changes in quality of intrapartum care, criterion-based audits of case files will be conducted at each facility during baseline and the 9th-12th implementation months, among a randomly-selected subgroup of labouring women, and among births by emergency caesarean section and vacuum extraction. Case files will be identified and included during the continual daily collection of birth outcome data. Results of the situational analysis will inform final sample size calculations.

Qualitative evaluation

Relevant pathways and contextual factors contributing to the potential measurable changes in care and outcomes related to the PartoMa intervention will be unpacked. As in the situational analysis, this investigation will initially be limited to two of the facilities, but may broaden if emerging findings suggest relevance of studying other of the facilities (e.g. major differences in intervention use). This component will engage birth attendants and hospital leadership using different qualitative methods, a qualitative researcher will facilitate a series of task-based investigations: i. examine barriers and facilitators in the implementation of the intervention through individual interviews; ii. elicit 'stories of change' through photography and diary keeping; and iii. unpack collective experiences and perspectives on the strengths and limitations of the PartoMa guidelines and associated activities through focus group discussions.

Economic evaluation

A pragmatic cost-effectiveness analysis will be conducted from a health care and societal perspective, considering incremental costs and effects related to the intervention. Cost data will be collected for both the intervention co-creation and the implementation. Cost will relate to transport, time, commodities for trainings as well as running and capital costs for training facilities. Data will be collected by participant questionnaires and original receipts for direct and indirect costs as well as daily reports on activities of facilitators, trainers and experts throughout co-creation and implementation. Fixed hourly rates based on national salary scales will be used to calculate the costs per hour. Costs will be adjusted by fixed-time exchange rates and purchasing power parity. Effects will be measured in natural units related to maternal complications, intrahospital stillbirths, Apgar score below 7 and non-medically indicated caesarean sections. A discount rate of 3% will be applied to both costs and effect. Changes in birth outcomes (stillbirths and low Apgar score) will be transformed to Disability-Adjusted Life Years' (DALY's) of a birth cohort by using latest Burden of Disease disability weights and construction of a Markov model to estimate DALYs lost or averted in a lifetime perspective. Finally, the incremental cost-effectiveness ratio (ICER) will be calculated and related to alternative strategies for improving care during birth in a low-resource context, including the frequently used World Health Organization threshold of three times national Gross Domestic Product per DALY to estimate cost-effectiveness. Sub-group analysis will be conducted for mode of birth (vaginal, caesarian section, instrumental delivery). Base case of costs and effect will be explored in univariate sensitivity analysis. Non-parametric bootstrap method will be used to estimate the sampling distribution of the ICER and subsequently compute the cost-effectiveness acceptability curve across a range of cost-effectiveness thresholds.

Inclusion Criteria:

• All deliveries recorded at the five hospital sites during the entire study period.
• All women in labour delivering at the five study sites.
• All health care providers in delivery wards at the five study sites during the baseline and intervention period.
• For the different substudies, sub-groups are selected (please see the secondary outcomes for a description)

(Please notice that all women and newborn children will be included in the the study, irrespectively of their health status)

Exlusion Criteria:

• There are no exclusion criteria.