Place-specific factors associated with adverse maternal and perinatal outcomes in Southern Mozambique: A retrospective cohort study

listen audio

Study Justification:
The study aimed to identify and measure the place-specific determinants associated with adverse maternal and perinatal outcomes in the southern region of Mozambique. This information is crucial for designing targeted interventions that address the social determinants of maternal health in the study area. By understanding the specific factors contributing to adverse outcomes, policymakers and healthcare providers can develop strategies to improve maternal and perinatal health in the region.
Highlights:
– The study conducted a retrospective cohort analysis to examine the relationship between various variables and adverse maternal and perinatal outcomes.
– Six variables were found to be statistically significant in explaining the combined outcome, including geographic isolation, flood proneness, access to an improved latrine, average age of reproductive age woman, family support, and fertility rates.
– The ordinary least squares model explained 69% of the variation in adverse outcomes, and geographically weighted regression increased the adjusted R2 to 71%, accounting for geographic variability.
– The study emphasized the importance of considering context-specific determinants of maternal health and the need for multisectoral collaboration in addressing these determinants.
Recommendations:
– Design targeted interventions that address the place-specific social determinants of maternal health in the study area.
– Collaborate with various sectors, including healthcare, infrastructure, and community support, to implement comprehensive strategies.
– Prioritize interventions that address geographic isolation, flood proneness, access to improved latrines, and family support.
– Consider the average age of reproductive age women and fertility rates when developing interventions.
– Use geographically weighted regression to account for geographic variability and tailor interventions to specific areas.
Key Role Players:
– Ministry of Health: Responsible for coordinating and implementing interventions related to maternal and perinatal health.
– Community Health Workers: Play a crucial role in delivering targeted interventions and providing support to pregnant women and their families.
– Infrastructure Development Agencies: Involved in improving access to healthcare facilities, latrines, and transportation in the study area.
– Non-Governmental Organizations: Provide additional support, resources, and expertise in implementing interventions and addressing social determinants of health.
Cost Items for Planning Recommendations:
– Infrastructure Development: Budget for improving access to healthcare facilities, latrines, and transportation in the study area.
– Training and Capacity Building: Allocate funds for training community health workers and healthcare providers on delivering targeted interventions.
– Community Engagement and Awareness: Set aside a budget for community outreach programs, workshops, and campaigns to raise awareness about maternal and perinatal health.
– Data Collection and Monitoring: Allocate resources for collecting and analyzing data on maternal and perinatal outcomes to monitor the effectiveness of interventions.
– Research and Evaluation: Set aside funds for further research and evaluation to assess the impact of interventions and identify areas for improvement.
Please note that the cost items provided are general categories and not actual cost estimates. The specific budget requirements will depend on the scale and scope of the interventions implemented.

The strength of evidence for this abstract is 7 out of 10.
The evidence in the abstract is moderately strong. The study is a retrospective cohort study conducted as part of a larger trial, which adds to its credibility. The study identified six statistically significant variables associated with adverse maternal and perinatal outcomes. The adjusted R2 value of 0.69 indicates a good fit of the model. However, the abstract does not provide specific details about the sample size, data collection methods, or statistical analysis techniques used. To improve the evidence, the abstract should include these details and provide more information about the study’s limitations.

Objectives To identify and measure the place-specific determinants that are associated with adverse maternal and perinatal outcomes in the southern region of Mozambique. Design Retrospective cohort study. Choice of variables informed by literature and Delphi consensus. Setting Study conducted during the baseline phase of a community level intervention for pre-eclampsia that was led by community health workers. Participants A household census identified 50 493 households that were home to 80 483 women of reproductive age (age 12-49 years). Of these women, 14 617 had been pregnant in the 12 months prior to the census, of which 9172 (61.6%) had completed their pregnancies. Primary and secondary outcome measures A combined fetal, maternal and neonatal outcome was calculated for all women with completed pregnancies. Results A total of six variables were statistically significant (p≤0.05) in explaining the combined outcome. These included: Geographic isolation, flood proneness, access to an improved latrine, average age of reproductive age woman, family support and fertility rates. The performance of the ordinary least squares model was an adjusted R 2 =0.69. Three of the variables (isolation, latrine score and family support) showed significant geographic variability in their effect on rates of adverse outcome. Accounting for this modest non-stationary effect through geographically weighted regression increased the adjusted R 2 to 0.71. Conclusions The community exploration was successful in identifying context-specific determinants of maternal health. The results highlight the need for designing targeted interventions that address the place-specific social determinants of maternal health in the study area. The geographic process of identifying and measuring these determinants, therefore, has implications for multisectoral collaboration.

The study was conducted as part of the feasibility study for the Community-Level Interventions in Pre-Eclampsia Trial (CLIP) in 2014 in Mozambique. CLIP was a community-based cluster randomised controlled trial aimed at reducing all-cause maternal and perinatal mortality and morbidity in the study region. CLIP was led by the University of British Columbia, in partnership with the Centro de Investigação em Saúde de Manhiça in Mozambique. The feasibility study for CLIP was conducted in 36 administrative regions termed localities within two provinces in the southern part of the country. There were four core aspects of this project that are summarised in figure 1: (1) gathering data on community perspectives of the determinants of maternal health, (2) prioritising variables through a Delphi consensus, (3) collecting primary empirical data on the variables and (4) conducting spatial and statistical analyses to explore the association of these variables with adverse maternal outcomes. Design overview. Ten focus groups discussions (FGDs) were conducted in 4 of the 12 clusters in the CLIP study area: Messano, 3 de Fevereiro, Ilha Josina+Calanga and Chongoene. These FGDs involved pregnant women, women of reproductive age, matrons (local birth attendants), male partners, community leaders and community-based health workers. Using purposive sampling combined with snowball sampling techniques, participants for the FGDs were recruited with the assistance of community gatekeepers. The FGDs covered topics regarding the sociocultural, environmental and economic factors thought to be related to adverse maternal events. Semistructured interviews were conducted with the chiefs in all 12 administrative posts in the study region to better understand the historical context (eg, civil wars, natural disasters, foreign aid and microfinance) of the communities and how these could impact maternal health. The FGDs and interviews were conducted in a local language (Changana) following a guide of open-ended questions reflecting the topics, were audio-recorded, transcribed and translated verbatim into Portuguese before a final translation into English. The full details concerning data collection, coding of the data and thematic analysis have been previously published.24 A Delphi consensus meeting by teleconference was conducted to prioritise the variables for statistical analysis. The Delphi technique helps with ‘achieving convergence of opinion concerning real-world knowledge solicited from experts within certain topic areas’.25 The panel of 17 experts had a range of relevant backgrounds, including obstetrics (n=2), epidemiology (n=2), demography (n=2), health geography (n=1), environmental health (n=1), spatial statistics (n=1), health equity (n=1), health systems research (n=1), medical anthropology (n=4) and mobile health (n=2). A structured questionnaire that had been designed based on an extensive literature review was used as a guide for the Delphi process during the teleconference. The same questionnaire was sent to members of the Delphi group that could not make the call. Consensus was reached after the first round as many of the variables tabled before the experts were backed by literature. The context-specific variables identified for consideration of their association with a combined maternal and perinatal adverse outcome were collected through a household census conducted as part of the CLIP feasibility study.26 The census included information on all women who had been pregnant in the 12 months prior to the census, as well as women of reproductive age who had died. Data collected included individual-level variables (eg, age, education and pregnancy history), as well as community characteristics (eg, availability of the household head and community support initiatives). All reports of maternal, fetal or perinatal deaths were followed up with verbal autopsy27 to classify the cause. The census identified 50 493 households that were home to 80 483 women of reproductive age (age 12–49 years). Of these women, 14 617 had been pregnant in the 12 months prior to the census, of which 9172 (61.6%) had completed their pregnancies. For the mother, there were 18 deaths (204.6 MMR) of which the verbal autopsy identified that 38% were from direct causes and 62% from indirect causes. For the baby, there were 288 (3.0%) miscarriages, 466 (4.9%) stillbirths and 8796 (92.1%) live births, of which there were 117 neonatal deaths. A full description of the health and sociodemographic profile of the women of reproductive age in the study area has been published.26 In addition, we collected five geographical variables using geographical information systems. There were three travel times to (1): primary health facilities, (2) secondary health facilities and (3) tertiary health facilities, using mixed transport modes for public transport and ambulances. Walking times to the nearest main road (4) were calculated to measure the degree to which communities were geographically isolated. Finally, an indicator for flood proneness (5) was designed based on flood and precipitation records from the previous year.28 These variables and other community-level estimates for the variables captured in the census were calculated for each locality in the study area as described in table 1. Both the census and geographical data were aggregated into community-level averages at the locality level for each of the chosen variables, as ethical approval did not allow to analyse the location data at the level of the individual woman. Community level variables potentially associated with the rates of adverse maternal outcomes The primary outcome for this study was a combined maternal and perinatal outcome that included maternal, fetal and neonatal deaths. The denominator was the total number of live births. A composite outcome was chosen as powering the study for maternal death alone would have required a prohibitively large sample size. There is clinical plausibility in combining the three outcomes as both fetal and early neonatal outcomes are related to the woman’s condition during the antenatal and intrapartum periods, while her environment and sociocultural circumstances have an impact on late neonatal outcomes.29 30 Furthermore, a significant proportion of miscarriages are related to placental dysfunction, as are stillbirths and neonatal deaths.31 32 The spatial statistics module within ArcGIS software33 was used for exploratory regression to further prioritise variables and to create the global ordinary least squares (OLS) regression model. The exploratory regression exercise evaluated different combinations of our explanatory variables for their fit for an OLS model and how these explained trends in our outcome variable. This method implements the exploration by screening variables in a forward stepwise sequence, exploring how different combinations of variables fit and perform in the regression model. Using criteria that assessed p values significance, multicollinearity measured by the variance inflation factor (VIF), normality of residuals and clustering of residuals in space (table 2), we selected the variables that best explained the outcome and met the criteria of a well-specified regression model and explored these through a more rigorous OLS modelling exercise. Criteria for variable selection prior to regression modelling The performance of the OLS models chosen from the exploratory regression were assessed based on the magnitude of the adjusted R2 values. In addition, we checked for significance of p values for the model coefficients. Multicollinearity between different variables in a model was checked using the lower VIF threshold of five. The Koenker statistic (p<0.01) was used to check if the relationships being modelled were consistent (either due to non-stationarity or heteroskadisticity), while the Wald statistic was used to assess overall model significance. The Jarque Bera test (p<0.01) was used to check if model predictions were biased (ie, if the model residuals were normally distributed). The model that performed best and met these criteria was selected for further analysis to create a locally specified model. The geographically weighted regression (GWR) technique was used to develop a second model, which extended the output from OLS, to explore spatial non-stationarity of effect of the variables. This allowed for the new model to account for spatial structure in estimating local rather than global model parameters.34 35 We foresee this to be an important step to creating interventions that are locally specific and an important part of more precisely targeting interventions. As part of the modelling process, the spatial weights based on the geographic proximity of observation are applied to give more weight to values that were closer together. GWR4 software36 was used for this part of the project. The geographic variability test was conducted to assess if there was significant non-stationarity in the coefficients after applying GWR. This test compares the geographically varying parameters with those in the fixed global model, where a negative difference (abbreviated ‘DIFF OF CRITERION’ in GWR4), indicates significant variation in parameter estimates across space.36 We also assessed the performance of the GWR model using the newly calculated values of the adjusted R2. Previous research24 that elicited community perceptions of risk factors related to adverse maternal outcomes from women, their male counterparts and community leaders in the study area informed the choice for some of the variables. This study used FGDs and in-depth interviews and has been published through other avenues. The same results have also been disseminated through an outreach workshop to the ministry of health in Mozambique. Further research that has received recent funding will communicate the identified risk factors through the use of a mobile app.

Based on the information provided, here are some potential innovations that could improve access to maternal health:

1. Mobile Health (mHealth) App: Develop a mobile application that communicates the identified risk factors for adverse maternal outcomes to pregnant women and their families. This app could provide educational resources, reminders for prenatal care appointments, and access to telemedicine consultations with healthcare providers.

2. Community-Based Health Workers: Expand the role of community health workers in the study area to include maternal health interventions. These workers could provide home visits, education, and support to pregnant women and their families, particularly in remote or isolated areas.

3. Targeted Interventions: Design and implement targeted interventions that address the place-specific social determinants of maternal health in the study area. This could involve collaborating with local community leaders, organizations, and government agencies to address factors such as geographic isolation, flood proneness, and access to improved sanitation facilities.

4. Improving Access to Healthcare Facilities: Explore strategies to improve access to primary, secondary, and tertiary healthcare facilities in the study region. This could include improving transportation infrastructure, increasing the availability of ambulances, and reducing travel times to healthcare facilities.

5. Geographically Weighted Regression (GWR): Utilize the findings from the GWR analysis to develop locally specific interventions. By understanding the spatial non-stationarity of the determinants of adverse maternal outcomes, interventions can be tailored to the specific needs and challenges of different communities within the study area.

These innovations aim to address the identified determinants of adverse maternal and perinatal outcomes in Southern Mozambique and improve access to maternal health services in the region.
AI Innovations Description
Based on the information provided, the recommendation to develop an innovation to improve access to maternal health in Southern Mozambique is to design targeted interventions that address the place-specific social determinants of maternal health in the study area. The study identified several variables that were statistically significant in explaining adverse maternal and perinatal outcomes, including geographic isolation, flood proneness, access to an improved latrine, average age of reproductive age woman, family support, and fertility rates. These variables showed significant geographic variability in their effect on rates of adverse outcomes. Therefore, it is important to consider the specific context and characteristics of each locality when developing interventions. This can be achieved through multisectoral collaboration and the use of geographically weighted regression to account for spatial non-stationarity of the variables. By targeting the social determinants of maternal health in a place-specific manner, interventions can be more effective in improving access to maternal health services and reducing adverse outcomes.
AI Innovations Methodology
Based on the information provided, here are some potential recommendations to improve access to maternal health:

1. Improve geographic accessibility: Address the issue of geographic isolation by implementing strategies to improve transportation infrastructure and access to healthcare facilities in remote areas. This could include building new roads, increasing the availability of public transportation, or implementing telemedicine services.

2. Enhance community support initiatives: Strengthen community-based health worker programs and community engagement to provide support and education to pregnant women. This could involve training and empowering local community members to provide basic maternal healthcare services, conduct health education sessions, and facilitate referrals to healthcare facilities.

3. Increase access to improved sanitation facilities: Promote the construction and use of improved latrines in communities to reduce the risk of infections and improve overall maternal health outcomes.

4. Address social determinants of health: Develop targeted interventions that address the specific social determinants of maternal health in the study area. This could involve initiatives to improve education and awareness about maternal health, address cultural practices that may negatively impact maternal health, and provide economic support to vulnerable populations.

To simulate the impact of these recommendations on improving access to maternal health, a methodology could be developed as follows:

1. Data collection: Gather data on key indicators related to maternal health, such as maternal mortality rates, access to healthcare facilities, availability of transportation, and community support initiatives. This data can be collected through surveys, interviews, and existing health records.

2. Establish a baseline: Analyze the current state of access to maternal health services and outcomes in the study area. This will serve as a baseline against which the impact of the recommendations can be measured.

3. Develop a simulation model: Use statistical and spatial analysis techniques to develop a simulation model that incorporates the identified recommendations and their potential impact on access to maternal health. This model should consider factors such as population distribution, geographic features, transportation networks, and community characteristics.

4. Simulate scenarios: Run the simulation model with different scenarios that reflect the implementation of the recommendations. This could involve adjusting variables such as transportation infrastructure, community support initiatives, and availability of improved latrines. The model should simulate the potential changes in access to maternal health services and outcomes based on these scenarios.

5. Analyze results: Evaluate the simulated outcomes of each scenario to assess the potential impact of the recommendations on improving access to maternal health. Compare the results to the baseline to determine the effectiveness of the recommendations in achieving the desired outcomes.

6. Refine and iterate: Based on the analysis of the simulation results, refine the recommendations and simulation model as needed. Iterate the simulation process to further optimize the strategies for improving access to maternal health.

By following this methodology, policymakers and healthcare providers can gain insights into the potential impact of different recommendations on improving access to maternal health. This can inform decision-making and help prioritize interventions that are most likely to have a positive impact on maternal health outcomes in the study area.

Share this:
Facebook
Twitter
LinkedIn
WhatsApp
Email