Linking Facility-based Mortality Audits With Community Engagement in Gilgit-Baltistan, Pakistan

A Pilot Study: Linking Facility-based Mortality Audits With Community Engagement to Improve Maternal and Newborn Outcomes in Gilgit-Baltistan, Pakistan

Aga Khan Foundation, Pakistan, Aga Khan Health Services, Centre for Global Child Health, SickKids Research Institute - Toronto, Canada, Aga Khan Development Network - Islamabad, Pakistan, Department of Health Gilgit-Baltistan - Pakistan, Aga Khan University, Nairobi, Kenya, Aga Khan Foundation, Canada

Pakistan is one of the countries in South Asia where neonatal mortality rates remain stagnant. Babies born in Pakistan encounter the highest risk of dying; of every 1,000 babies born, 46 die before the end of their first month (UNICEF, 2018). Some of the highest perinatal and neonatal mortality rates in Pakistan are found in districts of Pakistan's mountainous northern region (Bhutta ZA, 2013), where geography, climate and security risks make it challenging for women in remote communities to reach health services in a timely manner. According to 2013 PDHS, the neonatal and perinatal mortality rate in the northern area of Gilgit Baltistan was 39/1,000 and 37/1,000, respectively. In the rural area of Khyber Pakhtunkhwa, the neonatal and perinatal mortality rate was 42/1,000 and 63/1,000, respectively. Implementation of a health facility mortality audit cycle has proved successful in reducing perinatal mortality by upto 30% in other LMICs. Meanwhile evidence suggests that the most common factors contributing to high mortality rates are due to phase-one delays (delay in the decision to seek care). This study will attempt to operationalize linkages between the community and facility to not only improve facility-based quality of care, but to bring change in the community through community-feedback meetings to mitigate phase one and two delays and improve maternal, perinatal and neonatal outcomes. Data from this study will inform MoH policy decisions about standardized mortality audits with community feedback. Given the geographical location of Gilgit-Baltistan (GB) and accompanying constraints such as terrain and security, this study will attempt to operationalize linkages between the community and facility to not only improve facility-based quality of care, but to bring change in the community through community-feedback meetings to mitigate phase one and two delays and improve maternal, perinatal and neonatal outcomes. Data from this study will inform MoH policy decisions about standardized mortality audits with community feedback.

Pakistan has the highest global newborn mortality rate (45.6/1,000 live births), with 1 in 22 babies dying before the end of their first month (UNICEF, 2018). Some of the highest perinatal and neonatal mortality rates are found in districts of Pakistan's mountainous northern region, where geography, climate and security risks make it challenging for women in remote communities to reach health services in a timely manner. Implementation of a health facility mortality audit cycle has proved successful in reducing perinatal mortality by up to 30% in other LMICs. Meanwhile evidence suggests that the most common factors contributing to high mortality rates are due to phase-one delays. Given the geographical location of GB and accompanying constraints such as terrain and security, this study will attempt to operationalize linkages between the community and facility to not only improve facility-based delivery services, but to effect change at the community, through community-feedback meetings to mitigate phase one and two delays and improve maternal, perinatal and neonatal outcomes. Data from this study may help inform government policy decisions about standardized mortality audits with community feedback. Objectives of the study Broad objective: The overall aim of the study is to assess whether implementation of facility-based maternal, perinatal and neonatal mortality audits (systemic clinical reviews of near misses and deaths), in combination with targeted community engagement and awareness activities, increases the recognition of danger signs during pregnancy, the number of facility deliveries and encourages the discussion of a birthing plan, thereby reducing phase -one, two and three delays as compared to implementing facility-based mortality audits only. This will be assessed by monitoring births and outcomes in public and private health facilities and their associated catchment communities in GB. The study will also evaluate whether these targeted community-engagement activities, informed by the clinical audits improve the quality of delivery services through better utilization of local resources and improvements in clinical signal functions. Specific objectives Primary objective To determine the effect of a combined approach of facility-based audits (clinical reviews) linked to community engagement on core community behaviours/practices to improve obstetric, perinatal and neonatal outcomes as compared to standard perinatal and neonatal facility-based audits only and no facility-based audits. Key practices include: Recognition of danger signs among women during pregnancy, labour and delivery, after delivery and in newborns Actions taken for a birthing preparation plan Secondary objectives Assess facility deliveries Assess severe maternal morbidity outcomes during delivery (including maternal near misses ) Assess changes in quality of delivery services as measured through the delivery room checklist Assess changes in first delay (care-seeking decision) I. Delay due to lack of knowledge II. Lack of empowerment (sociocultural factors/barriers such as women's decision making, women's status) Assess changes in second delay (identification and reaching health facility) Assess changes in third delay (receiving adequate care and treatment at facilities) Assess perinatal mortality rates (fresh still birth or neonatal death in the first week of life) Assess neonatal mortality rates (deaths in the first 28 days of life)

This will be a three-arm cluster randomized controlled study to evaluate a novel intervention. Each facility is a cluster and includes all deliveries (including stillbirths and live births) in health facilities and respective catchment areas that provide obstetric and post-natal services.

Health facilities (clusters) will be the units of randomization. Within each stratum, health facilities will be randomly allocated to one of the three arms. Arm 1: includes morbidity and mortality health facility audits-only; arm 2: includes morbidity and mortality health facility audits in addition to community-engagement; and arm 3: will serve as the control arm and collect outcome data only.

Formalized audit teams with monthly meetings at each facility. Audit Intervention phases: Identification of facility leadership (i.e physicians or leading health care providers) to be trained on best practices in obstetric and perinatal care in the implementation of the audits. Training of identified audit leaders (main investigator is typically a physician) a. 5-day training workshop to include: i. Day 1: maternal death and near misses, perinatal, neonatal deaths and morbidity outcomes ii. Day 2: 'Three-delay' framework and identifying modifiable factors iii. Day 3: data collection and setting up the audit system iv. Days 4-5: mentoring on the identification of the audit teams and initiating audit implementation Creating audit team (composed of at least two obstetricians, 2 pediatricians plus the main investigator) Establishing and launching the audit cycle (monthly meetings) Annual re-certification of audit leaders

This intervention will implement the audits as described in arm 1; however, key community representatives (approximately 3-5 members) will be identified to attend monthly follow-up meetings with audit team leaders to discuss the community aspects (phase one and two-delays) that could have prevented the death, near miss or severe adverse outcome. Information regarding the cause(s) of death in the community will be collected by the LHW or CMW, who reports to the health facility on a monthly basis.

Monthly outcome data will be collected from health facilities where no audits will be conducted. Data from this group will help evaluate the effectiveness of the intervention arms on perinatal and neonatal mortality.

This intervention will implement the audits as described in arm 1; however, key community representatives (approximately 3-5 members) will be identified to attend monthly follow-up meetings with audit team leaders to discuss the community aspects (phase one and two-delays) that could have prevented the death, near miss or severe adverse outcome. Information regarding the cause(s) of death in the community will be collected by the LHW or CMW, who reports to the health facility on a monthly basis.

Change in proportion of women who can correctly identify at least 3 danger signs at each stage across the continuum of care

Change in proportion of women who can correctly identify at least 3 danger signs at each stage across the continuum of care

Change in proportion of women for which at least 5 birthing plan actions were taken for the birth of the index child

Change in proportion of women for which at least 5 birthing plan actions were taken for the birth of the index child: Discussed place of delivery, Discussed who will perform delivery, Set aside funds for delivery, Set aside alternate funds for costs of skilled and emergency care, Made arrangements for transport, Identified a blood donor, Discussed accompaniment to planned delivery location.

Change in severe maternal morbidity outcomes: proportion of assisted deliveries, emergency caesarean sections, transfusions and hysterectomies.

Inclusion Criteria: Women aged 15-49 years who have had a pregnancy in the last 12 months who reside in the five targeted districts. Public and private health facilities in five districts in GB that offer obstetric and postnatal care, and respective catchment areas are supported by LHWs, Lady Health Volunteers (LHVs), CMWs, and CHWs will be included in the study. Data for all maternal 'near misses', perinatal and neonatal mortality and morbidity outcomes will be recorded for all women and newborns who deliver at home (through LHW, LHV, CMW monthly reports) and who contact the health facility within 42 days post-delivery, regardless of whether or not they delivered in the health facility. Exclusion Criteria: Women aged 15-49 years who have not had a pregnancy in the last 12 months, who reside in the five targeted districts. Health facilities in five districts of GB that do not provide any obstetric and postnatal care and are not affiliated with any LHWs, or CMWs will be excluded.