Spatial patterns of delivery service utilisation: using a spatial interaction model approach to reach every woman in Eastern Region, Ghana.
Lead Research Organisation:
University of Southampton
Department Name: School of Geography
Abstract
Interventions for reducing the maternal mortality ratio to less than 70 per 100 000 live births by 2030 require information to improve service quality, bridge access and utilisation gaps, and adequately plan maternal health services. This study would be of immense benefit to health service delivery not only in Ghana, but in other countries with similar healthcare information as well as being of academic interest. It would provide the health service with information that might change previous perceptions of utilisation patterns, thereby enhancing plans to reduce maternal deaths. For instance, understanding the interaction between healthcare quality and utilisation would enable targeted improvement in quality care, hence saving costs associated with wholesale implementation of interventions. The output of this study would inform a fair distribution of skilled delivery services to reach every woman, thereby averting maternal deaths.
Utilisation patterns among hospitals have been explored in the MSc work but primary lower level health facilities contribute more than one third of total births in the region. Hence further analysis of data would reveal the effect of primary health facilities on the overall utilisation patterns. The MSc report revealed a high influx of clients from the periphery of Greater Accra region closer to Nsawam hospital. Advanced Bayesian statistical models would be implemented to predict utilisation of delivery services in the Eastern region. This approach would provide a framework for estimation of utilisation patterns.
Objectives
The goal of this study is to use spatial interaction modelling approach to understand utilisation of delivery services. This goal would be achieved through analysis of data to achieve the following objectives:
1. To formulate a Bayesian model that estimates the probability of utilisation of delivery services.
2. To estimate the quality of maternal health care delivery in health facilities.
3. To delineate realistic catchments for delivery services based on observed flows.
4. To quantify the bypassing of skilled delivery facilities and determine individual factors that promote it.
Methods
Data sources
The following sources would be consulted for data to meet the set objectives:
1. With the support of Dr. Quansah, Ghana health service would be contacted for routine delivery records
2. Health facilities would be surveyed to create quality of care indices
3. Gridded surfaces of pregnancy counts would be modelled in collaboration with WorldPop.
4. Women who have interacted with delivery services would be interviewed to determine individual limitations to accessing care and calibrate drive times based on type of transportation.
Utilisation patterns among hospitals have been explored in the MSc work but primary lower level health facilities contribute more than one third of total births in the region. Hence further analysis of data would reveal the effect of primary health facilities on the overall utilisation patterns. The MSc report revealed a high influx of clients from the periphery of Greater Accra region closer to Nsawam hospital. Advanced Bayesian statistical models would be implemented to predict utilisation of delivery services in the Eastern region. This approach would provide a framework for estimation of utilisation patterns.
Objectives
The goal of this study is to use spatial interaction modelling approach to understand utilisation of delivery services. This goal would be achieved through analysis of data to achieve the following objectives:
1. To formulate a Bayesian model that estimates the probability of utilisation of delivery services.
2. To estimate the quality of maternal health care delivery in health facilities.
3. To delineate realistic catchments for delivery services based on observed flows.
4. To quantify the bypassing of skilled delivery facilities and determine individual factors that promote it.
Methods
Data sources
The following sources would be consulted for data to meet the set objectives:
1. With the support of Dr. Quansah, Ghana health service would be contacted for routine delivery records
2. Health facilities would be surveyed to create quality of care indices
3. Gridded surfaces of pregnancy counts would be modelled in collaboration with WorldPop.
4. Women who have interacted with delivery services would be interviewed to determine individual limitations to accessing care and calibrate drive times based on type of transportation.
Publications
Aheto JMK
(2022)
Multilevel analysis of predictors of multiple indicators of childhood vaccination in Nigeria.
in PloS one
Boateng LA
(2020)
Healthcare-seeking behaviour in reporting of scabies and skin infections in Ghana: A review of reported cases.
in Transactions of the Royal Society of Tropical Medicine and Hygiene
Dotse-Gborgbortsi W
(2022)
Distance is "a big problem": a geographic analysis of reported and modelled proximity to maternal health services in Ghana
in BMC Pregnancy and Childbirth
Dotse-Gborgbortsi W
(2020)
Spatial inequalities in skilled attendance at birth in Ghana: a multilevel analysis integrating health facility databases with household survey data.
in Tropical medicine & international health : TM & IH
Dotse-Gborgbortsi W
(2022)
Delineating natural catchment health districts with routinely collected health data from women's travel to give birth in Ghana.
in BMC health services research
Dotse-Gborgbortsi W
(2022)
Dam-mediated flooding impact on outpatient attendance and diarrhoea cases in northern Ghana: a mixed methods study
in BMC Public Health
Dotse-Gborgbortsi W
(2019)
The influence of distance and quality on utilisation of birthing services at health facilities in Eastern Region, Ghana.
in BMJ global health
Dzodzomenyo M
(2018)
Sachet water quality and product registration: a cross-sectional study in Accra, Ghana.
in Journal of water and health
Rice BL
(2021)
Variation in SARS-CoV-2 outbreaks across sub-Saharan Africa.
in Nature medicine
Semey M
(2020)
Characteristics of packaged water production facilities in Greater Accra, Ghana: implications for water safety and associated environmental impacts
in Journal of Water, Sanitation and Hygiene for Development
Utazi C
(2021)
District-level estimation of vaccination coverage: Discrete vs continuous spatial models
in Statistics in Medicine
Utazi CE
(2021)
District-level estimation of vaccination coverage: Discrete vs continuous spatial models.
in Statistics in medicine
Yu W
(2021)
Mapping access to basic hygiene services in low- and middle-income countries: A cross-sectional case study of geospatial disparities
in Applied Geography
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| ES/P000673/1 | 01/10/2017 | 30/09/2027 | |||
| 2123367 | Studentship | ES/P000673/1 | 01/10/2018 | 31/12/2022 | Winfred Dotse-Gborgbortsi |
| Description | The outputs of my research predicted the probability of skilled attendance at birth in Ghana as a map layer, with a fine 100 by 100 square meter spatial resolution. The detailed map identifies specific communities with inadequate coverage of essential birthing services. In addition, the high-resolution mapping enables aggregation of the results to any administrative unit. In exploring the relative roles of health service quality versus geographic access to healthcare, the study concluded that, given the expansion of primary care facilities in Ghana, higher quality healthcare services, as opposed to closer proximity of facilities to women, are needed to widen skilled birth uptake and improve maternal health. |
| Exploitation Route | Implementing recommendations made in research publications and policy briefs. Public health practitioners can replicate the methods used in my project to analyse service utilisation. For instance, the UNFPA is using the methods to analyse data for sexual and reproductive health services in several other countries. |
| Sectors | Healthcare |
| Description | Summary of non-academic impact My ESRC-funded research uses geospatial methods to investigate spatial patterns of birthing services in Ghana. It aims at improving access and utilisation of health services to promote health and wellbeing (SDG Goal-3). My published MSc (1+3) dissertation (five citations) developed approaches to integrate several datasets to understand maternal health service utilisation. The analysis provides insights into the relative importance of health service quality versus geographic access to healthcare facilities. Furthermore, the research reveals spatial inequalities in access to essential obstetric care services, thus improving access to quality essential health services (SDG Target-3.8). Although my research impact is most directly in Ghana, the methods are applicable in similar low-income settings. The main impacts are 1. National impact on maternal and newborn healthcare delivery in Ghana: My research has informed decisions by Ghana Health Services (GHS) to improve health services' distribution and quality to reduce maternal and neonatal mortality (SDG Target-3.1 and 3.2). 2. Instrumental international impact on UNFPA practice and policy: The UNFPA highlighted what they describe as "innovative approaches" from our paper in an international policy document and used these to investigate access and use of reproductive health services in several countries. 3. Capacity building: I shared skills learnt through the PhD by training public health staff in two districts in Ghana on routine health data management. Likewise, I facilitated mass nationwide health information management training for public health officers in Ghana. Secondary impacts are: 4. Impact on public understanding of infectious disease outbreaks: I featured in a BBC documentary discussing interdisciplinary approaches to outbreak investigation. I spoke about spatial epidemiology, a skill I improved through the PhD. 5. Open data uptake: Through collaboration with Ghana Health Service and mapping volunteers, the PhD project generated freely available spatial data on road network quality to improve travel time models in Ghana. 37. Extent and significance of the impact The extent of the impact 1. National impact on maternal and newborn healthcare delivery in Ghana: My PhD research feeds directly into recent interventions to reorient the health system collaboration, resource sharing and referral patterns. Our research shows the relative importance of quality and proximity on the use of maternal health services in all 261 districts in Ghana. The results are used by GHS to make decisions as we identify locations with poor maternal health quality, communities living far from services and how it affects service use. The Deputy Director General of GHS describes my latest paper on delineating health districts as "strong evidence underpinning a new initiative to group health facilities and communities into a network of practice for better health service delivery." 2. Instrumental international impact on UNFPA practice and policy Building on my PhD research, my analysis demonstrated how national averages in human resource capacity and health outcomes can be unmasked using geospatial analysis. A presentation to UNFPA led to me. contributing to WHO/UNFPA's "2021 State of the World's Midwifery" report on these methods (page 44). UNFPA Technical Specialist Sainan Zhang notes "we identified methods in Winfred's study on modelling spatial inequalities in access to skilled birth in Ghana as being of key value for a multi-country project. Winfred's research contributes to reducing maternal mortality," 3. Sharing expertise and Capacity-building: Drawing on the skills developed through my PhD, I trained ten public health staff face-to-face in two districts in Ghana. I then facilitated online training with 1,320 participants as part of continuous professional development accreditation for public health practice in Ghana (https://www.gaphto.org/2021/08/31/continuous-professional-development-cpd/). I network with other early career researchers particularly from West Africa for example by supporting a PhD student from The Gambia who studies at LSHTM, leading to publications (doi: 10.1371/journal.pone.0245297). |
| First Year Of Impact | 2021 |
| Sector | Healthcare |
| Impact Types | Policy & public services |
| Description | Ghana Association of Public Health Technical Officers CPD Training |
| Geographic Reach | National |
| Policy Influence Type | Influenced training of practitioners or researchers |
| Impact | About 1,320 officers attended the online training in August 2021. The trainings are part of the requirements to renew registration for all allied health professionals in Ghana. The trainings were held online with participants from all 16 regions nationwide. |
| URL | https://www.gaphto.org/2021/08/31/continuous-professional-development-cpd/ |
| Description | Supporting PhD student from Gambia |
| Geographic Reach | Africa |
| Policy Influence Type | Influenced training of practitioners or researchers |
| Impact | According to the Plos journal metrics, the paper has been viewed 2,209 times and cited four times. This findings are used to improve neonatal care in Nigeria. |
| URL | https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0245297 |
| Title | Integration of multiple data sources |
| Description | Our analysis of data demonstrated for the first time how large demographic datasets can be integrated with health facility census and routine data. |
| Type Of Material | Data analysis technique |
| Year Produced | 2020 |
| Provided To Others? | No |
| Impact | This analysis technique has led to discussions with the UNDP Headquarters to apply similar techniques to analysing access to and use of sexual and reproductive health services. |
| URL | https://onlinelibrary.wiley.com/doi/10.1111/tmi.13460 |
| Description | Countdown2030 |
| Organisation | African Population and Health Research Center |
| Country | Kenya |
| Sector | Academic/University |
| PI Contribution | The Countdown to 2030 collaboration in Ghana includes the University of Ghana, the Ghana Health Service, and WorldPop at the University of Southampton. It aims to enhance the capacity of the Ghana Health Service to provide quality evidence through targeted analysis during national and subnational reviews of overall sector progress, as well as performance review of the country's RMNCAH+N plan. |
| Collaborator Contribution | The Countdown consortium allocates funding for activities and facilitates collaboration between other organisations, universities and research institutions involved in the project. |
| Impact | Outcomes Acheived 1. Trained ten public health staff in two districts in Ghana. The training was on routine health data management and annual report writing. The training developed the capacity of Ghana Health Service staff to improve routine reporting of health outcomes in the selected districts. 2. A paper has been drafted on socio-demographic inequalities in proximity to health services in Ghana 3. Participating in a special issues on access to health in urban cities in Africa. I review abstracts and papers from Dar es Salam, Tanzania. Expected outputs Two analytical workshops, one on denominator estimation and mapping. An analytical report which will detail progress and performance made towards national and international targets for key indicators at the national and district levels. The analysis will focus on recent coverage, trends, inequalities and geospatial analysis on selected indicators One policy brief A peer-reviewed journal article on critical topics from the workshop. Routine annual progress reports incorporating new methodologies and corresponding estimates. Inputs to performance-based management in the context of the universal healthcare roadmap MLE framework. An analytical report to track country and international targets for key indicators based on surveys, health facility data, and administrative data, focusing on recent trends and inequalities at national and subnational levels (up to sub-district level), including longterm projections. |
| Start Year | 2020 |
| Description | Final Year South Coast DTP Conference |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Postgraduate students |
| Results and Impact | The audience were from Brighton, Portsmouth and Southampton University. They were from different backgrounds such as social statistics, geography, psychology, linguistics, economics, business, education and other social science backgrounds. There were discussions after my presentation which was focused on how I network and engage with stakeholders involved in my studies in Ghana. |
| Year(s) Of Engagement Activity | 2021 |
| Description | School Seminar on maternal health |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | The seminar was for the school of Geography. There were academics and research students present. I shared results from my data collection in Ghana and discussed interesting findings. I shared strategies of collecting data in low resource settings. There were discussions on how to improve the research using other sources of data. The presentation abstract is below: Speaker: Mr Winfred Dotse-Gborgbortsi - Postgraduate Research Student within Geography and Environmental Science at the University of Southampton. Title: Quality of emergency obstetric care in Ghana Abstract: Although women giving birth in health facilities has increased over the years in Ghana, maternal deaths are still unacceptably high. Women dying in health facilities while giving birth brings to question the quality of emergency obstetric care. Are emergency obstetric services available and ready when there is a complication? Emergency obstetric and newborn care (EmONC) signal functions are the standard indicators for assessing health facility readiness in emergency cases. However, recent literature has proposed a broader range of indicators including routine services and the capacity of health facilities, all of which may affect patient choice of healthcare facility. In this seminar, we will discuss a recent maternal health service provision assessment (SPA) of 151 health facilities. The SPA was used to develop a quality care index based on a broader set of domains. Two approaches, an index of multiple deprivation and principal component analysis were used to analyse the data. Preliminary visualisations of the origins and destinations of over 40,000 women giving birth in these health facilities will also be discussed. Initial results show mostly secondary care facilities in higher quality quintiles, where domains measured capacity or physical size of health facilities. However, there is more variation in domains assessing training, water sanitation and hygiene, privacy, and staff motivation. Facility size or tier thus serves as a poor proxy for these aspects of service quality. Going forwards, the study will assess how these quality indicators affect healthcare utilisation through spatial interaction modelling. |
| Year(s) Of Engagement Activity | 2021 |
| Description | Visit to Ghana Health Service, Eastern Regional Office |
| Form Of Engagement Activity | Participation in an open day or visit at my research institution |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Professional Practitioners |
| Results and Impact | I visited my study area's regional public health office where I met with regional officers. The meetings include the regional director of health services, the deputy regional director in charge of public health, the regional research officers, the regional maternal health focal persons, and the regional health information office. I engaged with 11 persons in total. We discussed research findings from my previous papers that benefit the region. Furthermore, the regional officers made inputs to my data collection by suggesting relevant areas and indicators to collect. This ensured that the project does not only narrow on academic outcomes but to also improve health service delivery within the region. |
| Year(s) Of Engagement Activity | 2021 |