https://orcid.org/0000-0002-1572-8104
Abstract
Consideration of health equity is fundamental to enhancing the health of those who are economically/socially disadvantaged. A vital characteristic of health equity and therefore health disparity is the level of spatial access to health services and its distribution among populations. Adequate knowledge of health disparity is critical to enhancing the optimal allocation of resources, identification of underserved populations and improving the efficiency and performance of the health system. The provision of such insight for sub-Saharan African (SSA) cities is a challenge and is severely limited in the literature. Accordingly, this study examined the disparities in potential spatial access to health services for four selected urban areas in Ghana based on: (1) the number of physicians per population; (2) access score based on a weighted sum of access components; (3) travel time to health services and (4) the combined evaluation of linkages between travel distance, settlement area, population and economic status. The overall spatial access to health services is low across all selected cities varying between 3.02 and 1.78 physicians per 10 000 persons, whereas the access score is between 1.70 and 2.54. The current number of physicians needs to be increased by about five times to satisfy the World Health Organization’s standard. The low spatial access is not equitable across and within the selected cities, where the economically disadvantaged populations were found to endure longer travel distances to access health services. Inequities were found to be embedded within the selected cities where economically poor populations are also disadvantaged in their physical access to healthcare. The health facilities in all cities have reasonable travel distances separating them but are inadequately resourced with physicians. Thus, increasing the physician numbers and related resources at spatially targeted existing facilities would considerably enhance spatial access to health services.
Synthesized the disparities in potential spatial access to health services.
Established the linkages between access disparities to health services and economic status.
Linked spatial access to health services to settlement area, population and economic status.
Introduction
According to Healthy People (2020), health disparity is defined as ‘a particular type of health difference that is closely linked with economic, social, or environmental disadvantage’ The commitment to eliminate these disparities in health and its determinants, including social determinants, has been referred to as health equity (Braveman and Gruskin, 2003; Braveman, 2014). In relation to the constitution of the World Health Organization (WHO), Whitehead (1992) and Whitehead and Dahlgren (2006) noted that equity in health ‘implies that ideally, everyone could attain their full health potential and that no one should be disadvantaged from achieving this potential because of their social position or other socially determined circumstance’. Health equity and health disparity are interconnected such that the latter is the metric that is used to measure the former, and that greater equity is achieved by improving the health of those who are economically/socially disadvantaged, not by a worsening of the health of those in advantaged groups (Braveman and Gruskin, 2003; Braveman, 2014).
One key aspect of health disparity and therefore health equity is the level of physical or spatial accessibility to health services and its distribution among population groups. Accessibility to health services has been recognized as a health priority worldwide as it facilitates the population health of communities (Penchansky and Thomas, 1981; Luo and Wang, 2003; Guagliardo, 2004; Humphreys and Smith, 2009; Wong et al., 2012). Spatial access has been described using several dimensions (Luo and Wang, 2003; Guagliardo, 2004; Wang and Luo, 2005; Ngui and Apparicio, 2011; Apparicio et al., 2017), but the focus of this study is potential and the spatial dimensions of access—potential spatial accessibility which is the extent to which people in a given area can reach health services and facilities (Guagliardo, 2004; Apparicio et al., 2017). The potential spatial access accounts for population demand in space and time for existing health services. Spatial access to health services is greatly influenced by factors such as travel time, travel distance and mobility infrastructure, and has been widely recognized to be complex, requiring adequate consideration of geography, socioeconomic and cultural factors (Penchansky and Thomas, 1981; Guagliardo, 2004; Humphreys and Smith, 2009). That is, for populations to receive healthcare, health facilities must be within reasonable distances for timely access.
Evaluating the disparities in spatial access to health services is critical to enhancing the adequate allocation of resources (Syed et al., 2013; Russell et al., 2019; Parker, 2021; Ujewe and van Staden, 2021; Jia et al., 2022), identification of underserved populations (Rosero-Bixby, 2004; Pan et al., 2016; Richard et al., 2016; Russell et al., 2019; Ashiagbor et al., 2020) and enhance efficiency and performance evaluation of the health system (Amer, 2007; Agbenyo et al., 2017; Ghorbanzadeh et al., 2020; Shao and Luo, 2022). While equity of health access is a challenge worldwide, the situation is much more severe in sub-Sahara Africa (SSA) due to inadequate health facilities (Dumedah et al., 2021; Ashu et al., 2022), low access (Anselmi et al., 2015; Ujewe and van Staden, 2021) and high costs of healthcare (Anselmi et al., 2015; Ujewe and van Staden, 2021; Ashu et al., 2022). Some work has been carried out to identify underserved areas in health access in SSA (Lawal and Anyiam, 2019; Ouko et al., 2019; Asare-Akuffo et al., 2020; Ashiagbor et al., 2020; Stewart et al., 2020; Dumedah et al., 2021). However, disparities in health access across geographical scales, in particular, inter-urban comparison of potential spatial access to health services in SSA, are limited in the literature. City-wide assessment of health access is important, but inter-city evaluations are also vital in accessing health resource allocation and service quality across cities. That is, the disparities in access to health services across cities are important to determine whether there are disparities in investment at certain locations, and to provide a consistent framework like a scorecard-type assessment to cities.
The linkages between human health and poverty are widely established (Whitehead et al., 2001; Berner, 2017; Chi et al., 2022; United Nations, U, 2023). Poverty has been described as deprivation of material essentials such as food, shelter, drinking water and clothing, and is associated with the lack of education, freedom and dignity (Berner, 2017; Yapa, 2017). Poverty can be characterized as absolute where it is often specified as those with income below a certain threshold, or relative in relation to lack of benefits from facilities and services available to others. The Human Poverty Index and Multidimensional Poverty Index (MPI) are examples of relative measure of poverty. Global MPI measures multiple deprivations in the areas of education, health and living standards, with the capability to reflect the incidence of poverty (i.e. proportion of poor persons) and its intensity (i.e. average number of deprivations that each poor person experiences) (Abbie Erler, 2012; Aryeetey, 2015; Berner, 2017; United Nations, U, 2023).
Relatedly, life expectancy across the suburbs of New Orleans has been shown to vary by a 25 year gap (Bambra, 2016; 2018). Literature has shown that human health can be linked to compositional factors such as individual-level demographic, behavioural and socioeconomic variables, and contextual factors in terms of the economic, social and physical environment (Cummins et al., 2007; Bambra, 2018). In addition, the notion of the relational approach emphasizes the interaction of people with the wider environment and the political economy approach which recognizes the role of the wider macro-political, economic and societal context (Bambra, 2016; 2018; 2022). That is, the disparities in health access across cities will directly relate to the interaction among key drivers of health including people, place, environment and political economy. Moreover, the evaluation of potential spatial access to health services in SSA is uniquely complicated by limitations in data, consistent standards and a mix of semi-informal services (Agbenyo et al., 2017; Asare-Akuffo et al., 2020; Ujewe and van Staden, 2021; Ashu et al., 2022).
Accordingly, this study examines disparities in the potential spatial access to health services in four selected urban areas in Ghana: Greater Accra, Kumasi area, Tamale area and Sekondi-Takoradi area. Assessment of spatial access to health services is generally limited and often based on a single urban area (Agbenyo et al., 2017; Asare-Akuffo et al., 2020; Ashiagbor et al., 2020; Dumedah et al., 2021). The study by Ashiagbor et al. (2020) found that about 81% of the population had access to primary healthcare in the Ashanti region of Ghana, while about 61% have access to secondary-level, and 14% to tertiary care. Notwithstanding the relatively high rates of accessibility, about 30% of the population has to travel far to access primary care. Low spatial access of about 2.34 physicians per 10 000 persons has been found in the Greater Kumasi of Ghana, where the health facilities were associated with detailed geographic coverage, low numbers of physicians and high population demand (Dumedah et al., 2021). Assessment of spatial access to three levels of health services in the Wa West District of Ghana showed that about 50% of communities have high accessibility to Community-Based Health Planning and Services (CHPS) compounds and health centres, whereas only 4% have high physical access to district hospitals (Agbenyo et al., 2017). Overall, these studies have shown that there is low spatial access to health services in Ghana, but they have employed different methodologies and are based on a single urban area.
The health system in Ghana is mainly divided into five tiers, comprising CHPS compound at the community level, health centres in sub-districts, district hospitals at the district level, regional hospitals at the regional level and specialized/teaching hospitals at the national level (Agbenyo et al., 2017). Recently, the national government of Ghana has embarked on Agenda111, which is aimed to undertake a major investment in healthcare infrastructure by constructing 88 hospitals in the districts without hospitals, six new regional hospitals in the six new regions and the rehabilitation of the Effia Nkwanta Hospital in Sekondi (Agenda111, 2022). Agenda111 is envisioned to ensure that Ghanaians in every district and region in the country have access to quality healthcare services. While this is a commendable objective, there are limited research works in Ghana that investigate the disparities and therefore the equity of spatial access to health services or facilities. Such national investigations are critically needed to evaluate the performance of existing health services and to inform health resource allocation for both current and future health facilities.
Material and methods
Study area
The study was carried out in four selected urban areas in Ghana: Greater Accra, Kumasi area, Tamale area and Sekondi-Takoradi area which are presented in Figure 1. These areas were chosen because they represent the leading top-four functional urban areas in Ghana by population. Each of the individual urban areas includes its corresponding regional capital, with Greater Accra having the city of Accra as the country’s capital. That is, each selected urban area includes an urban core and the surrounding outskirts which are less urbanized.
The study area shows the four selected urban areas in Ghana: Greater Accra, Kumasi area, Tamale area and Sekondi-Takoradi area
Greater Accra covers a land area of about 3245 km2 representing 1.4% of Ghana’s total land area with a population of about 5.5 million inhabitants, according to Ghana’s 2021 Population and Housing census. In Greater Accra, a total of 437 health units were obtained for the Greater Accra, comprising 324 Primary Health Care (PHC) units, 104 Secondary Health Care (SHC) units and 9 Tertiary Health Care (THC) units. The PHC comprises 247 clinics, 48 maternity homes, 23 health centres and 6 CHP. The SHC has 104 hospitals, whereas the THC is made up of 8 health training institutes and one teaching hospital.
The area referred to, in this study, as the Kumasi area is a combination of administrative areas comprising the Kumasi metropolis at the core and surrounded by several municipalities which are Asokore Mampong, Oforikrom, Old Tafo, Kwadaso, Suame, Asokwa, Atwima Nwabiagya North and South, Afigya Kwabre, Kwabre East, Atwima Kwanwoma, Bosomtwe and Ejisu. The chosen Kumasi area covers a land area of about 3044 km2 representing 1.3% of Ghana’s total land area and comprises 2.7 million inhabitants. In the Kumasi area, there is a total of 327 health units comprising 259 PHC units, 64 SHC units and 4 THC units. The PHC comprises 161 clinics, 64 maternity homes, 31 health centres and 3 CHP. The SHC has 64 hospitals, whereas the THC is made up of 3 health training institutes and one teaching hospital.
The area referred to, in this study, as the Tamale area is a combination of administrative areas comprising the Tamale metropolis and the municipality of Sagnerigu. The Tamale area covers a land area of about 1135 km2 representing 0.5% of Ghana’s total land area and comprises 716 455 inhabitants. In the Tamale area, there is a total of 29 health units comprising 22 PHC units, 4 SHC units and 3 THC units. The PHC comprises 5 clinics, 2 maternity homes, 11 health centres and 4 CHP. The SHC has 4 hospitals, whereas the THC is made up of 2 health training institutes and one teaching hospital.
The area referred to, in this study, as the Sekondi-Takoradi (or Takoradi) area is a combination of administrative areas comprising the Sekondi-Takoradi metropolis and the municipality of Shama. The Takoradi area covers a land area of about 260 km2 representing 0.12% of Ghana’s total land area and comprises 362 606 inhabitants. In the Takoradi area, there is a total of 60 health units comprising 52 PHC units, 7 SHC units and one THC unit. The PHC comprises 42 clinics, 5 maternity homes, 3 health centres and 2 CHP. The SHC has 7 hospitals, whereas the THC is made up of one health training institute.
Data sources
Data on health facilities were obtained from the Ghana Health Service through Ghana’s Medical and Dental Council’s permanent register for 2020. It included details on the number of physicians, facility type, type of services offered and ownership (private or government). Health facilities were categorized into PHC, SHC and THC. The PHC units comprise CHPS, Clinics, Polyclinics, maternity homes and health centres; the SHC units comprise District health directorates, Hospitals and District Hospitals and THC units were considered to include Teaching hospitals, training institutions, psychiatric hospitals and regional hospitals.
Population data were obtained from the Humanitarian Data Exchange (https://data.humdata.org), at about 100 m spatial resolution for 2020. The population density data set forms the basis to spatially quantify the demand for health services at any location in the selected urban areas. The high-resolution 100 m population data allow a spatially distributed quantification of the demand instead of a uniform assignment of demand to larger administrative areas. Economic status data based on Relative Wealth Index (RWI) estimated by Chi et al. (2022) at 2.4 km spatial resolution were obtained from the Humanitarian Data Exchange. The RWI estimates the relative standard of living within countries using anonymized connectivity data, satellite imagery and other non-traditional data sources. The RWI by Chi et al. (2022) is an asset-based Demographic and Health Survey (DHS) wealth index, which is a single continuous measure of relative household wealth with a mean value of zero and a standard deviation of one. The RWI has been observed to be suitable for informal communities where large portions of the population do not earn formal wages and measures of income are especially unreliable (Chi et al., 2022). As a composite measure of a household’s cumulative living standard, the RWI places individual households on a continuous scale of relative wealth allowing identification of economically disadvantaged/poor and advantaged households.
The road network data set was obtained from OpenStreetMap (OSM) but required preprocessing for its use in transportation analytics in this case routing. The preprocessing included the correct identification of nodes which are decision-making points in a transport network and representative topology of road segments. The QGIS software was used to facilitate the download, preprocessing of the road network data and subsequent derivation of the travel distances towards evaluating the potential spatial access to health services.
Indicative population demand in the selected cities
Typically, population growth leads to increasing demand for health services. How the population is distributed in space can directly impact the level of access to health services (Wang et al., 2013; Ingvardson and Nielsen, 2018). Each of the selected urban areas includes a diverse range of land cover types that distribute their populations in a certain way. To relate the proportion of land area covered by urban land, the population density is presented for the selected study cities in Figure 2. In Greater Accra, 35.5% of urban land is mostly concentrated on the south-western side, indicating the areas with the highest demand for health services. In the case of the Kumasi area, the population is highest at the central location and spreads outwards in all directions. The 32.7% of urban land in the Kumasi area is mostly concentrated at the central location, indicating the areas with the highest demand for health services.
The population density in the four selected urban areas in Ghana
The population distribution in the Tamale area shows that the population is highest at a small speckle at the urban core which depicts a near-rectangular shape. Generally, 6.4% of urban land in the Tamale area is concentrated at the urban core and the peri-urban area indicating the areas with the highest demand for health services. In the case of the Takoradi area, the population is highest at the urban core which is located on the south-western side. Surrounding the urban core are pockets of densely populated peri-urban areas which spread northward and eastward. The 30.1% of urban land in the Takoradi area is mostly concentrated at the south-western side, indicating the areas with the highest demand for health services. In general, all the selected urban areas and their distributions of the population are indicative of the demand areas for health services.
Estimation of potential spatial access to health services
Spatial access to health services has continued to receive intense investigation with considerations of several of its drivers. Among its drivers, it is widely recognized that the evaluation of spatial access to health services is complex, and that it requires a consideration of geography, socioeconomic and cultural factors (Guagliardo, 2004; Guagliardo et al., 2007; Apparicio et al., 2017; Masiano et al., 2019). Current research into spatial access to health services has shown two leading approaches: Three-Step Floating Catchment Area (3SFCA) developed by Wan et al. (2012), and Rational Agent Access Model (RAAM) which was developed by Saxon and Snow (2020).
3SFCA—the Three-Step Floating Catchment Area
The 3SFCA is a special case of the gravity model, which has improved on the family of Floating Catchment Areas (FCA) approaches. The 3SFCA accounts for the supply and demand aspects of healthcare access and has been widely applied in the literature as a robust option to estimate the spatial accessibility to health services (Delamater, 2013; McGrail and Humphreys, 2014; 2015; Neutens, 2015; Langford et al., 2016; Dumedah et al., 2021). The 3SFCA method has three major steps, which are described as follows.
Step 1: A threshold catchment area for each population demand location, i is defined for an area within a 30 min driving zone, which in turn is divided into three travel time zones 1–3. This first step searches all the supply locations within the catchment to assign a Gaussian weight to each supply location based on the travel time zone in which it lies. The assigned weights are then used to calculate a selection weight between each supply location and population demand location by using equation (1).
In equation (1), |${G_{ij}}$| is the selection weight between population demand location i and supply location j derived from a Gaussian function; di, k is the travel time between i and any supply location k within the catchment; Dr is the rth travel time zone (r = 1–3) within the catchment; |$W({t_{ij}})$| and |$W({t_{ik}})$| are the distance weights for j and k, respectively, defined using a Gaussian function.
Step 2: For each supply location (j), all population demand locations (k) within the threshold travel time zone (Dr) from location j are searched, where the weighted physician-to-population ratios Rj is estimated using equation (2).
In equation (2), Pk is the population of grid cell k falling within the catchment j (|${d_{kj}} \in {D_r}$|), Sj is the number of health facilities at location j, dkj is the travel time between k and j and Dr is the rth travel time zone (r = 1–3) within the catchment. Wr is the distance weight for the rth travel time zone estimated from the Gaussian function, which describes the distance decay of access to the health facility j.
Step 3: For each population location i, all health facilities locations (j) which are within the threshold 30 min travel time zone from location i are searched, where the sum of all the health service-to-population ratios Rj at these locations is determined using equation (3).
In equation (3), |$A_i^F$| represents the accessibility of the population at location i to health service, Rj is the health service-to-population ratio at health facility location j which falls within the catchment centred at population i (i.e. |${d_{kj}} \in {D_r}$|) and dij is the travel time between i and j locations. To account for the distance decay for different travel time zones, the weights estimated from the Gaussian functions as in Step 1 were used, where |${G_{ij}}$| is the selection weight between i and j, and Wr is the Gaussian weight for the rth travel time zone (r = 1–3) Dr. Additional information on the analytical process of 3SFCA can be found in the work by Wan et al. (2012).
RAAM—the Rational Agent Access Model
The RAAM approach optimizes the allocation of patients to providers by minimizing travel times and congestion at the provider. This is achieved by accounting for users’ response to congested supply locations, where congestion is the ratio of demand to supply, normalized by the area mean (Saxon and Snow, 2020; Saxon et al., 2022). The RAAM uses access score—a weighted sum of access components like distance to provider and relative importance of provider type, to assign the level of access to health services where a lower access score is better.
In the RAAM, the cost of service is considered as the sum of congestion and travel costs. The congestion cost accounts for the total demand relative to supply, and is normalized by a factor ρ. The travel cost is as determined before in the 3SFCA case, where tij is the travel cost between demand location i and supply location j, and normalized by a factor τ. The allocation of demand to supply is expressed as the solution to the optimization problem in equation (4).
From equation (4), Dij is the demand from origin i allocated to supply location j; Sj is the supply at the destination j; tij is the travel cost between i and j; τ sets the travel costs relative to congestion and ρ is taken as a reference inverse of service-to-user ratio. Through minimization of the optimization problem, the RAAM index for the allocation of demand i to the optimal location l is presented in equation (5). The RAAM index accounts for the sum over r reflecting all the other demand points allocated to supply point l. Additional information on the analytical process of RAAM can be found in the work by Saxon and Snow (2020).
To apply both methods (3SFCA and RAAM), the supply was determined as the location of each health facility and its associated number of physicians. The demand was estimated as the location of each community and its associated population. The travel costs between demand and supply locations were estimated as the travel distance along the shortest pathway on a road network. These shortest travel paths were estimated using Geographic Information Systems (GIS) in QGIS software. These inputs were used to estimate the spatial access using 3SFCA and RAAM methods.
Travel time estimation from Openrouteservice (ORS)
Aside from 3SFCA and RAAM methods, travel time to access health facilities was estimated using Openrouteservice (ORS). The travel time was based on the time spent travelling on the shortest travel path on a road network from origin (i.e. community location) to destination (health facility) using a car under typical traffic conditions. The coordinate values for origin and destination pairs were supplied as input into Openrouteservice (ORS) [https://openrouteservice.org/] to estimate the travel time to access health services. ORS is a free web service that uses publicly available geographic data from OpenStreetMap to provide travel directions, travel time and travel-related information. The estimated travel times were obtained in minutes where they were categorized into groups to determine the number of communities associated with each.
Results and discussion
Potential spatial accessibility by 3SFCA
The 3SFCA estimations of the potential spatial access to health services for the selected areas are presented in Figure 3 showing the spatial differentiation of access. In Greater Accra, the estimated spatial access has an average of 2.06 physicians per 10 000 persons and a standard deviation of 1.46. These estimates indicate relatively moderate spatial access in comparison to the other selected urban areas. Geographically, the highest access is found in the northern portion of Greater Accra which is sparsely populated. The lowest spatial access is found at the extreme eastern side and the north-west side of Greater Accra which are also sparsely populated. The south-western side which is associated with the highest population demand has a relatively moderate spatial access to health services, around the overall average.
The estimated potential spatial accessibility is based on the 3SFCA approach for the four cities in Ghana
In the Kumasi area, the estimated spatial access has an average of 2.24 physicians per 10 000 persons and a standard deviation of 1.11. These estimates indicate moderate spatial access in comparison to the other selected urban areas. The highest spatial access is found centrally at the urban core with high population demand and at large areas of the south-east which is associated with low population demand. The lowest spatial access is found mostly on the outskirts with low population demand mostly on the western side, south side and some spotted areas in the north. The rest of the urban core is associated with high levels of access; overall, areas with higher population demand have higher levels of spatial access.
In the Tamale area, the estimated spatial access has an average of 3.02 physicians per 10 000 persons and a standard deviation of 1.45. These estimates show relatively higher spatial access in comparison to the other selected urban areas. The highest spatial access to health services is found at the urban core, the peri-urban areas and some areas surrounding it, which together make up the highest population demand. The lowest spatial access is found predominately on the eastern side and some dotted sections on the south-west and north sides. Overall, enhanced spatial access is at the urban core and peri-urban area where it decreases uniformly outward in all directions. In other words, the level of spatial access almost decreases consistently away from the urban core.
In the Takoradi area, the estimated spatial access has an average of 1.78 physicians per 10 000 persons and a standard deviation of 0.29. These estimates show relatively lower spatial access in comparison to the other selected urban areas. The highest spatial access to health services is found at the urban core and the peri-urban areas which together make up the highest population demand. The lowest spatial access is found at the eastern side where there is low population demand. Overall, enhanced spatial access is at the urban core and peri-urban area where it decreases almost uniformly towards the outskirts in the north and eastern directions.
Potential spatial accessibility by RAAM
The RAAM estimations of the potential spatial access to health services for the selected areas are presented in Figure 4, showing the spatial differentiation of access. It is noted that the lower the access score, the better the level of access to health services. In Greater Accra, the estimated spatial access score has an average of 2.35 and a standard deviation of 1.14. These estimates show relatively moderate spatial access in comparison to the other selected urban areas. The highest spatial access is found at the urban core and some dotted areas on the north and east sides. The lowest spatial access is found at the extreme eastern side and the north-west side of Greater Accra which are sparsely populated. In general, the spatial pattern of access to health services indicates the highest access at the urban core and almost uniformly decreases towards the outskirts.
The estimated potential spatial accessibility is based on the RAAM approach for the four cities in Ghana
In the Kumasi area, the estimated spatial access has an average of 1.70 and a standard deviation of 0.80. These estimates indicate relatively higher spatial access in comparison to the other selected urban areas. The highest spatial access is found at the urban core, its northern outskirts and some areas on the south-east side. The lowest spatial access is found on the western side and some areas on the south side of Kumasi area which are sparsely populated. In general, the urban core associated with the highest population demand has different levels of spatial access varying from highest to moderate.
In the Tamale area, the estimated spatial access has an average of 2.54 and a standard deviation of 0.91. These estimates indicate relatively lower spatial access in comparison to the other selected urban areas. The highest spatial access is found at the urban core, its peri-urban area and some outskirts on the south side. The lowest spatial access is found on the eastern side and some areas on the western side which are mostly sparsely populated. In general, the urban core and the peri-urban areas which are associated with the highest population demand uniformly have the highest spatial access. The spatial pattern of access in the Tamale area indicates the highest access at the urban core and almost uniformly decreases towards the outskirts mostly in the east and west directions.
In the Takoradi area, the estimated spatial access has an average of 2.42 and a standard deviation of 1.11. These estimates indicate relatively lower spatial access in comparison to the other selected urban areas. The highest spatial access is found at the urban core and its peri-urban area, which is associated with the highest population demand. The lowest spatial access is found on the eastern side which is mostly sparsely populated. In general, the spatial pattern of access in the Takoradi area indicates the highest access at the urban core and almost uniformly decreases towards the outskirts to the north and east directions.
Comparison between 3SFCA and RAAM estimates
The estimations of potential spatial access to health services obtained from 3SFCA and RAAM indicate that there is a moderate agreement between the two approaches. This agreement is predicated on the consistency of the potential spatial access between the estimations from 3SFCA and RAAM. In general, both estimations showed consistency in the locations identified with higher or lower levels of spatial access. In addition, the relative spatial access in comparison between the selected urban areas is consistent across both methods. For example, both 3SFCA and RAAM estimations indicated that the Kumasi area has relatively higher spatial access in comparison to the other selected urban areas. Similarly, the Takoridi area was consistently identified in both 3SFCA and RAAM estimations as having relatively lower spatial access in comparison to the other selected urban areas.
Nevertheless, there are notable differences between the 3SFCA and RAAM estimations of the potential spatial access to health services for the four selected urban areas. To evaluate these comparisons, the linear relationship between the 3SFCA and RAAM estimations is assessed using a one-to-one plot and the coefficient of determination (R-square or R2) presented in Figure 5. The R2 values for the selected urban areas are 0.049 for Greater Accra, 0.145 for the Kumasi area, 0.578 for the Tamale area and 0.744 for the Takoradi area. These results indicate that the linear agreement between 3SFCA and RAAM estimations is strong for Takoradi, moderate for Tamale and weak for Greater Accra and Kumasi. The estimates for Takoradi and Tamale are indicative of their relatively smaller population densities in comparison to those of Greater Accra and Kumasi areas. That is, a stronger agreement between the 3SFCA and RAAM estimations is strongly associated with smaller population densities. It is noted that smaller population density areas are relatively more likely to be associated with less congestion at the health service providers in comparison to the higher population density areas.
Comparison of potential spatial access estimates between 3SFCA and RAAM for the four cities in Ghana
However, the estimates for Greater Accra and Kumasi areas are indicative of their relatively higher population densities. The weak linear relationship between the 3SFCA and RAAM estimations is strongly associated with higher population densities. Congestion at the health service provider is most likely related to higher population density areas, where it is specifically accounted for in the RAAM estimation compared with the 3SFCA approach. That is, in cases where congestion at the health service provider is considerable, the RAAM approach is preferable. In addition, the one-to-one plots for Greater Accra and Kumasi areas show that the differences between 3SFCA and RAAM estimations are largest for cases where the RAAM shows a higher spatial access whereas the 3SFCA produces a smaller spatial access. That is, for smaller spatial access, there is consistent strong agreement between the 3SFCA and RAAM estimates whereas there is weak agreement at higher spatial access.
Travel time to health services by the proportion of communities
A further comparison between the four selected urban areas was carried out to evaluate the travel time (from ORS) to accessing health services by the proportion of communities in Figure 6. The information in Figure 6 shows that for 10 minutes and 20 minutes of travel time, Tamale and Takoradi are the leading cities with larger proportions of communities to reach health services followed by Kumasi and then Greater Accra. This pattern is reversed for 45 minutes or greater of travel time where Greater Accra followed by Kumasi have larger proportions of communities to reach health services, than Takoradi and Tamale. For shorter travel times to health services Tamale and Takoradi have a larger proportion of communities whereas for longer travel times they are associated with a smaller proportion of communities. That is, a greater proportion of the communities in Tamale and Takoradi can reach their health services within shorter travel times. This pattern of travel time to reach health services is reversed for the Kumasi area and Greater Accra where a greater proportion of their communities can only reach their health services at longer travel times. Across all the travel times to health services, the decreasing order of performance is the Tamale area, Takoradi area, Kumasi area and Greater Accra.
Comparison of travel time in minutes from communities to health facilities for the four selected cities in Ghana
In relation to the decreasing order of performance for travel time by community, it is worth noting the relative differences in population densities for the four selected urban areas. High population densities are often associated with higher road traffic congestion where even shorter travel distances require longer travel times. Greater Accra and Kumasi areas are associated with higher population densities and are known for their higher road traffic congestion, which is quite severe for Greater Accra. That is, the poor performance of travel time by community to health services for Greater Accra and Kumasi areas is greatly attributable to their higher levels of road traffic congestion. This emphasizes the relevance of mobility infrastructure in accessing health services (Dumedah et al., 2022).
Linkages to settlement area, population and economic status
Furthermore, a comparison between the four selected urban areas was carried out by simultaneously evaluating the following variables: the land area covered, the population served and their economic status based on RWI who are within specific travel distance categories to health services. The travel distance categories based on a service area estimation for health facilities are presented in Figure 7 for the four selected urban areas. Travel distance was used instead of travel time as the former is fairly objective and not typically variable and subject to other variables. For any given travel distance, the~travel time can be estimated based on the travel mode and road traffic conditions. Travel time is dynamic and has been known to be influenced by several factors including travel mode, personal mobility, travel costs, transport availability and scheduling, natural barriers, age, gender and other socioeconomic, cultural and organizational variables (Kaur Khakh et al., 2019; Dumedah et al., 2021).
Travel distances to health services are estimated at consistent intervals for the four cities in Ghana
The eight travel distance categories in kilometres are 0–0.5; 0.5–1.0; 1–2; 2–3; 3–5; 5–7; 7–10 and greater than 10 (or ≫10). The information in Figure 7 indicates the spatial distribution of travel distance for each selected urban area. Typically, travel distances are small for areas close to the locations of health facilities and larger for locations further away from health facilities. The eight travel distance categories were used to extract their corresponding land area covered, the population served and their corresponding RWI. This information is presented in Figure 8, for each of the four urban areas showing the travel distances on the horizontal axis, the percentages of the land area and the population on the primary vertical axis and the RWI on the secondary vertical axis. As noted, RWI values can range from negative to positive values. Among the four variables: travel distance, land area, population served and RWI their cross-correlation coefficient (r) values are presented for each urban area in Table 1. The information from Figure 8 and Table 1 is summarized for the four selected urban areas in the following.
Comparison of travel distances (horizontal axis) to health services against percentages of population served and land area (primary vertical axis), and relative wealth index (secondary vertical axis) for the four cities in Ghana
Estimated correlation coefficient (r) values across travel distance, land area, population served and relative wealth index for each selected urban area
| Urban areas | Variable | Travel distance | Land area | Population | RWI |
|---|---|---|---|---|---|
| Greater Accra | Travel distance | 1.00 | 0.83 | −0.78 | −0.96 |
| Land area | 1.00 | −0.43 | −0.78 | ||
| Population | 1.00 | 0.83 | |||
| RWI | 1.00 | ||||
| Kumasi area | Travel distance | 1.00 | 0.60 | 0.15 | −0.61 |
| Land area | 1.00 | 0.85 | −0.95 | ||
| Population | 1.00 | −0.74 | |||
| RWI | 1.00 | ||||
| Tamale area | Travel Distance | 1.00 | 0.60 | 0.52 | −0.89 |
| Land area | 1.00 | 0.95 | −0.73 | ||
| Population | 1.00 | −0.64 | |||
| RWI | 1.00 | ||||
| Takoradi area | Travel Distance | 1.00 | 0.62 | −0.02 | −0.95 |
| Land area | 1.00 | 0.49 | −0.72 | ||
| Population | 1.00 | 0.04 | |||
| RWI | 1.00 |
| Urban areas | Variable | Travel distance | Land area | Population | RWI |
|---|---|---|---|---|---|
| Greater Accra | Travel distance | 1.00 | 0.83 | −0.78 | −0.96 |
| Land area | 1.00 | −0.43 | −0.78 | ||
| Population | 1.00 | 0.83 | |||
| RWI | 1.00 | ||||
| Kumasi area | Travel distance | 1.00 | 0.60 | 0.15 | −0.61 |
| Land area | 1.00 | 0.85 | −0.95 | ||
| Population | 1.00 | −0.74 | |||
| RWI | 1.00 | ||||
| Tamale area | Travel Distance | 1.00 | 0.60 | 0.52 | −0.89 |
| Land area | 1.00 | 0.95 | −0.73 | ||
| Population | 1.00 | −0.64 | |||
| RWI | 1.00 | ||||
| Takoradi area | Travel Distance | 1.00 | 0.62 | −0.02 | −0.95 |
| Land area | 1.00 | 0.49 | −0.72 | ||
| Population | 1.00 | 0.04 | |||
| RWI | 1.00 |
Estimated correlation coefficient (r) values across travel distance, land area, population served and relative wealth index for each selected urban area
| Urban areas | Variable | Travel distance | Land area | Population | RWI |
|---|---|---|---|---|---|
| Greater Accra | Travel distance | 1.00 | 0.83 | −0.78 | −0.96 |
| Land area | 1.00 | −0.43 | −0.78 | ||
| Population | 1.00 | 0.83 | |||
| RWI | 1.00 | ||||
| Kumasi area | Travel distance | 1.00 | 0.60 | 0.15 | −0.61 |
| Land area | 1.00 | 0.85 | −0.95 | ||
| Population | 1.00 | −0.74 | |||
| RWI | 1.00 | ||||
| Tamale area | Travel Distance | 1.00 | 0.60 | 0.52 | −0.89 |
| Land area | 1.00 | 0.95 | −0.73 | ||
| Population | 1.00 | −0.64 | |||
| RWI | 1.00 | ||||
| Takoradi area | Travel Distance | 1.00 | 0.62 | −0.02 | −0.95 |
| Land area | 1.00 | 0.49 | −0.72 | ||
| Population | 1.00 | 0.04 | |||
| RWI | 1.00 |
| Urban areas | Variable | Travel distance | Land area | Population | RWI |
|---|---|---|---|---|---|
| Greater Accra | Travel distance | 1.00 | 0.83 | −0.78 | −0.96 |
| Land area | 1.00 | −0.43 | −0.78 | ||
| Population | 1.00 | 0.83 | |||
| RWI | 1.00 | ||||
| Kumasi area | Travel distance | 1.00 | 0.60 | 0.15 | −0.61 |
| Land area | 1.00 | 0.85 | −0.95 | ||
| Population | 1.00 | −0.74 | |||
| RWI | 1.00 | ||||
| Tamale area | Travel Distance | 1.00 | 0.60 | 0.52 | −0.89 |
| Land area | 1.00 | 0.95 | −0.73 | ||
| Population | 1.00 | −0.64 | |||
| RWI | 1.00 | ||||
| Takoradi area | Travel Distance | 1.00 | 0.62 | −0.02 | −0.95 |
| Land area | 1.00 | 0.49 | −0.72 | ||
| Population | 1.00 | 0.04 | |||
| RWI | 1.00 |
In Greater Accra, with increasing travel distance to health services the land area also consistently increased. The population served remained almost constant from 0–2 km, and then consistently decreased from 3–8 to 7–10. The RWI has a direct inverse relationship to the travel distance, where an increase in the latter shows a decrease in the former. Notably, RWI values in Greater Accra are positively higher in comparison to the other urban areas, which indicates a higher economic status. These relationships are supported by r values where travel distance is strongly related to land area (0.83), and negatively to both population (−0.78) and RWI (−0.96). Population and land area were found to be moderately and negatively related. In addition, RWI is found to be strongly predicated on travel distance, land area (−0.78) and population (0.83).
In the case of the Kumasi area, with increasing travel distance to health services the land area also consistently increased. But the land area corresponding to ≫10 km is considerably higher in comparison to the rest of the travel distances. On the contrary, the population served consistently decreased with increasing travel distance until ≫10 km. As was the case for land area, the population corresponding to ≫10 km is considerably higher in comparison to other travel distances. The RWI has an overall inverse relationship to the travel distance, where an increase in the latter shows a decrease in the former with few exceptions. Particularly, RWI values in the Kumasi area are mostly close to zero and negatively higher at the ≫10 km travel distance; these values indicate a relatively lower economic status. The linear relationships among the variables are supported by r values where the travel distance is moderately related to land area (0.60), weak for population (0.15) and moderately negative for RWI (−0.61). Land area was found to be strongly correlated with population (0.85) and inversely with RWI (−0.95). In addition, RWI is strongly and inversely related to travel distance, land area and population (−0.74).
The information on the Tamale area indicates that land areas consistently increase with increasing travel distance to health services. But the land area corresponding to ≫10 km is considerably higher in comparison to the rest of the travel distances. The relationship is relatively mixed between travel distance and population where the population increased with increasing travel distance from 0–0.5 to 1–2 km and then decreased with increasing travel distance from 2–3 to 7–10 km. As was the case for land area, the population corresponding to ≫10 km is considerably higher in comparison to other travel distances. The RWI has an overall inverse relationship to the travel distance, where an increase in the latter corresponds to a decrease in the former with an exception at 2–3 travel distance. Notably, RWI values in the Tamale area are mostly negative and higher indicating the lowest economic status in comparison to the other urban areas. The linear relationships among the variables are supported by r values where travel distance is moderately related to land area (0.60), population (0.52) and strongly negative for RWI (−0.89). Land area was found to be strongly correlated with population (0.95) and inversely with RWI (−0.73). Consistently, RWI is strongly and inversely related to travel distance, land area and population (−0.64).
In the case of the Takoradi area, land areas consistently increase with increasing travel distance to health services. But the land area corresponding to ≫10 km is considerably higher in comparison to the rest of the travel distances. The relationship is mixed between travel distance and population where the population increased with increasing travel distance from 0–0.5 to 1–2 km and then decreased with increasing travel distance from 3–5 to 7–10 km. The RWI has a direct inverse relationship to the travel distance, where the RWI decreases with increasing travel distance to health services. Notably, RWI values in the Takoradi area are positively higher in comparison to the other urban areas, which indicates a higher economic status. These relationships are supported by r values where the travel distance is moderately related to land area (0.62), weak for population (−0.02) and strongly negative for RWI (−0.95). Population and land area were found to be moderately and positively related (0.49). In addition, RWI is inversely and strongly predicated on travel distance and land area (−0.72), but weak for population (0.04).
Summary of findings
The comparison of the potential spatial access to health services for the four selected urban areas illustrates disparities at multiple levels including (1) the number of physicians per population by the 3SFCA approach; (2) access score based on the RAAM approach; (3) travel time to health services and (4) concurrent evaluation of the linkages between travel distance, settlement area, population and economic status. Based on the 3SFCA approach, the highest overall spatial access by the number of physicians per 10 000 persons was found to be the Tamale area (3.02) followed by the Kumasi area (2.24), Greater Accra (2.06), and Takoradi area (1.78) being the lowest. The specific ratios are 1:4854 for Greater Accra, 1:4464 for the Kumasi area, 1:3311 for the Tamale area and 1:5618 for the Takoradi area. That is, by considering the supply of health services based on the number of facilities and physicians, demand based on population and the travel costs based on travel distance on a roadway, the Tamale area is considered the highest performing, and the Takoradi area is the lowest. According to the Ghana Ministry of Health (2022), the national physician-to-patient ratio in 2020 is 1:6355, whereas their estimates for the regions which host the selected urban areas are 1:2619 for Greater Accra, 1:6007 for Ashanti, 1:8945 for Northern and 1:17 577 for Western. The WHO standard is 1:1000. The 3SFCA estimates show that all selected urban areas have better access compared with the national average and their respective regions. Greater Accra which has a matching area differs by about 2235 persons.
However, when congestion at the health service provider was specifically accounted for in addition to supply, demand and travel costs factors, the RAAM approach showed that the decreasing order of performance is Kumasi area (1.7), Greater Accra (2.35), Takoradi area (2.42) and Tamale area (2.54) based on the access score. When 3SFCA and RAAM are considered independent measures of potential spatial access, the overall decreasing order of performance is the Kumasi area, followed by Greater Accra and Tamale areas having equivalent performance, and the Takoradi area being the lowest. On account of travel time to health services by the proportion of communities, the decreasing order of performance is the Tamale area, Takoradi area, Kumasi area and Greater Accra. Primarily, this finding is a direct response to the level of road traffic congestion in the various urban areas with Greater Accra being the most congested city.
In addition, the findings based on linkages between travel distance to health services, settlement area, population and economic status illustrate further disparities in health access. Consistently, shorter travel distances to health services were found at the urban core where population demand is typically highest. Consistently across all four selected urban areas, the settlement area increases with increasing travel distance to health services supported by an overall high r-value of about 0.66. On the contrary, the relationship between population and travel distance to health services is mixed across all four selected urban areas. Similarly, the relationship between settlement area and population is mixed with a strong positive correlation in most cases but a negative correlation in the case of Greater Accra.
Economic status based on RWI was found to exhibit a consistent relationship to travel distance to health services and settlement areas across the four selected urban areas. Consistently across all four selected urban areas, RWI was found to decrease with increasing travel distance to health services supported by an overall r value of −0.85. Clearly, this implies that the most vulnerable populations based on economic status typically travel longer distances to access health services. That is, the disadvantage of longer travel distances to health services faced by vulnerable populations defined by lower RWI is consistent across all the selected urban areas. Also, RWI was found to be consistently inversely correlated to settlement area with an overall r value of −0.80. The linear relationship between RWI and population is mixed mostly strongly related either positively or negatively across the selected urban areas.
Conclusions and implication of findings
This study investigated disparities in the potential spatial access to health services for four selected urban areas in Ghana based on a variety of criteria. These criteria included: (1) the number of physicians per population by 3SFCA approach; (2) access score based on RAAM approach; (3) travel time to health services and (4) concurrent evaluation of the linkages between travel distance, settlement area, population and economic status. Each criterion highlighted unique information providing greater insight into the disparities in healthcare access within and across the selected urban areas. The overall spatial access to health services is low across all selected urban areas varying between 3.02 for the Tamale area and 1.78 physicians per 10 000 persons for the Takoradi area. To satisfy the WHO standard of one (1) physician to 1000 persons, the current number of physicians needs to be increased by five (5) times in Greater Accra, five (5) times in the Kumasi area, four (4) times in Tamale area and six (6) times in Takoradi area. The access score based on RAAM indicates the highest performing urban area as 1.7 for Kumasi, and 2.54 for Tamale being the least performing. Overall the findings show that the potential spatial access to health services is low and not equitable across and within the selected urban areas. In each urban area, the economically disadvantaged populations defined based on RWI were found to endure longer travel distances to access health services; this pattern is consistent across all the selected urban areas.
However, urban areas associated with higher economic status (i.e. RWI) are not necessarily advantaged in their access to health services. Based on RWI, Greater Accra and Takoradi areas are associated with relatively higher economic status but their corresponding spatial access to health services is not superior in comparison to the other urban areas. That is, the disparities in spatial access to health services between the urban areas are not attributable to differences in economic status but driven by other factors. While economically advantaged populations within each urban area were favoured in their spatial access to health services, this does not directly translate between the selected urban areas. That is, spatial access to health services is equitable at the inter-urban level, but not equitable at the intra-urban level. Each urban area has embedded disadvantaged populations whose spatial access to health services needs to be enhanced instead of inter-city improvement. It is acknowledged that urban areas with lower overall spatial access (e.g. Takoradi area) would need relatively higher health investment to level up to other higher-performing urban areas (e.g. Kumasi area).
Moreover, the range of findings shows the vital role of transport and mobility and their related environmental factors in enhancing spatial access to health services. For example, the high level of road traffic congestion associated with Greater Accra greatly weakens its spatial access in comparison to those in the Tamale area which has relatively lower road traffic congestion. This finding is in agreement with those from Bambra (2018; 2022) and Cummins et al. (2007) who acknowledged the role of the wider macro-political, economic and societal context. In addition, the overall low spatial access to health services in all the urban areas is partly attributable to a low number of physicians within the context of high population demand. The health facilities in all the selected urban areas have reasonable travel distances but are inadequately resourced with enough physicians and related resources. Increasing the physician numbers at spatially targeted existing facilities which already have relatively high proximity will considerably enhance the spatial access to health services.
The implication of economically disadvantaged populations travelling longer distances to access health services is far-reaching in all the selected urban areas. This finding indicates that potential spatial access to health services is not equitable within each urban area, and raises a question of the criteria for siting health facilities and healthcare provision in Ghana. For example, economically disadvantaged populations may be unable to afford housing near health facilities due to higher costs of living in those communities. Economic status and access to health services are central to meeting Sustainable Development Goals (SDGs, e.g. goals 3 and 10), with this study showing that inequities are embedded within the selected urban areas, and that economically poor populations are also disadvantaged in their physical access to healthcare. The extent to which economically poor populations are disadvantaged in their physical access to healthcare needs to be critically considered in the siting and resourcing of health facilities and healthcare provision.
However, there are remaining questions that need further investigation. This study employed potential spatial access to measure the extent to which populations are able to physically reach health services. Future studies should include realized or actual access to health services to account for all barriers (e.g. transport costs, traffic congestion, etc.) that need to be overcome to receive or actualize health services. It is widely acknowledged that there are differences between potential and realized access to services (Guagliardo, 2004; Lättman et al., 2016; 2018), and that addressing these disparities would enrich the findings of future studies. In addition, the economic status of populations was measured with RWI at a relatively larger spatial resolution; thus, collecting primary data to disaggregate the RWI data to finer spatial resolution would provide further insight into its linkages to healthcare access. Disadvantaged populations were measured economically based on RWI; social variables such as ethnicity, gender and other economic variables should be accounted for in further studies to expand the implications of unequal access to health services.
Authorship contributions
All persons who meet authorship criteria are listed as authors, and all authors certify that they have participated sufficiently in the work to take public responsibility for the content, including participation in the concept, design, analysis, writing or revision of the manuscript. Furthermore, each author certifies that this material or similar material has not been and will not be submitted to or published in any other publication before its appearance in Health Policy and Planning.
The following presents the specific contributions made by each author.
| Contribution Type | Contributing Authors |
|---|---|
| Conceptualization | Gift Dumedah, Samuel Adu-Prah, Sinead English |
| Methodology | Samuel Adu-Prah, Gift Dumedah, Seidu Iddrisu |
| Formal analysis | G. Dumedah, Seidu Iddrisu, Christabel Asare |
| Investigation | Seidu Iddrisu, Gift Dumedah, Sinead English |
| Resources | Christabel Asare, Samuel Adu-Prah |
| Data Curation | Seidu Iddrisu, Christabel Asare |
| Writing - Original Draft | Gift Dumedah, Samuel Adu-Prah, Sinead English, Seidu Iddrisu |
| Writing - Review & Editing | Gift Dumedah, Christabel Asare, Samuel Adu-Prah |
| Supervision | Gift Dumedah, Samuel Adu-Prah, Sinead English |
| Contribution Type | Contributing Authors |
|---|---|
| Conceptualization | Gift Dumedah, Samuel Adu-Prah, Sinead English |
| Methodology | Samuel Adu-Prah, Gift Dumedah, Seidu Iddrisu |
| Formal analysis | G. Dumedah, Seidu Iddrisu, Christabel Asare |
| Investigation | Seidu Iddrisu, Gift Dumedah, Sinead English |
| Resources | Christabel Asare, Samuel Adu-Prah |
| Data Curation | Seidu Iddrisu, Christabel Asare |
| Writing - Original Draft | Gift Dumedah, Samuel Adu-Prah, Sinead English, Seidu Iddrisu |
| Writing - Review & Editing | Gift Dumedah, Christabel Asare, Samuel Adu-Prah |
| Supervision | Gift Dumedah, Samuel Adu-Prah, Sinead English |
| Contribution Type | Contributing Authors |
|---|---|
| Conceptualization | Gift Dumedah, Samuel Adu-Prah, Sinead English |
| Methodology | Samuel Adu-Prah, Gift Dumedah, Seidu Iddrisu |
| Formal analysis | G. Dumedah, Seidu Iddrisu, Christabel Asare |
| Investigation | Seidu Iddrisu, Gift Dumedah, Sinead English |
| Resources | Christabel Asare, Samuel Adu-Prah |
| Data Curation | Seidu Iddrisu, Christabel Asare |
| Writing - Original Draft | Gift Dumedah, Samuel Adu-Prah, Sinead English, Seidu Iddrisu |
| Writing - Review & Editing | Gift Dumedah, Christabel Asare, Samuel Adu-Prah |
| Supervision | Gift Dumedah, Samuel Adu-Prah, Sinead English |
| Contribution Type | Contributing Authors |
|---|---|
| Conceptualization | Gift Dumedah, Samuel Adu-Prah, Sinead English |
| Methodology | Samuel Adu-Prah, Gift Dumedah, Seidu Iddrisu |
| Formal analysis | G. Dumedah, Seidu Iddrisu, Christabel Asare |
| Investigation | Seidu Iddrisu, Gift Dumedah, Sinead English |
| Resources | Christabel Asare, Samuel Adu-Prah |
| Data Curation | Seidu Iddrisu, Christabel Asare |
| Writing - Original Draft | Gift Dumedah, Samuel Adu-Prah, Sinead English, Seidu Iddrisu |
| Writing - Review & Editing | Gift Dumedah, Christabel Asare, Samuel Adu-Prah |
| Supervision | Gift Dumedah, Samuel Adu-Prah, Sinead English |
Reflexivity Statement
The authors include two females and three males, and span multiple levels of seniority especially the inclusion of two young and early-career researchers. Three of authors are local experts raised and living in Ghana, while the remaining two reside outside Ghana but conduct extensive research in the Global South. Among the two authors living outside Ghana, one was raised and has lived in Ghana in the past. All authors have experience conducting equity of health access research in the Global South.
Ethical approval
Ethical approval for this type of study is not required by our institute.
Conflict of interest statement
None declared.
