Multilevel geospatial analysis of factors associated with unskilled birth attendance in Ghana.

BACKGROUND: Globally, about 810 women die every day due to pregnancy and its related complications. Although the death of women during pregnancy or childbirth has declined from 342 deaths to 211 deaths per 100,000 live births between 2000 and 2017, maternal mortality is still higher, particularly in sub-Saharan Africa and South Asia, where 86% of all deaths occur.
METHODS: A secondary analysis was carried out using the 2014 Ghana Demographic and Health Survey. A sample total of 4,290 women who had a live birth in the 5 years preceding the survey was included in the analysis. GIS software was used to explore the spatial distribution of unskilled birth attendance in Ghana. The Geographic Weighted Regression (GWR) was employed to model the spatial relationship of some predictor of unskilled birth attendance. Moreover, a multilevel binary logistic regression model was fitted to identify factors associated with unskilled birth attendance.
RESULTS: In this study, unskilled birth attendance had spatial variations across the country. The hotspot, cluster and outlier analysis identified the concerned districts in the north-eastern part of Ghana. The GWR analysis identified different predictors of unskilled birth attendance across districts of Ghana. In the multilevel analysis, mothers with no education, no health insurance coverage, and mothers from households with lower wealth status had higher odds of unskilled birth attendance. Being multi and grand multiparous, perception of distance from the health facility as not a big problem, urban residence, women residing in communities with medium and higher poverty level had lower odds of unskilled birth attendance.
CONCLUSION: Unskilled birth attendance had spatial variations across the country. Areas with high levels of unskilled birth attendance had mothers who had no formal education, not health insured, mothers from poor households and communities, primiparous women, mothers from remote and border districts could get special attention in terms of allocation of resources including skilled human power, and improved access to health facilities.

Background

Globally, about 810 women die every day due to pregnancy and its related complications [1]. Although the death of women during pregnancy or childbirth has declined from 342 deaths to 211 deaths per 100,000 livebirths between 2000 and 2017, maternal mortality is still higher, particularly in sub-Saharan Africa and South Asia, where 86% of all deaths occur. The lifetime risk of obstetric causes of maternal deaths in developing countries is 33 times higher than deaths in developed countries. Evidence depicts that 94 per cent of maternal and child deaths occur in low-and-middle-income countries, including Ghana [1]. However, Ghana has chalked major successes in its quest to reduce maternal and neonatal mortality. For instance, the maternal mortality ratio (MMR) has declined from 760 deaths per 100,000 live births in 1990 to 350/100,000 live births in 2010 and 319/100,000 livebirths in 2015 [2]. Similarly, neonatal deaths (NMR) have steadily declined from 59 deaths per 1,000 live births in 1970 to 23 per 1,000 live births in 2019. These reductions are also due to the substantial increase in institutional healthcare deliveries from 46% in 2003 to 78% in 2018 [2]. However, births assisted by skilled personnel in Ghana are below 70%, compared to the global 81 per cent coverage [3].

Skilled birth attendance (SBA) is defined as births assisted by healthcare trained professionals such as doctors, midwives, and nurses to manage uncomplicated pregnancies, childbirth, the postpartum period (within 24 hours), and new-born health [4]. SBA has been identified as one of the surest ways of extenuating the risks of maternal deaths during pregnancy [3]. Therefore, it is recommended that SBA be relied upon throughout the continuum of care as it has clinical rationales–prevention of direct and indirect obstetric complications including haemorrhage, eclampsia, obstetric fistula, obstructed labour, unsafe abortions [5, 6], and HIV and malaria, respectively. Further, neonatal deaths are lowest in countries with the highest coverage of skilled birth attendance [7], and it is found that SBA reduces stillbirth by 23% [8].

In Ghana, appreciable progress has been made to promote SBA. For instance, the National Health Insurance Scheme (NHIS) covers SBA, and pregnant women are further exempt from paying NHIS premiums [9]. Considerable improvements have also been achieved in Ghana’s Universal access to healthcare as Community-based Health Planning Services (CHPS) have been extended to reach hitherto remote areas in the country. Therefore, it is not surprising that SBA has increased from 44% in 1993 to 68% in 2011 [10]. However, MMR of 319/100,000 and NMR of 30 deaths per 1,000 live births are still unexpectedly high, and a major contributor of MMR and NMR is unskilled birth attendance, defined as births assisted by a traditional birth attendant (TBA), traditional health volunteer, community/village health volunteer, neighbours/ friends/relatives, and other untrained individuals [11]. Such women are usually assisted by either traditional birth attendants or relatives and friends [12], and this has dire consequences for both the mother and her unborn child or children. Unskilled birth attendance is undoubtedly a major drawback to achieving SDG 3 target 3.1, aiming to reduce the global maternal mortality ratio to less than 70 per 100,000 [3].

Several factors have been identified in Ghana to influence women’s decision to utilize unskilled birth attendance. These factors include low socioeconomic status, no national health insurance (NHIS) coverage, residing in rural areas, higher-order parity, ethnicity, traditional religion, previous contact with the health system, and low levels of education [13-15]. Budu [13] found that as parity increased, home delivery was more likely to occur, and women with no NHIS were more likely to deliver at home. In other reports, Apanga and Awoonor-Williams [16] found that physical inaccessibility to health care, socio-cultural and economic factors, and health care system challenges were the major causes of low levels of skilled birth attendance in Ghana. In his equity analysis also, Zere et al. [17] showed that many maternal health services, including access to SBA and caesarean section services, healthcare facilities, use of modern contraceptives, and intermittent treatment for malaria, are not equitably distributed as women in wealthier households are more likely to utilize Maternal and Child Health (MCH) services, compared to their counterparts from poorer households. Evidence of North-South variations in SBA has been reported as well.

Skilled birth attendance (SBA) uptake is higher in southern Ghana [13, 18, 19] compared to Northern parts of Ghana [20]. Such variations and uneven use of services blurs national progress and continuously perpetuates inequalities in health. Cues must be taken because Ghana failed to achieve some Millennium Development Goals (MDGs) targets, including those related to maternal health. We, therefore, analyse the current Demographic and Health Survey data of Ghana to determine individual, community, and regional levels factors predicting unskilled birth attendance in Ghana. We further contribute to the maternal and child health literature in Ghana by highlighting hotspots for unskilled birth attendance and the factors associated with such observations. Previous geospatial studies have focused on inadequate variables [21] or focused on skilled birth attendance [22]. Understanding the predictors of unskilled birth attendance and its geospatial coverage in Ghana is of high relevance to both local and public health understanding of unskilled birth attendance and highlight areas where national interventions will yield cost-effective results.

Methods

Source of data

The Ghana Demographic and Health Survey (GDHS) used a standard Demographic and Health Survey (DHS) model questionnaire developed by the Measure DHS programme. This study used the most recent DHS data and a cross-sectional study design. The DHS are national surveys carried out every five years in low-and middle-income countries globally. The surveys concentrate on maternal and child health, physical activity, sexually transmitted infections, fertility, health insurance, tobacco use, and alcohol consumption. They provide data to monitor the demographic and health profiles of the respective countries. For the study, women with birth history who had given birth up to five years before the survey were included. Only the last birth of the women aged 15–49 years preceding the survey was included in the study. A sample of 4,290 women with complete data required for our analysis participated in this study. Permission to use the data set was given by the MEASURE DHS following the assessment of our concept note. This study is a secondary analysis of the de-identified 2014 Ghana Demographic and Health Survey (GDHS), a publicly available dataset. Therefore, ethics approval or consent to participate is not applicable. The datasets are freely available to the public at www.measuredhs.com.

Description of variables

Outcome variables

Outcome variable. The primary outcome variable was unskilled birth attendance. The outcome variable was derived from the response to the question "who assisted with the delivery?" Responses were categorized under Health Personnel and Other Person. Health personnel included doctor, nurse, nurse/midwife, auxiliary midwife, and other people consisting of a traditional birth attendant, traditional health volunteer, community/village health volunteer, neighbours/ friends/relatives, etc. For this study, unskilled birth attendance referred to births assisted by a traditional birth attendant, traditional health volunteer, community/village health volunteer, neighbours/ friends/relatives, other [11].

Explanatory variables. Eleven explanatory variables were used. These were grouped into individual and community level variables. The individual characteristics consist of level of education, age, parity, health insurance coverage, wealth status, distance to a health facility, and media exposure. The community-level characteristics comprised the type of residence, community socioeconomic status, community literacy, and region of residence.

Statistical analysis

We employed both descriptive and inferential analytical approaches. First, we computed the proportion of women who utilized the service of unskilled birth attendants during delivery. This ensued with bivariate analysis between individual characteristics (level of education, age, parity, health insurance coverage, wealth status, distance to a health facility, and media exposure), community characteristics (Type of residence, Community socioeconomic status, Community literacy, Region of residence) and utilization of unskilled birth attendants (see Table 1). Following the hierarchical nature of the data set, the Multilevel Logistic Regression Model (MLRM) was employed. This comprises fixed effects and random effects [23]. The fixed effects of the model were gauged with binary logistic regression, which resulted in odds ratios (ORs) and adjusted odds ratios (aORs) (see Table 2). Model 1 was an empty table, where model 2 looked at the relationship between the individual variable and the outcome variable. Model 3 looked at the relationship between community variables and the outcome variable. Model 4 was the complete model that looked at the relationship with both the individual and community variables and the outcome variable. The random effects, on the other hand, were assessed with Intra-Cluster Correlation (ICC) [23] (Table 3). The sample weight (v005/1,000,000) was applied in all the analyses to control for over and under-sampling. All the analyses were carried out using STATA version 14.

Table 1

Measurement of variables.

No	Variable	Description/Question	Coding
Outcome variable
1	Unskilled Birth Attendance	Who assisted with the delivery?	0 = skilled birth attendant
			1 = unskilled birth attendant
Explanatory/independent variables
Individual variable
1	Level of Education	Level of Education	0 = no education
			1 = primary
			2 = secondary
			3 = higher
2	Age	Age of respondents	1 = 15–19
			2 = 20–24
			3 = 25–29
			4 = 30–34
			5 = 35–39
			6 = 40–44
			7 = 45–49
3	Parity	Birth history of women	1 = Primiparous
			2 = Multiparous
			3 = Grand multiparous
4	Health insurance coverage	Covered by health insurance	0 = no
			1 = yes
5	Getting medical help for self: distance to a health facility	Whether the respondents have a problem with distance, getting medical help	1 = big problem
			2 = not a problem
6	Media exposure	Media exposure looks at the exposure of respondents to listening to the radio, watching television, and reading the newspaper/magazine	1 = low
			2 = medium
			3 = high
Community variables
7	Type of residence	Type of residence	1 = urban
			2 = rural
8	Community socioeconomic status	Community socioeconomic status	1 = low
			2 = high
9	Community literacy	Community literacy	1 = low
			2 = medium
			3 = high
10	Region of residence	Region of residence of respondents	1 = Western
			2 = Central
			3 = Greater Accra
			4 = Volta
			5 = Eastern
			6 = Ashanti
			7 = Brong Ahafo
			8 = Northern
			9 = Upper East
			10 = Upper West

Source: GDHS, 2014

Table 2

Background characteristics and unskilled birth attendant.

Variable	Frequency N = 4,290	Percentage (%)	Proportion of unskilled birth attendant
Individual Characteristics
Level of education
No education	1,117	26.0	44.7
Primary	822	19.7	29.9
Secondary	2,133	49.7	13.0
Higher	198	4.6	1.3
Age
15–19	190	4.5	21.1
20–24	728	17.0	25.5
25–29	1,037	24.2	22.0
30–34	1,006	23.4	25.1
35–39	808	18.8	20.5
40–44	396	9.2	27.8
45–49	125	2.9	39.3
Parity
Primiparous	977	22.8	13.2
Multiparous	2,231	52.0	22.4
Grand multiparous	1,982	25.2	37.3
Covered by health insurance
No	1,425	33.2	31.3
Yes	2,865	66.8	20.4
Wealth status
Poorest	901	21.0	50.4
Poorer	871	20.3	37.3
Middle	857	20.0	21.7
Richer	844	19.7	5.1
Richest	817	19.0	2.8
Getting medical help for self: distance to a health facility
Big problem	1,138	26.5	37.4
Not a big problem	3,152	73.5	19.2
Media exposure
Low	405		49.5
Medium	1,107		33.2
High	2,778		16.7
Community variables
Type of residence
Urban	1,981	46.2	8.7
Rural	2,309	53.8	37.2
Community socioeconomic status
Low	2,953	68.8	35.6
High	1,337	31.2	7.3
Community literacy
Low	1,440	33.6	44.8
Medium	1,437	33.5	26.3
High	1,413	32.9	8.3
Region of residence
Western	442	10.3	22.3
Central	472	11.0	27.3
Greater Accra	698	16.3	6.8
Volta	327	7.6	29.1
Eastern	403	9.4	30.7
Ashanti	764	17.8	12.4
Brong Ahafo	388	9.0	19.7
Northern	497	11.6	60.5
Upper east	184	4.3	15.2
Upper west	115	2.7	32.5

Source: GDHS, 2014

Table 3

Multilevel logistic regression of unskilled birth attendants among women in Ghana.

Variables	Model 1	Model II aOR (95% CI)	Model III aOR (95% CI)	Model IV aOR (95% CI)
Level of education
No education		Ref		Ref
Primary		0.77**(0.60, 0.97)		0.81(0.63, 1.03)
Secondary		0.49***(0.38, 0.63)		0.52***(0.40, 0.67)
Higher		0.07*(0.01, 0.53)		0.07*(0.01, 0.55)
Age
15–19		1.01(0.62, 1.66)		1.01(0.62, 1.66)
20–24		0.89(0.67, 1.19)		0.90(0.67, 1.20)
25–29		Ref		Ref
30–34		0.82(0.63, 1.08)		0.87(0.66, 1.14)
35–39		0.60**(0.44, 0.82)		0.64**(0.47, 0.87)
40–44		0.59**(0.40, 0.85)		0.65**(0.45, 0.94)
45–49		0.71(0.43, 1.16)		0.75(0.46, 1.23)
Parity
Primiparous		0.49***(0.36, 0.66)		0.51***(0.38, 0.70)
Multiparous		Ref		Ref
Grand multiparous		1.30*(1.01, 1.67)		1.26(0.97, 1.62)
Covered by health insurance
No		1.81***(1.49, 2.19)		1.80***(1.48, 2.19)
Yes		Ref		Ref
Wealth status
Poorest		Ref		Ref
Poorer		0.80(0.63, 1.03)		0.78(0.60, 1.01)
Middle		0.46***(0.34, 0.63)		0.53***(0.38, 0.70)
Richer		0.10***(0.06, 0.16)		0.16***(0.10, 0.27)
Richest		0.08***(0.04, 0.16)		0.17***(0.08, 0.36)
Getting medical help for self: distance to health facility
Big problem		1.21(0.99, 1.48)		1.10(0.90, 1.35)
Not a big problem		Ref		Ref
Media exposure
Low		1.42*(1.06, 1.89)		1.35*(1.01, 1.80)
Moderate		0.99(0.80, 1.24)		0.96(0.77, 1.20)
High		Ref		Ref
Community level
Place of Residence
Urban			0.33***(0.23, 0.48)	0.52**(0.35, 0.77)
Rural			Ref	Ref
Community socioeconomic status
Low			Ref	Ref
High			0.48**(0.30, 0.77)	0.83(0.50, 1.34)
Community literacy
Low			3.91***(2.51, 6.10)	2.06**(1.32, 3.23)
Medium			2.25***(1.54, 3.28)	1.60*(1.10, 2.34)
High			Ref	Ref
Region of residence
Western			0.99(0.49, 2.04)	0.92(0.45, 1.89)
Central			1.44(0.72, 2.89)	1.07(0.53, 2.17)
Greater Accra			Ref	Ref
Volta			1.75(0.85, 3.60)	1.17(0.56, 2.37)
Eastern			1.64(0.81, 3.32)	1.16(0.56, 2.37)
Ashanti			0.72(0.35, 1.48)	0.59(0.28, 1.23)
Brong Ahafo			0.62(0.30, 1.29)	0.42*(0.21, 0.91)
Northern			3.13**(1.15, 6.47)	1.58(0.75, 3.36)
Upper east			0.40*(0.19, 0.85)	0.20***(0.09, 0.43)
Upper west			0.79(0.37, 1.69)	0.49(0.22, 1.07)
Random effect result
PSU variance (95% CI)	2.65(2.09, 3.35)	1.17(0.89, 1.54)	0.85(0.62, 1.15)	0.71(0.51, 0.98)
ICC	0.45	0.26	0.20	0.18
LR Test	χ2 = 854.06	χ2 = 281.75	χ2 = 185.05	χ2 = 130.24
	p = 0.0000	p = 0.0000	p = 0.0000	p = 0.0000
Wald chi-square		356.45	302.07	463.11
Model fitness
Log-likelihood	-2082.78	-1864.82	-1940.67	-1810.84
BIC	4182.28	3905.29	4006.80	3906.06
AIC	4169.55	3771.65	3911.33	3690
N	4,290	4,290	4,290	4,290

*p<0.05

**p<0.01

*** p<0.001

Source: GDHS, 2014

Model fit and specifications

We assessed the fitness of all the models with the Likelihood Ratio (LR) test. The presence of multicollinearity between the independent variables was checked before fitting the models. The variance inflation factor (VIF) test revealed the absence of high multicollinearity between the variables (Mean VIF = 2.28).

Spatial analysis

In the conduct of the survey, instead of mapping outhouses in which the data were collected, clusters were mapped to protect the actual identity and location of respondents [24]. These clusters are developed to suit the district-level data, making it easy to merge the household records with spatial data. During the data collection period, there were 216 administrative districts in Ghana; however, not all districts had respondents drawn from for the survey. This aided in the merger of the data gathered with the district shapefiles obtained from the Department of Geography and Regional Planning, University of Cape Coast, Ghana. This was done to permit the analysis to be made at a district level. The data is best analysed at the district level since the information is more representative at the cluster level [24]. This study extracted the required variables from the 2014 GDHS. The extracted data maintained the mapped clusters information. This mapped cluster information was used to help join the extracted non-spatial data to the coordinates gathered for the clusters. All the data required (GDHS data and 216 district boundary) were projected into the projected coordinate system of Ghana Meter Grid to aid in the spatial analysis. The extracted GDHS data were merged with coordinate, and a spatial join was undertaken to transfer the cluster point to the 216-district boundary (polygon) layer using ArcMap version 10.5. This activity enabled us to easily identify and trace where each case is located within a district. It was identified that some of the district boundaries had more than one cluster. In such cases, the data from the clusters were aggregated, and their means were computed to represent the respective district they fell within [24].

With regards to the geospatial analyses, four spatial statistical tools were applied to analyse the data. These tools were spatial autocorrelation (Global Moran’s I), hot spot analysis (Getis-Ord G), outlier and cluster analysis, and Geographically Weighted Regression. The spatial autocorrelation was used to assess whether unskilled birth attendance in Ghana had a clustering or dispersion pattern at the district level. This study hypothesized that unskilled birth attendance is randomly distributed across various districts in the country. The null hypothesis is rejected if a calculated p-value is small (95% confidence interval), which implies an unlikely situation that the observed spatial pattern results from random processes [24]. Further, hot spot analysis (Getis-Ord G) was used to ascertain statistically significant spatial variations in unskilled birth attendance [24, 25]. This analysis was conducted to determine districts with high prevalence against areas of the low prevalence of unskilled birth attendance. In addition, an outlier and cluster analysis was conducted to identify districts that appeared as outliers. Outlier districts could either be a hot spot district that is surrounded by cold spot districts and vice-versa. The geographically weighted regression (GWR) modelling was conducted after ascertaining the hot spot and cluster and outlier analysis of unskilled birth attendance, the geographically weighted regression (GWR) modelling was conducted. This spatial regression modelling was performed to identify which explanatory variables best account for the observed spatial patterns of unskilled birth attendance [25]. To be specific, the GWR uses the OLS coefficient from the clusters concerning its nearest neighbours in modelling the predictability of the explanatory variable. The output shows how the strength of each explanatory variable changed across space. Therefore, maps of the statistically significant coefficients were generated.

Results

Background characteristics and unskilled birth attendant

Forty–four per cent of women with no education utilized the service of an unskilled birth attendant during delivery. A high proportion of women with grand multiparous (37.3%), poorest wealth status (50.4%), low media exposure (49.5%), and those who saw the distance to a health facility as a big problem (37.4%) utilized the service of an unskilled birth attendant during delivery (Table 2). Three in ten women from rural areas and communities with low socioeconomic status utilized the service of an unskilled attendant during delivery (Table 2).

Multilevel logistic regression of unskilled birth attendants

Results for all the models were presented in Table 2. From the final model (Model IV), our study found out that level of education, age, parity, health insurance coverage, wealth status, media exposure, place of residence, community literacy had a significant relationship with unskilled birth attendants.

Women with higher education had a lesser likelihood (aOR = 0.07, CI = 0.01, 0.55) of utilizing the services of unskilled birth attendants compared to those with no education. Primiparous women have a lesser odd (aOR = 0.51, CI = 0.38, 0.70) utilization of the service of unskilled birth attendants during delivery compared to multiparous women (see Table 2). Women aged 35–39 years were less likely (aOR = 0.64, CI = 0.47, 0.87) to utilize the service of unskilled birth attendants compared to those 25–29 years. Women who were not covered by health insurance had a higher likelihood of utilizing delivery services from unskilled birth attendants (aOR = 1.80, CI = 1.48, 2.19) compared to those covered by health insurance. We also found that women with the richest wealth status had a lesser odd of utilizing delivery services from unskilled birth attendants (aOR = 0.17, CI = 0.08, 0.36) compared to those with the most deficient wealth status (Table 3). Women with low media exposure had a higher likelihood of utilizing delivery services from unskilled birth attendants (aOR = 1.35, CI = 1.01, 1.80) compared to those with high media exposure (Table 3).

Furthermore, women from urban centres had a lesser odd of utilizing delivery services from unskilled birth attendants (aOR = 0.52, CI = 0.35, 0.77) compared to those from the rural areas. Women from communities with low literacy had a higher chance of utilizing delivery services from unskilled birth attendants (aOR = 2.06, CI = 1.32, 3.23) compared to those from communities with high literacy levels (Table 3).

Spatial distribution on unskilled birth attendance

The spatial distribution of unskilled birth attendance results from Moran’s I spatial autocorrelation analysis (S1 Appendix) revealed that unattended skilled birth was clustered. This implies that unskilled birth attendance was not randomized. Therefore, the non-randomised distribution of unskilled birth attendance led to exploring further to visualize the distribution of unskilled birth attendance in Ghana using Hotspot Analysis (GetisOrd Gi).

The hotspot analysis shows areas of statistically significant high and low intensity of the distribution of a phenomenon under study. According to Fig 1, there is a statistically significant (99% confidence level) high clustering of unskilled birth attendance in some parts of north-eastern Ghana (areas shown in red colour). This comprises districts such as Mion, Savelugu-Nanton, Nanimba North, Gushiegu, Kpandai, Zabzugu, Nanumba South, Saboba, Nkwanta South, Yendi Municipal, Karaga, Nkwanta North and Krachi Bchumuru. However, some parts of the south comprising of districts such as Akwapim North, New Juaben Municipal, Shai Osudoku, Ga South Municipal, La-Nkwantanang-Madina, Adenta Municipal, Kpone Katamanso, Accra Metropolis, Lower Manya Krobo, Ashaiman Municipal, Tema Metropolis, Ningo Prampram, Ga East Municipal and La-Dade-Kotopon Municipal showed a cold spot indicating high skilled birth attendance (areas shown in blue colour). Although the hotspot analysis identified statistically significant clustered areas, the cluster and outlier analysis was conducted to show districts with high unskilled birth attendance but were not captured in the hotspot analysis. The hotspot analysis does a cluster analysis and zones areas as hot or cold spots when an area has a similar occurrence of a phenomenon with its neighbours. This necessitated the conduct of the cluster and outlier analysis to identify outlier areas.

Fig 1

Hotspot analysis of unskilled birth attendance in Ghana.

Source: Authors’ construct (2021).

The cluster and outlier analysis (Fig 2) revealed that about 12 districts (marked in the colour red); Ahafo Ano South, Ayensuano, Shai Osudoku, Wa East, Jaman North, Tano North, Atwima Nwabiagya, Wassa Amenfi East, Bosome Freho, Sekyere South, Lower Manya Krobo and Ho West, that are randomly distributed have high unskilled birth attendance but could not be captured as a hotspot since they were found closer to areas with low unskilled birth attendance. For instance, Shai Osudoku and Lower Manya Krobo, identified as cold spot after the cluster and outlier analysis, have revealed that they have high unskilled birth attendance but are surrounded by neighbours with high skilled birth attendance. Also, two districts (marked in colour blue), Krachi Nchumuru and Sene West, were found to have low unskilled birth attendance but closer to districts with high unskilled birth attendance. This result reveals that relying on hotspot analysis is reasonable when considering areas with statistically significant similar characteristics to their neighbours, whereas cluster and outlier analysis aids in identifying areas statistically significant with distinctive characteristics.

Fig 2

Cluster and outlier analysis of unskilled birth attendance in Ghana.

Source: Authors’ construct (2021).

Further analysis was conducted to assess factors that account for the spatial distribution of unskilled attendance in Ghana. These factors were assessed using the Geographically Weighted Regression (GWR) analysis. However, to conduct GWR, the ordinary least square (OLS) analysis was conducted to identify significant factors that explain the existing distribution of unskilled birth attendance in Ghana. The robust probability results of the OLS (Table 4) revealed that four out of five factors could geographically explain the spatial distribution of unskilled birth attendance in Ghana.

Table 4

OLS of factors that accounts for unskilled birth attendance in Ghana.

Variable	Coefficient	StdError	t-Statistic	Probability	Robust_SE	Robust_T	Robust Probability
Intercept	-0.013	0.038	-0.335	0.738	0.031	-0.409	0.683
Distance to Facility	0.272	0.057	4.789	0.000*	0.062	4.403	0.000*
Media exposure	0.214	0.101	2.118	0.036*	0.121	1.767	0.079
NHIS Subscription	0.193	0.068	2.837	0.0053*	0.068	2.86	0.005*
Community economic status	0.122	0.043	2.857	0.005*	0.032	3.761	0.000*
Community literacy	0.097	0.037	2.619	0.010*	0.046	2.115	0.036*

Source: GDHS, 2014.

The four significant predictors identified from the OLS results were used to conduct the Geographically Weighted Regression (GWR) analysis (S2 Appendix). The GWR is a local model that estimates the different responses of the dependent variable on independent variables based on locational effects. Table 3 shows the GWR model for unskilled birth attendance in Ghana. Since GWR localizes the results of a phenomenon under study, it improves the model fit when the relationship between the predictors and short birth interval is non-stationary [25]. This is evident from the R2 and adjusted R2 values where the OLS value were 0.352 and 0.335 while the GWR values were 0.623 and 0.509, respectively. The GWR’s adjust R2 value explains about 50% of the phenomenon under study (S3 Appendix). The subsequent Figs 3–6 show the spatial predictive power of the explanatory variables towards the unskilled birth attendance in Ghana. Areas shaded red have a strong predictive power, while areas shaded blue have a low predictive power of the explanatory variable.

Fig 3

Distance to facility GWR coefficient for predicting unskilled birth attendance in Ghana.

Source: Authors’ construct (2021).

Fig 6

Community literacy level GWR coefficient for predicting unskilled birth in Ghana.

Source: Authors’ construct (2021).

Regarding the distance to a health facility (Fig 3), the north-eastern part of Ghana (shaded red) had a high positive relationship with unskilled birth attendance. This implies that as the proportion of women who identified distance to a health facility as a problem increased, the occurrence of unskilled birth attendance in the north-eastern part of Ghana increases about 44%-77%.

Also, media exposure was a strong predictor of unskilled birth from the north-east through to the mild belt of Ghana (Fig 4). A unit decrease in media exposure would result in about a 30 to 65% increase in unskilled birth from the north-east through to the mild belt of Ghana.

Fig 4

Media exposure GWR coefficient for predicting unskilled birth attendance in Ghana.

Source: Authors’ construct (2021).

National health insurance subscription was a predictor from west of the northern region through to the southeast of Ghana (Fig 5). A unit increase in national health insurance subscriptions will result in about a 49 to 72% increase in skilled birth attendance in Ghana.

Fig 5

NHIS subscription GWR coefficient for predicting unskilled birth attendance in Ghana.

Source: Authors’ construct (2021).

Finally, it was found that community literacy level serves as a high predictor of unskilled birth attendance in the north-eastern and southwestern parts of Ghana (Fig 6). As community literacy reduces, unskilled birth attendance increases by about 17 to 25%.

Discussion

Among the variables which adjusted for in the ordinary least squares (OLS), the study found that distance to health facility, national health insurance subscription, community economic status, and community literacy as strong predictors of unskilled birth attendance in Ghana for the spatial analysis (i.e. OLSM) while maternal educational level and age, parity, national health insurance subscription, household wealth, media exposure, place of residence, community-level literacy and region of residence were key findings associated with unskilled birth attendance in the non-spatial model. The nonspatial model explained the variation and clustering that exist in unskilled birth attendance in Ghana. For this reason, appreciating the spatial pattern of the analysis of unskilled birth attendance was deemed essential since it helps visualise the phenomenon under study. Therefore, to complement our analysis, we fitted a spatial model to identify the established pattern further and to enable us to provide some justification for the variation in unskilled birth attendance across the country. Generally, the findings (the predictor variables found in the GWR analysis) were like the multilevel analysis conducted in this study. However, our discussions will be based on the results of the spatial model (GWRM) because that is the focus of this study.

The study’s findings suggested that unskilled birth attendance has a significant spatial variation in Ghana. The hot spot analysis found a total of 16 statistically significant districts at various levels of significance with a high prevalence of unskilled birth attendance. Significant hotspots of unskilled birth attendance were seen in the north-eastern part of Ghana. This finding is in line with findings from Dickson and Amu [20]. They reported that women in the Northern and Upper West regions of Ghana were less likely to access skilled assistance during delivery. Variations in skilled birth attendance have been reported in most developing countries [26, 27]. These variations could be due to the physical inaccessibility and several socio-economic hardships [16], coupled with inequalities in the distribution of scarce resources [17], including inadequately skilled health professionals in remote/border areas of Ghana. In addition, socio-cultural beliefs and practices among women in different districts are potential drivers for the prevalence of unskilled birth participation in Ghana [27, 28]. Furthermore, mothers from border districts may have limited access to information concerning maternal health services and other services such as access to school/education that could affect their chances of accessing the services of the skilled birth attendant, as was also reported by Antwi and his colleagues [9].

The cluster and outlier analysis was used to verify and complement the hot spot analysis as it allows us to detect both groupings and areas where anomalies exist [29]. The cluster and outlier analysis results show aspects that may have been overlooked in the hot spot analysis but are interesting highlights, especially in those areas where diverse types of groupings coexist to influence the study’s outcome. Districts labelled high-low outliers are anomalous districts with a high hotspot for unskilled birth attendance, but districts with high skilled assisted deliveries surround them. The reason for this outcome is rooted in the uneven distribution of scarce resources. Most areas with fewer populations do not get the needed infrastructural facilities and personnel since some professionals refuse to post to rural communities in Ghana [12]. A similar explanation can be extended to low-high outlier areas were districts with high unskilled birth attendance surround districts with low unskilled birth attendance.

The GWR analysis revealed a strong positive relationship between mothers who perceived distance to the health facility as a big problem, those not covered by health insurance, those who resided in communities with low literacy and socio-economic levels with unskilled birth attendance in different districts in Ghana. However, exposure to social media could not predict unskilled birth attendance among various districts in the country. There was a higher likelihood of not visiting the health care facility for those women who were not covered by health insurance in the hotspot districts. This finding is in line with studies conducted in Ghana [30], and elsewhere including Ethiopia [10] and Kenya [27]. A possible explanation could be because of additional costs incurred if payment for healthcare services is made out of pocket. Similarly, women who are not covered by health insurance in hotspot areas are highly likely to seek alternative health care services different from their peers from cold spot areas. Our study supports this—areas with higher proportions of people with health insurance subscriptions tend to be cold spots for unskilled births.

The study further found that district-level characteristics are key influencers of unskilled birth delivery in Ghana. A corroboratory to NHIS coverage was the finding that women residing in districts with low socio-economic levels are more likely to access the services of unqualified and unskilled birth attendants than women residing in districts with medium or higher socio-economic levels. This finding is supported by a study done in Nigeria, where the utilisation of unskilled birth attendance services was high among women living in rural communities with low socioeconomic status [26] and, Ethiopia where women from the poorer and poorest households used less skilled assistance during delivery [27]. Generally, women in districts with higher poverty levels are less likely to pay for health insurance subscriptions [31]. The unavailability of health care facilities to readily cater to the health needs of women in poorer districts significantly influences their utilisation of SBA services.

Moreover, participants who perceived distance from the health facility in hotspot districts as a major problem were less likely to seek services for skilled attendant delivery. This is in line with studies done in Ethiopia [27] and rural Nigeria [26]. This is because the distance to and from health facilities and lack of transport are crucial factors preventing mothers from seeking and using skilled maternity services.

Conclusion

In Ghana, unskilled birth attendance had spatial variations and clustering across different parts of the country. About 13 randomly distributed districts have high unskilled birth attendance. Districts with high proportions of unskilled birth attendance were detected in the North-Eastern parts of the country. In the GWR analysis, districts with higher proportions of low socio-economic class, low literacy status, highly perceived distance to and from the health facility, and not covered by health insurance were found to have an increased probability of influencing unskilled birth attendance. Due to such spatial variations and clustering, interventions aimed at promoting skilled birth attendance should include primary health care accessibility, particularly at the north-eastern part of Ghana, increase in media coverage, particularly at the north-eastern part through to the middle belt of Ghana, increase in national health insurance subscription particularly at the northern through to the south eastern part of Ghana, and improvement in community literacy level particularly at the north eastern and southwestern parts of Ghana.

Spatial autocorrelation of unskilled birth in Ghana.

(DOCX)

Click here for additional data file.

OLS diagnostics.

(DOCX)

Click here for additional data file.

GWR model for unskilled birth attendance in Ghana.

(DOCX)

Click here for additional data file.

7 May 2021

PONE-D-21-07780

Multilevel Geospatial Analysis of unskilled birth attendant in Ghana

PLOS ONE

Dear Dr. Bediako,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Jun 19 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Kannan Navaneetham, PhD

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. Your ethics statement should only appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please move it to the Methods section and delete it from any other section. Please ensure that your ethics statement is included in your manuscript, as the ethics statement entered into the online submission form will not be published alongside your manuscript.

4. We note that Figures 1-7 in your submission contain map images which may be copyrighted. All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For these reasons, we cannot publish previously copyrighted maps or satellite images created using proprietary data, such as Google software (Google Maps, Street View, and Earth). For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.

We require you to either (1) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (2) remove the figures from your submission:

4.1.    You may seek permission from the original copyright holder of Figures 1-7 to publish the content specifically under the CC BY 4.0 license.

We recommend that you contact the original copyright holder with the Content Permission Form (http://journals.plos.org/plosone/s/file?id=7c09/content-permission-form.pdf) and the following text:

“I request permission for the open-access journal PLOS ONE to publish XXX under the Creative Commons Attribution License (CCAL) CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). Please be aware that this license allows unrestricted use and distribution, even commercially, by third parties. Please reply and provide explicit written permission to publish XXX under a CC BY license and complete the attached form.”

Please upload the completed Content Permission Form or other proof of granted permissions as an "Other" file with your submission.

In the figure caption of the copyrighted figure, please include the following text: “Reprinted from [ref] under a CC BY license, with permission from [name of publisher], original copyright [original copyright year].”

4.2.    If you are unable to obtain permission from the original copyright holder to publish these figures under the CC BY 4.0 license or if the copyright holder’s requirements are incompatible with the CC BY 4.0 license, please either i) remove the figure or ii) supply a replacement figure that complies with the CC BY 4.0 license. Please check copyright information on all replacement figures and update the figure caption with source information. If applicable, please specify in the figure caption text when a figure is similar but not identical to the original image and is therefore for illustrative purposes only.

The following resources for replacing copyrighted map figures may be helpful:

USGS National Map Viewer (public domain): http://viewer.nationalmap.gov/viewer/

The Gateway to Astronaut Photography of Earth (public domain): http://eol.jsc.nasa.gov/sseop/clickmap/

Maps at the CIA (public domain): https://www.cia.gov/library/publications/the-world-factbook/index.html and https://www.cia.gov/library/publications/cia-maps-publications/index.html

NASA Earth Observatory (public domain): http://earthobservatory.nasa.gov/

Landsat: http://landsat.visibleearth.nasa.gov/

USGS EROS (Earth Resources Observatory and Science (EROS) Center) (public domain): http://eros.usgs.gov/#

Natural Earth (public domain): http://www.naturalearthdata.com/

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Reviewer’s Comments

Multilevel geospatial analysis of unskilled birth attendant in Ghana

The authors examined an important aspect maternal health especially in Ghana. However, I have some few comments for consideration.

General observations

The topic should be revised to “Multilevel geospatial analysis of factors associated with unskilled birth deliveries in Ghana”

The manuscript requires proofreading to improve upon the English Language

Referencing has been done haphazardly. There is no coherence right from the start of writing. E.g. in the Background, the first reference was numbered (41), then followed with (36) and (13). This inconsistency is obvious throughout the whole work.

Background

Paragraphs 2 and part of paragraph 3 should not be part of the background. These paragraphs highlights issues on the definition of skilled births and some related statistics of skilled births. The focus of the study is unskilled births. Therefore, the background should succinctly concentrate on unskilled births. This will distinguish it from studies that have examined skilled births in Ghana.

Also, the opening sentence of the last paragraph of the background suggests a special interest in a particular ecological zone in Ghana. This I believe is the case. The study considered the country. This too needs to be revised. Rather highlight unskilled birth situations in the country.

The whole background should be revised in a coherent manner to inform the reader exactly what the study is meant to examine, thus unskilled births in Ghana.

Methods

Source of data

“For the study, women with birth history who had given birth up to five years before the survey were included”. These inclusion criteria is not clear. With a period of five years, a woman can comfortably give birth to at least two children assuming there are no any medical complications. And these births can occur at home or at health facility or both at home or both at a health facility. So, which one of these children should the study be on? For uniform selection of women, only the last birth of the women aged 15-49 years preceding the survey should be included in the study. If this done, it will certainly affect the current sample size of 4,290 women; meaning re-analysis is required.

Outcome variable

How was the outcome variable coded?

Explanatory variable

It will be interesting for the authors to indicate the various responses of the variables as captured in the GDHS and then indicate how each of them was recoded, where necessary.

Statistical analysis

Authors should indicate how variables were selected for the regression models. It seems preference was given to some variables at each stage of the modelling.

Spatial analysis

This section is too detailed. It should be summarized in two paragraphs for purposes of sustaining the interest of reader and also making the content simple.

Results

Background characteristics

• To refer to a table in text, state it as (Table 1) and not as (see Table 1).

• On Table 1, proportion for ‘Richest’ category of wealth status is missing.

Spatial distribution on unskilled births attendants

Some of the spatial figures and tables can be on supplementary sheets. Some of these include:

a. Fig. 1: Spatial autocorrelation of unskilled births in Ghana

b. Table 4: OLS Diagnostics

c. Table 5: GWR model for unskilled birth attendance in Ghana

The authors failed to mention the specific districts that are associated with unskilled births in Ghana. It is important to mention these districts to better inform policy makers and readers where unskilled births are high within the country.

The discussion was well presented however the referencing style is not appropriate.

Conclusion too was well presented.

Reviewer #2: Authors have adequate literature review and analysis of the data in this study. This study is novel enough to be published.

Small typographical errors and a few unclear sentences to correct. Standardize use of Fig. and Figures. Abbreviations and acronyms should only be spelled out at first use.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: Reviewer comments.doc

Click here for additional data file.

Submitted filename: PONE-D-21-07780_reviewer_Kelley.pdf

Click here for additional data file.

1 Jun 2021

In submitting our revised manuscript entitled “Multilevel Geospatial Analysis of Factors Associated with Unskilled Birth Attendance in Ghana”, we would like to draw the attention of the journal to a copyright issue which is related to the maps used in the manuscript. We state emphatically that all maps in the manuscript were designed by the authors based on the data used for the study. All the authors are in the know and consent to this fact.

Thank you.

Thank you.

Vincent Bio Bediako

(Corresponding Author)

Submitted filename: Response to Reviewers.doc

Click here for additional data file.

9 Jun 2021

Multilevel Geospatial Analysis of Factors Associated with Unskilled Birth Attendance in Ghana

PONE-D-21-07780R1

Dear Dr. Bediako,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Kannan Navaneetham, PhD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

18 Jun 2021

PONE-D-21-07780R1

Multilevel Geospatial Analysis of Factors Associated with Unskilled Birth Attendance in Ghana.

Dear Dr. Bediako:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Prof. Kannan Navaneetham

Academic Editor

PLOS ONE