Empirical Analysis of Equity in Maternal and Child Health Services Delivery and Access in Ghana: A Cross-Sectional Study


 Background

Inequities in the distribution of and access to maternal and child health care services is pervasive in Ghana. Understanding the drivers of inequity in maternal and child health (MCH) is important to achieving the universal health coverage component of the sustainable development goals and poverty reduction in Ghana and other developing countries. However, there is increasing disparities in MCH services, especially in rural -urban and income quintiles. The study aimed to examine the disparities in maternal and child health care services in Ghana for policy intervention.
Methods

Data for this study was extracted from the nationally representative Ghana Statistical Service (GSS) Multiple Indicator Cluster Survey (MICS) round 4, 2011. Respondents of this survey were women of reproductive age 15–49 years with a sample size of 10,627 households. The models were estimated using multivariate regression analysis together with concentration index (CI) and risk ratio (RR) to assess the distribution of MCH indicator groups across the household wealth index.
Results

Higher educational attainment played an important role in MCH. Women with secondary school level and above were more likely to receive family planning, prenatal care, and delivery by a skilled health professional than those without formal education. Mothers with low level of educational attainment were 87% more likely to have their first pregnancy before the age of 20 years, and 78% were more likely to have children with under-five mortality, and 45% more likely to have children who had diarrhoea. Teenage pregnancy, under five mortality, child underweight, reported diarrhoea, and suspected pneumonia were more concentrated in the poorer than in the richer households. The RR between the top and bottom quintiles ranged from 0.77 for child underweight to 0.82 for child wasting.
Conclusion

Geographic location, income status and formal education are key drivers of maternal and child health inequities in Ghana. Implementing health policies to address inequalities in MCH services through primary health care, and resource allocation skewed towards rural areas and the lower wealth quintile can bridge the inequality gaps and improve MCH outcomes in Ghana.

Research shows that educated women are more likely to start antenatal care (ANC) 92 visits earlier than less educated women(9), and utilisation of delivery care depends largely on 93 the women's educational level (10). Most maternal health studies in Ghana. (11,12) 119 Family planning has been an integral component of government of Ghana's maternal health 120 programs for decades. Family Planning is an important factor in the population management 121 and national development outlined in many national development plans (15).

122
Family planning aims to assist couples and individuals of reproductive age to achieve their 123 reproductive aspirations. Despite the high premium placed on family planning programs in 124 Ghana, funding remains a daunting challenge. Family planning intake is highest (69%) among 125 women between 15-19 years and lowest (33%) among women within 45-49 years. The demand 126 for family planning is also highest (59%) among women in rural areas. Those women with 127 at least primary or high school education use more family planning services and women in 128 the middle three quintiles (60-61% ), (16).

129
Although there is a huge progress in family planning services intake, there is still bout 130 50% unmet need for family planning services in Ghana, especially among young women within 131 the 15-19 years (51%) and lowest among women aged 45-59 (14%). Also women in rural 132 dwelling have slightly higher (31%) unmet needs of family planning than their counterparts in 133 urban areas ( 29% ) (16).

134
Delivery by skilled health personnel is another key indicator of maternal health. There 135 has been progress in this indicator, about 68% of all births in the last two years preceding the 136 MICS survey round 6 were delivered by skilled personnel. Education plays an important role 137 in deliveries by skilled health personnel. Educated woman were more likely to have assisted 138 delivered by a skilled health personnel. Assisted delivery by skilled health personnel for 139 mothers with no formal education constituted only 44% of all deliveries compared to 95% for 140 women with secondary or higher education. Also, poor women were less likely to deliver using 141 skilled personnel (39%), compared to rich women (98%). Despite the progress made in 142 delivery by skill personnel deliveries at home is still highly significant as 1 in 3 births take 143 8 place at home without a skilled health personnel (17).This needs to be addressed to reduce 144 preventable maternal mortality which might emanate from complications or blood loss.

145
Promoting and ensuring deliveries in health facilities can reduce the health risks to both the 146 mother and the baby. Proper medical attention and hygienic conditions during delivery can also 147 reduce the risks of complications and infection that can cause morbidity and mortality to either 148 the mother or the baby (17).

150
The health of children is a global concern. Over the years, many countries and institutions have 151 worked towards improving the health of children to reduce infant mortality. Despite the 152 significant investments and improvement in child health in the past few decades, many children 153 still lose their lives to diseases before their 5th birthday globally, and inequity in health is still

169
The main intervention in treating pneumonia is antibiotics. Of the 3% suspected pneumonia 170 cases reported in Ghana, 41% of them were taken to an appropriate health provider and 56% 171 received antibiotics. Children in rural areas and/or poor homes are disadvantaged in terms of 172 care seeking behaviour (17).

173
WHO estimates that at least 10 million deaths were prevented between 2010 and 2015 globally 174 due to vaccinations; and many lives were protected from suffering and disability associated 175 with diseases such as pneumonia, diarrhoea, whooping cough, measles, and polio (20).  Stunting becomes more common as children get older, peaking at 28% among children aged 24-35 186 months. Stunting affects a significantly higher percentage of males (20%) than females (17%), and 187 stunting is more prevalent in rural areas (22%) than in urban areas (15%). Stunting rates vary by area, 188 ranging from 10% in Greater Accra to 33% in the Northern region. Education and income are inversely 189   205 We measured inequities in maternal and child health outcomes and access to health care 206 interventions by three steps: i) Identification of the health outcome or intervention whose 207 distribution is to be measured; ii) classification of the population into different strata by a 208 selected equity stratifier; and iii) measuring the degree of inequality (22). Finally, we tried to 209 understand the drivers of these inequities in MCH utilization. The variables of interest, 210 maternal and child health outcomes and interventions are listed in Table 1. In the Multiple 211 Indicator Cluster Survey, the socio-economic stratifier used is household wealth, which is 212 derived from the household ownership of assets such as television, car etc. and dwelling 213 characteristics such as flooring material and source of drinking water. In this study, we have 214 used wealth quintiles that are provided in the MICS 4 report (17). Each asset was assigned a 215 weight (factor score) generated through principal components analysis, and the resulting asset 216 scores were standardised in relation to a normal distribution with a mean of zero and standard 217 deviation of one. Each household was then assigned a score for each asset, and the scores were 218 summed for each household; individuals were ranked according to the total score of the 219 household in which they resided. The sample was then divided into quintiles from one (lowest) 220 to five (highest). A single asset index was developed for the whole sample; separate indices 221 were not prepared for the urban and rural populations (23).   The index is bounded between -1 and 1. For a discrete living standards variable, it can be 245 written as:

Measurement of inequities
where is the health sector variable, is its mean, and = , N is the fractional rank of  258 Health equity is the absence of unjust, avoidable differences in health care access, quality, or 259 outcomes. Measuring health inequalities allows us to identify differences that can be acted on 260 and can be used to measure progress toward achieving health equity. Disaggregating health 261 indicators using equity stratifiers can identify inequalities between subpopulations. An equity 262 stratifier refers to a characteristic such as a demographic, social, economic, racial, or 263 geographic descriptor that can identify population subgroups for the purpose of measuring 264 differences in health and health care that may be considered unfair or unjust (33).

265
To assess wealth, the study used selected assets and durables in a sample household, because 266 asset ownership tends to fluctuate less than individual income or expenditure. The assets 267 13 considered in houses were permanent floors, roofs, or walls; flush or pour-flush toilets; 268 transportation -including bicycles, motorcycles, cars or trucks; and electrical equipment, 269 including radios, televisions, line or mobile telephones, refrigerators and computers.

270
Households with these assets were considered richer than those without. The study also used

284
The study also compared the prevalence of health outcomes and the coverage of the 285 MCH interventions between the richest and the poorest subgroups using a risk ratio (RR). All 286 households were ranked according to their wealth indices, which was divided equally into 287 quintile (5) and decile (10)  between the richest and poorest quintiles and deciles.

299
Economic disparities in health outcomes 300 We discovered substantial differences between income classes and geographic areas, 301 suggesting that household wealth has a significant impact on child survival and that the poor 302 have a higher risk of child mortality. Teenage pregnancy, low birth weight, infant malnutrition, 303 and child disease all showed economic inequalities in MCH. The poorer subgroups were more 304 likely to have negative health effects (as shown by the negative CIs in Table 2 and 3). The poor 305 had the highest concentration, which was statistically important for child underweight. The CI 306 for stunting and wasting in children was negative, but this had no statistical significance. In 307 terms of magnitude of concentration, teen pregnancy was ranked third among the poor.

308
Children under the age of 5 years old with suspected pneumonia and diarrhoea were also more 309 prevalent among the poor (Table 2 and

321
The results indicate that the primary MCH interventions were spread more evenly across 322 economic strata than the health outcomes (  (Tables 1 and 2); both had more than one for comparisons between the first and fifth 335 quintiles, as well as between the first and tenth deciles.      Table 3 and 4 also summarizes the urban-rural and educational disparities in MCH, as 368 reflected by the RR. The four vaccines (BCG, MMR, DPT and yellow fever) and the coverage 369 indicator appropriate provider for pneumonia were all concentrated in the urban areas than 370 rural areas (thus, by 47%, 52%, 53% and 56% respectively) while low birth, under five 371 mortality, underweight, stunting, wasting, child illness (diarrhoea and suspected pneumonia) 372 and coverage indicator ORS/ORT for diarrhoea were more concentrated in rural than in urban 373 areas. The most profound health gap was under-five mortality, which was 33% more prevalent 374 in rural than in urban areas. Low birth weight and suspected pneumonia were equitably 375 distributed between rural and urban areas.

376
The urban-rural gap for MCH service coverage was quite large. For instance, women 377 living in rural areas were 30% less likely than those in urban areas to receive prenatal and 378 delivery care from a skilled health worker. Although family planning was concentrated in the 379 urban areas than rural, this was not statistically significant. Again, teenagers in urban areas 380 were 73% less risky in getting expose to teenage issues compared to their counterparts in the 381 rural areas. Also, there was a sharp gap between urban and rural women usage of 382 print/electronic media and technology, thus women in urban areas were 4.9-78% more likely 383 to use newspapers, radio, television, computer, and internet than their counterparts in the rural

Educational Inequity and MCH Disparity 387
Mothers' or caregivers' formal schooling is a significant determinant of MCH inequity. Our findings 388 show that more educated mothers or caregivers did better on all outcome indicators. The disparity was 389 most noticeable when it came to teenage pregnancy. Women with less than a secondary school 390 education were 87% more likely than those with a secondary school education to have their first 391 pregnancy before the age of 20. Mothers or caregivers with no formal education were more likely (78 392 percent and 45 percent, respectively) to have under-five mortality and children with diarrhoea than 393 those with a secondary education. Although there was little educational disparity among mothers of 394 children with low birth weight and wasting, these indicators were still more common in the subgroup 395 of mothers or caregivers who were uneducated (See Table 3). 396 Women with education beyond secondary school were 30-46% more likely than those without any 397 formal education to receive family planning, maternal care, and delivery by a professional health 398 worker or in a health facility. Higher educational attainment was also associated with a consistent 399 improvement in maternal care coverage, with a large difference (RR: 1.304-1.457; P <0.01). 400 Surprisingly, children born to mothers or cared for by someone with a post-secondary education were 401 57-66 percent more likely than those who were not in this subgroup to receive all forms of vaccination 402 before the age of one year.     of MCH service coveragebetween the wealthy and the poor, urban and rural populations (7).

480
Over the last few decades, the regional reach of district hospitals and sub-district health centers 481 has tended to increase in favor of the urban and wealthy. District health systems, which include 482 hospitals and health centers, are leading the way in offering a wide variety of curative, 483 preventative, and health-promotion programs, including MCH.

484
In Ghana, there is still some difference in child health results between the rich and the poor, as 485 well as between urban and rural areas. The country's CIs for diarrhoea, malnutrition, 486 underweight, and stunting, for example, are equivalent to the MICS for developing countries.

487
One of the most significant social determinants of health inequity is education. The education 488 disparities for measures of teenage pregnancy and child malnutrition were far greater than the 489 urban-rural differential, according to this report. As the mother's or caregiver's formal 490 education level increased, the prevalence of adolescent pregnancy and infant malnutrition 491 decreased. Teenage pregnancy was found to be much less common among those who had 492 completed high school. This and other studies (13)  interventions. Inequity at birth has long-term consequences; undernutrition, for example, is 504 linked to a loss of human capital (i.e., the skills and knowledge that enable people to work and 505 thus produce economic value). Since, this study was an empirical analysis of an existing data, it did not require ethical 526 clearance.

527
Availability of data for the study 528 The data that support the findings of this study are available from Ghana Statistical Service 529 but restrictions apply to the availability of these data, which were used under license for the 530 current study, and so are not publicly available. Data are however available from Mubarik 531 Salifu, (our author who managed the data) upon reasonable request and with permission of 532 Ghana Statistical Service.

533
Competing Interests 534 We the authors for this research declare that we have no competing interests.