Factors associated with non-use of maternal health services in Botswana.
The study investigated individual and household factors associated
with non-use of maternal health services in Botswana.
Nationally-representative data, drawn from the 1996 Botswana Family
Health Survey, were used. A weighted sample of 19,031 women, aged 15-49
years, who had at least one pregnancy history in the five years prior to
the survey was considered for analysis. Both simple cross-tabulations
and logistic regression were used for analyzing the data. Consistently,
the teenagers were less likely to seek prenatal care, to have their
babies delivered by a qualified person, and to seek postnatal check-up.
Using results from logistic regression analysis, it can be observed that
low-parity women were less likely to use maternal services. Another
consistent finding is that women with low educational level, those
residing in rural areas, and those with low socioeconomic status were
less likely to use maternal services. More focussed investigation is
needed, but understanding the differentials of the use of maternal
services allows policy-makers to identify problem areas that need
Key words: Maternal health services; Women's health; Maternal mortality; Botswana
Maternal health services (Surveys)
Mothers (Patient outcomes)
Rakgoasi, Serai Daniel
|Publication:||Name: Journal of Health Population and Nutrition Publisher: International Centre for Diarrhoeal Disease Research Bangladesh Audience: Academic Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2003 International Centre for Diarrhoeal Disease Research Bangladesh ISSN: 1606-0997|
|Issue:||Date: March, 2003 Source Volume: 21 Source Issue: 1|
|Geographic:||Geographic Scope: Botswana Geographic Code: 6BOTS Botswana|
The International Safe Motherhood Initiative was launched with the aim of reducing maternal mortality throughout the world (1). To achieve this goal, this initiative aspires to enable each woman to have a safe and healthy pregnancy and delivery by providing high-quality maternity services to all. These services include the provision of care by skilled health personnel before, during, and after childbirth. Pregnancy outcomes improve as a result of early and regular use of antenatal and postnatal care services (2). Despite fairly good distribution of maternal and child health/family-planning (MCH/FP) programmes that deliver antenatal and postnatal care in Botswana, some women do not use these facilities. Owuor-Omondi and Kobue noted that toxaemia accounts for an unusually large percentage of maternal deaths, which may imply poor coverage of antenatal care and/or lack of knowledge among women about the complications (3).
In some parts of Africa, the risk of dying due to a pregnancy-related condition is 1 in 16 women (4). According to the World Health Organization (WHO), each year 60 million deliveries take place worldwide in which the woman is cared for only by a family member, an untrained traditional birth attendant, or none at all, while less than 30% of women receive postpartum care (5). Low use of maternal health services is due mainly to long distance from health services, high costs, multiple demands for women's time, and lack of decision-making power of women within the family (5).
The maternal mortality rates in the great majority of sub-Saharan African countries range from 600 to 999 per 100,000 livebirths, while only 10 countries in the region have rates below 600. The maternal mortality rate in Botswana was reported to be 200 per 100,000 livebirths in the early 1990s (3). However, the State of World Population (2001) document reports a maternal mortality rate of 480 for Botswana (6). Human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS) further increases the maternal mortality rate. Albeit high, the comparatively low maternal mortality rate in Botswana, compared to other sub-Saharan African countries, is a result of good maternal services. The percentage of births receiving antenatal care, for instance, has risen from 90% in 1984 to 94% in 1996, while the percentage of medically-supervised deliveries also rose from 66% to 87% for the same period. Despite the fact that many women in Botswana use maternal services, the Government of Botswana desires to "increase the percentage of pregnant women attending antenatal care (at least once during their pregnancy) from 92% in 1988 to 100% in 2011 ..."(7).
No research had been carried out to understand why some women do not use maternal health services that are known to improve the health of both mother and her baby. To achieve 100% coverage, one needs to understand the factors that inhibit the use of maternal services so that appropriate actions can be taken to remove barriers; this is the reason why this study was undertaken. We investigated individual and household factors associated with the non-use of maternal services in Botswana. The study investigated: (i) whether the woman had an institutional delivery; (ii) whether the woman had prenatal check-up; (iii) whether the woman had a qualified delivery assistant; (iv) whether the woman was vaccinated against tetanus during pregnancy; and (v) whether the woman sought postnatal check-up.
MATERIALS AND METHODS
Data used in this study were derived from the 1996 Botswana Family Health Survey (BFHS-III) as part of a worldwide family of surveys, globally known as Demographic and Health Surveys, funded by the U.S. Agency for International Development. The study was designed to yield a nationally-representative random sample of 8,483 women aged 15-49 years. Topics covered in this survey included demographic characteristics, socioeconomic situation, fertility history, and information on health and use of healthcare services.
For the purpose of this paper, the BFHS-III sample was restricted to women who had at least one pregnancy history in the five years prior to the survey. This resulted in 1,805 cases, 621 of which had missing values on variables of interest. The 621 cases were dropped from analysis leaving 1,184 women. An analysis of the missing cases did not suggest any significant selection bias. Thus, women who did not provide a response to questions of maternal services are not uniquely different from the rest of the women in the sample.
Descriptive statistics and logistic regression were used for studying individual and household factors associated with non-use of maternal services.
In this study, five dependent variables and six independent variables were used for analysis. The independent variables were grouped into two broad categories: (i) individual factors (age, parity, education, marital status, and place of residence), and (ii) household factors (socioeconomic status). All the variables were coded in such a manner that the theoretically low-risk group was treated as a reference category. All the dependent variables (non-institutional delivery, no prenatal check-up, unqualified delivery assistance, no tetanus injection, and no postnatal check-up) were binary, while the independent variables were either dichotomous or categorical. Although other variables were fairly straightforward to understand, the household variable needs further explanation. Socioeconomic status is a composite variable created from a number of variables relating to household ownership of radio, television set, car, stand pipe, modern toilet facility (pit latrine or flush system) and whether a household uses electricity for cooking. Each of the above variables was recoded into binary variables with the value 1 indicating a household that did not own any of the items mentioned above and zero to indicate those households that did. A single scaled variable measuring socioeconomic status was then computed from the recoded variables by adding all the binary variables to produce a scaled variable with possible values between zero (for households that owned all of the items) and six (for households that owned none of the items). The categories of this variable were then further recoded into a new variable with three categories of socioeconomic status. Households that had a score between 5 and 6 were classified as 'low' socioeconomic status, those with a score of between 3 and 4 were classified as 'medium' socioeconomic status, while those with a score between 0 and 2 were classified as 'high' socioeconomic status. This composite variability provides a better measure of socioeconomic status than does any individual variable listed above. The WHO states that low status limits the access of women to the economic resources to pay for healthcare or for family-planning services (4).
Methods of data analysis
Simple cross-tabulations were used for examining the bivariate relationships between the independent variables and the dependent variables. Logistic regression analysis was used for evaluating the relationship between a group of predictor variables and the probability for non-use of certain maternal healthcare services, while controlling for other variables in the model. The logistic regression method was used because it provides an interpretable linear model for a categorical dependent variable. It also allows the significance of a given predictor to be tested for while controlling for all other predictors in the model (8).
Separate logistic regression models were used for evaluating the effects of individual and household factors on the probability for non-use of selected maternal services. For the [i.sup.th] individual, this model can be expressed as:
ln [P.sub.I]/(1-[P.sub.i]) = [[beta].sub.0] + [Sigma][[beta].sub.k][x.sub.ki]
where [P.sub.I] is the probability that the [i.sup.th] woman will not use a certain maternal service, [[beta].sub.0] is the baseline constant, [x.sub.ki] is an array of (k) independent variables, and [beta] is the corresponding vector of unknown regression coefficients. The SPSS-PC logistic programme was used for estimating regression coefficients through the maximum likelihood procedure (9,10).
The betas represent the change in the log odds due to unit increments in values of the predictors (8). Interpreting logistic regression results in terms of odds, [e.sup.B], is a summary statistic for the partial effect of a given predictor on the odds, controlling for other predictors in the model.
The results are presented under two sections: (i) individual factors and (ii) household factors. Table 1 presents the descriptive statistics on the proportions of women who were not using a particular maternal service. Table 2-6 show the results of logistic regression analysis on non-use of maternal services. Both odds ratios and confidence intervals are presented in these tables. It should be noted that although bivariate results were used for investigating the relationship between two variables, they may lead to false conclusions because they do not control for potential confounding variables. Therefore, where there appears to be a contradiction in the results from bivariate analysis and logistic regression analysis, more confidence is given to logistic regression results because potential confounding variables have been controlled for.
Individual factors influencing non-use of maternal services
The descriptive results in Table 1 show that unqualified persons attended about 19% of all deliveries by the teenage mothers compared to 12% and 17% by the mothers aged 20-34 years and the mothers aged 35 years and over respectively. About 30% of the teenage mothers had no postnatal check-ups compared to 17% and 16% for the mothers aged 20-34 years and the mothers aged 35 years and over respectively. Although a substantial proportion of the teenage mothers did not appear to use other maternal services, a significant majority of them had institutional deliveries. For example, 93% of all births to teenage mothers were institutional births compared to 90% and 79% among women aged 20-34 and over 35 years respectively. Over 90% of all mothers had at least one antenatal check-up. Antenatal check-up was the most used maternal service, and less than 12% of the women in each age group did not use antenatal check-up.
The results of logistic regression analysis which controls for potential confounders, such as parity, education, marital status, socioeconomic status, and place of residence, indicated that the relative odds of non-use of certain maternal services were significantly higher among the teenage mothers and the mothers aged 20-34 years compared to the mothers aged 35 years and over (Table 2-6). For example, the teenage mothers were 13 times more likely to have no antenatal check-up and were six times more likely to have no postnatal checkup compared to the women aged 35 years and over. The teenage mothers were 11 times more likely to have been assisted during delivery by an unqualified birth attendant compared to the women aged 35 years and above. All the above relationships were statistically significant.
The results showed that the non-use of maternal services was related to women's parity. The proportion of nonuse of maternal services among high-parity women was consistently higher than among low-parity women, except for postnatal check-up (Table 1). Thirteen percent of the women with four or more births did not receive a tetanus injection during their last pregnancy compared to 10% of the low-parity women. These 10% low-parity women were more likely to be teenage mothers. First births are more likely to have delivery-related complications, and, therefore, there is a need to have their deliveries in modern healthcare facilities where help will be available should the need arise.
The results from logistic regression analysis controlling for age, education, marital status, socioeconomic status, and place of residence, indicated that the relative odds of non-use of maternal services were higher among the low-parity women and decreased with increasing parity (Table 2-6). This finding is consistent with the results from a study by Letamo and Majelantle indicating that the "majority of antenatal care non-attenders are teenagers compared to women of other age groups" (11, p.376). The explanation from focus-group discussions revealed that the majority of teenagers often attempted to hide their pregnancies from their parents. The primiparous women were more likely than the multiparous women to have non-institutional deliveries, no antenatal or postnatal services and not to have received a tetanus injection than the multiparous women.
Education showed the strongest relationship with nonuse of institutional delivery. That is, the women with no education were less likely to use maternal services. For example, 36% of all births to women with no formal education were non-institutional compared to 16% and 4% of all births among women with primary and secondary or more education respectively (Table 1).
The results from logistic regression analysis controlling for age, parity, marital status, socioeconomic status, and place of residence indicated that women with no formal education and those with primary education were significantly more likely to have non-institutional births, no prenatal check-up, and had unqualified delivery assistance (Table 2-6). Compared to the women with secondary education or more, the women with no formal education and those with primary education were 11 times and four times more likely to have non-institutional births respectively.
Using descriptive statistics, there was virtually no difference in the proportion of non-use of maternal services between the ever-married and the never-married women (Table 1). However, the results of logistic regression analysis showed that the never-married women were (1.4 times) less likely to have received tetanus injection than the ever-married women (Tables 2 and 4), but the same women were significantly more likely to have institutional births and qualified delivery assistance.
A higher proportion of the women residing in rural areas did not use maternal services (Table 1). For instance, about one in five rural women had non-institutional deliveries compared to only one in 20 in towns. About 30% of the rural women were assisted by an unqualified birth attendant compared to only 6% of the women who resided in towns.
The urban-rural status of women displayed significant effects on patterns of non-use of maternal services, even after controlling for age, parity, education, marital status, and socioeconomic status. The results of logistic regression analysis indicated that the rural women and those residing in villages were less likely to use maternal services compared to the women living in towns (Table 2-6). For example, rural women were more than twice as likely to have had a non-institutional birth and no antenatal check-ups, twice as likely to have no postnatal check-up, and over five times more likely to have unqualified delivery assistance as those from towns. Rural women were 1.2 times more likely to have had no tetanus injection during the last pregnancy.
Household factors influencing non-use of maternal services
Women with higher socioeconomic status more often had resources and the ability to buy health. The majority of women who did not use maternal services were of low socioeconomic status. For instance, 23% of women with low socioeconomic status did not seek a postnatal check-up compared to 13% of women with a high socioeconomic profile (Table 1). Another notable observation was that 22% of women with low socioeconomic status had non-institutional delivery compared to only 2% of women in the high socioeconomic stratum.
Controlling for age, parity, education, marital status, and place of residence did not change the conclusions reached by descriptive statistics. Women with a low socioeconomic profile were four times less likely to have an institutional delivery compared to women from a high-socioeconomic group. The same conclusions could be drawn regarding having unqualified delivery assistance, tetanus injection, and receiving a postnatal check-up. There was, however, no difference in seeking an antenatal check-up between women coming from a low socioeconomic group compared to those coming from a high socioeconomic group.
This study investigated the individual and household factors that are associated with non-use of maternal services in Botswana. There is no doubt that the use of maternal health services improves reproductive health outcomes (11). The results of this study indicate that, although overall statistics showed a high percentage of women using maternal services, there were variations among different groups of women. Majelantle and Letamo found that some women preferred delivery at home and to be assisted during the delivery by traditional birth attendants because the traditional birth attendants were considered to be more compassionate and caring than modern healthcare providers (12). This, despite the fact that the training of traditional birth attendants alone, in the absence of back-up from a functioning referral system and support from professionally-trained health workers, is not effective in reducing maternal mortality (13). Various studies have shown that women who use antenatal care services during their pregnancies have more favourable pregnancy outcomes (2,11,14). However, some women still do not use these beneficial maternal services. More in-depth qualitative studies are needed to unravel the specific reasons for non-use of maternal services in those identified groups.
The current study has several major limitations. First, the study used secondary data that limited us from investigating the critical variables, such as distance from the nearest facility, attitude of healthcare providers towards clients, waiting time to receive assistance, and costs. These variables are thought to influence the decision to use maternal services or not. Second, since the data used were quantitative in nature, the authors are limited to providing insights into why some women prefer using traditional birth attendants rather than modern health facilities. These insights can only be obtained from qualitative approaches, such as focus-group discussions and in-depth interviews. It is, thus, important to conduct qualitative research to augment what is generated from the quantitative data.
The authors are grateful to the Central Statistics Office of Botswana for providing access to the Botswana Family Health Survey III. The authors would also like to thank anonymous reviewers for their valuable comments and some of our colleagues in the Department of Population Studies at the University of Botswana, who made valuable comments on the paper.
(1.) Rukaria KR, Cheserem E. Safe Motherhood Initiative in the African region in the health of women and safe motherhood: RH training materials (Popline abstract). Nairobi: Medical Women's International Association, 1993.
(2.) Letamo G, Majelantle RG. Health implications of early childbearing on pregnancy outcome in Botswana: insights from the institutional records. Soc Sci Med 2001;52:45-52.
(3.) Owuor-Omondi L, Kobue M. Determinants of maternal mortality in Botswana: an institutional, community and household perspective. Gaborone: Botswana Safe Motherhood Programme, 1993:1.
(4.) World Health Organization. Factors underlying the medical causes. (http://www.who.int/reproductive-health ... reduction_maternal_mortality_chap2.htm, accessed on 26 November 2001).
(5.) World Health Organization. World Health Day: safe motherhood. (www.who.int/archives/whday/en/ pages1998/whd98_01.html, accessed on 26 November 2001).
(6.) United Nations Population Fund. The state of world population, 2001: monitoring ICPD goals: selected indicators. (http://www.unfpa.org/swp/2001/english/ indicators/indicators1.html, accessed on 26 November 2001).
(7.) Botswana. Ministry of Finance and Development Planning. National Council on Population and Development. National population policy. Gaborone: Government Printer, 1997:26
(8.) DeMaris A. Logit modeling: practical applications. London: Sage, 1992. 87 p.
(9.) Agresti A, Finlay B. Statistical methods for the social sciences. 2d ed. San Francisco: Dellen, 1986. 556 p.
(10.) Hosmer DW, Jr., Lemeshow S. Applied logistic regression. New York: Wiley, 1989. 307 p.
(11.) Letamo G, Majelantle RG. Factors influencing low birth weight and pre-maturity in Botswana. J Biosoc Sci 2001:33:391-403.
(12.) Majelantle RG, Letamo G. The reproductive health problems of teenage childbearing in Botswana /edited by HM Yousif. Gaborone: National Council on Population and Development, Ministry of Finance and Development Planning, 1999. 36 p. (Research paper no. 1).
(13.) World Health Organization. Action for safe motherhood. (http://www.who.int/reproductivehealth ... reduction_maternal_mortality_chap3.htm, accessed on 26 November 2001).
(14.) Rahman MM, Barkat-e-Khuda, Kane TT, Mozumder KA, Reza MM. Determinants of antenatal care-seeking behaviour in rural Bangladesh. In: Kane TT, Barkat-e-Khuda, Phillips JF, editors. Reproductive health in rural Bangladesh: policy and programmatic implications, v. 1. Dhaka: International Centre for Diarrhoeal Disease Research, Bangladesh, 1997:86-104. (ICDDR,B monograph no. 7).
Gobopamang Letamo and Serai Daniel Rakgoasi
Department of Population Studies, University of Botswana, Private Bag 0022, Gaborone, Botswana
 Tetanus toxoid injections are given to women during pregnancy to protect infants from tetanus, a major cause of infant death due primarily to unsanitary conditions during childbirth. Two doses of tetanus toxoid during pregnancy offer full protection.
Correspondence and reprint requests should be addressed to:
Dr. Gobopamang Letamo
Department of Population Studies
University of Botswana
Private Bag 0022, Gaborone
Fax: (267) 585099
Table 1. Percentage of women who did not use specific maternal-child health services by selected individual and household characteristics Non-institutionalNo prenatal Characteristics birth check-up Age (years) <20 6.7 11.2 20-34 10.1 7.5 35+ 20.9 7.2 Parity 1 child 11.5 5.8 2-3 children 9.7 5.5 4 children 20.2 12.5 Education Never attended 36.1 11.1 Primary 15.6 10.7 Secondary+ 3.5 4.4 Marital status Ever-married 18.7 8.7 Never-married 11.4 7.2 Socioeconomic status High 2.4 4.9 Medium 5.4 7.8 Low 22.1 11.1 Residence Village-rural 19.6 11.5 Village-urban 9 9.9 Town 5.4 4.9 Unqualified No No postnatal Characteristics delivery tetanus check-up assistance injection Age (years) <20 18.8 14.2 29.7 20-34 12.3 9.7 16.6 35+ 17 11.4 16.1 Parity 1 child 12.5 9.8 17.8 2-3 children 13.2 9 16.4 4 children 17.2 13.1 17.2 Education Never attended 21 14.5 22.6 Primary 18.7 12.1 20.6 Secondary+ 9.1 8.4 13.2 Marital status Ever-married 18.8 11.7 14.7 Never-married 12.1 9.9 18.1 Socioeconomic status High 8.3 8.1 12.6 Medium 13.3 9 15.7 Low 19.8 15.9 22.6 Residence Village-rural 29.5 13.9 26.2 Village-urban 19.9 14.5 21.1 Town 5.9 6.5 11.8 Table 2. Relative odds that a woman would have a non-institutional delivery by selected individual and household characteristics (a): net effects model Odds Characteristics ratio Age (years) <20 0.439 (***) 20-34 1.147 (**) 35+ 1.000 Parity 1 child 5.344 (***) 2-3 children 2.535 (***) 4+ children 1.000 Education No formal education 10.579 (***) Primary 4.261 (***) Secondary+ 1.000 Marital status Ever-married 1.000 Never-married 0.520 (***) Socioeconomic status High 1.000 Medium 1.278 (**) Low 4.137 (***) Residence Village-rural 2.064 (***) Village-urban 1.031 Town 1.000 Age/parity interaction Term 1 0.619 (**) Term 2 0.305 (***) Term 3 0.354 (**) Characteristics 95% CI for exp (B) Lower Upper Age (years) <20 20-34 0.304 0.633 35+ 1.023 1.286 Parity 1 child 2-3 children 4.357 6.554 4+ children 2.215 2.902 Education No formal education Primary 9.348 11.973 Secondary+ 3.825 4.747 Marital status Ever-married Never-married Socioeconomic status 0.481 0.562 High Medium Low 1.058 1.543 Residence 3.450 4.961 Village-rural Village-urban 1.854 2.296 Town 0.896 1.187 Age/parity interaction Term 1 Term 2 0.398 0.961 Term 3 0.241 0.387 0.296 0.423 (a) The group with 1.000 is the reference category (b) Non-institutional delivery (births delivered outside the modern healthcare facilities) (***) Significant at p<0.01 (**) Significant at p<0.05 CI=Confidence interval Table 3. Relative odds that a woman would have no prenatal check-up by selected individual and household characteristics (a): net effects model Characteristics Odds ratio Age (years) <20 13.162 (***) 20-34 8.505 (***) 35+ 1.000 Parity 1 child 7.731 (***) 2-3 children 0.819 4+ children 1.000 Education No formal education 1.280 (**) Primary 1.550 (***) Secondary+ 1.000 Marital status Ever-married 1.000 Never-married 1.034 Socioeconomic status High 1.000 Medium 0.921 Low 1.011 Residence Village-rural 2.689 (***) Village-urban 1.844 (***) Town 1.000 Age/parity interaction Term 1 0.037 (***) Term 2 0.020 (***) Term 3 0.245 (***) Characteristics 95% CI for exp (B) Lower Upper Age (years) <20 8.548 20.266 20-34 6.820 10.605 35+ Parity 1 child 5.727 10.437 2-3 children 0.581 1.154 4+ children Education No formal education 1.037 1.582 Primary 1.349 1.781 Secondary+ Marital status Ever-married Never-married 0.905 1.182 Socioeconomic status High Medium 0.781 1.085 Low 0.836 1.221 Residence Village-rural 2.276 3.176 Village-urban 1.621 2.097 Town Age/parity interaction Term 1 0.022 0.062 Term 2 0.014 0.028 Term 3 0.169 0.356 (a) The group with 1.000 is the reference category (***) Significant at p<0.01 (**) Significant at p<0.05 CI=Confidence interval Table 4. Relative odds that a woman would have unqualified delivery assistance by selected individual and household characteristics (a): net effects model Characteristics Odds ratio Age (years) <20 10.880 (***) 20-34 2.638 (***) 35+ 1.000 Parity 1 child 1.849 (***) 2-3 children 2.323 (***) 4+ children 1.000 Education No formal education 2.384 (***) Primary 1.571 (***) Secondary+ 1.000 Marital status Ever-married 1.000 Never-married 0.617 (***) Socioeconomic status High 1.000 Medium 1.054 Low 1.314 (***) Residence Village-rural 5.211 (***) Village-urban 4.373 (***) Town 1.000 Age/parity interaction Term 1 0.112 (***) Term 2 0.283 (***) Term 3 0.186 (***) Characteristics 95% CI for exp (B) Lower Upper Age (years) <20 7.729 15.316 20-34 2.245 3.099 35+ Parity 1 child 1.403 2.438 2-3 children 1.935 2.788 4+ children Education No formal education 2.022 2.810 Primary 1.406 1.755 Secondary+ Marital status Ever-married Never-married 0.558 0.683 Socioeconomic status High Medium 0.926 1.199 Low 1.133 1.526 Residence Village-rural 4.540 5.982 Village-urban 3.907 4.896 Town Age/parity interaction Term 1 0.070 0.178 Term 2 0.207 0.386 Term 3 0.148 0.234 (a) The group with 1.000 is the reference category (c) Delivery by unqualified healthcare personnel (excludes medical doctors and trained nurses/midwives) (***) Significant at p<0.01 (**) Significant at p<0.05 CI=Confidence interval Table 5. Relative odds that a woman would not have a tetanus injection by selected individual and household characteristics (a): net effects model Odds Characteristics ratio Age (years) <20 0.014 (***) 20-34 1.772 (***) 35+ 1.000 Parity 1 child 6.299 (***) 2-3 children 0.729 (**) 4+ children 1.000 Education No formal education Primary 0.772 (**) Secondary+ 0.588 (***) Marital status Ever-married 1.000 Never-married 1.403 (***) Socioeconomic status High 1.000 Medium 0.989 Low 2.411 (***) Residence Village-rural 1.225 (**) Village-urban 2.014 (***) Town 1.000 Age/parity interaction Term 1 9.084 Term 2 0.051 (***) Term 3 0.704 (**) Characteristics 95% CI for exp (B) Lower Upper Age (years) <20 20-34 0.000 0.579 35+ 1.485 2.114 Parity 1 child 2-3 children 4.941 8.031 4+ children 0.576 0.922 Education No formal education Primary Secondary+ 0.648 0.919 Marital status 0.521 0.664 Ever-married Never-married Socioeconomic status 1.247 1.578 High Medium Low 0.864 1.133 Residence 2.074 2.803 Village-rural Village-urban 1.051 1.426 Town 1.807 2.245 Age/parity interaction Term 1 Term 2 0.220 374.600 Term 3 0.038 0.068 0.536 0.925 (a) The group with 1.000 is the reference category (***) Significant at p<0.01 (**) Significant at p<0.05 CI=Confidence interval Table 6. Relative odds that a woman would not have a postnatal check-up by selected individual and household characteristics (a): net effects model Characteristics Odds ratio Age (years) <20 6.360 (***) 20-34 1.374 (***) 35+ 1.000 Parity 1 child 3.607 (***) 2-3 children 1.507 (***) 4+ children 1.000 Education No formal education 1.590 (***) Primary 1.083 Secondary+ 1.000 Marital status Ever-married 1.000 Never-married 1.379 Socioeconomic status High 1.000 Medium 0.996 Low 1.368 (***) Residence Village-rural 1.977 (***) Village-urban 1.617 (***) Town 1.000 Age/parity interaction Term 1 0.064 (***) Term 2 0.199 (***) Term 3 0.575 (***) Characteristics 95% CI for exp (B) Lower Upper Age (years) <20 4.599 8.795 20-34 1.168 1.616 35+ Parity 1 child 2.884 4.511 2-3 children 1.269 1.790 4+ children Education No formal education 1.374 1.840 Primary 0.984 1.192 Secondary+ Marital status Ever-married Never-married 1.251 1.520 Socioeconomic status High Medium 0.894 1.111 Low 1.207 1.551 Residence Village-rural 1.758 2.223 Village-urban 1.479 1.768 Town Age/parity interaction Term 1 0.043 0.097 Term 2 0.152 0.259 Term 3 0.463 0.713 (a) The group with 1.000 is the reference category (***) Significant at p<0.01 (**) Significant at p<0.05 CI=Confidence interval
|Gale Copyright:||Copyright 2003 Gale, Cengage Learning. All rights reserved.|