Maternal and Child Health Journal

, Volume 20, Issue 1, pp 172–186 | Cite as

The Influence of Women’s Empowerment on Child Immunization Coverage in Low, Lower-Middle, and Upper-Middle Income Countries: A Systematic Review of the Literature

  • Sara Thorpe
  • Kristin VanderEnde
  • Courtney Peters
  • Lauren Bardin
  • Kathryn M. Yount
Article

Abstract

Objectives

An estimated 1.5 million children under five die annually from vaccine preventable diseases, and 17 % of these deaths can be averted with vaccination. Predictors of immunization coverage, such as maternal schooling, are well documented; yet, preventable under-five mortality persists. To understand these patterns, researchers are exploring the mother–child relationship through an empowerment framework. This systematic review assesses evidence of the relationship between women’s agency as a component of empowerment and vaccine completion among children <5 years in lower-income countries.

Methods

We searched in Socindex, Pubmed, Web of Science and Women’s Studies International for peer-reviewed articles focused on two measures of women’s agency—decision-making and freedom of movement—and child vaccination. Our initial search identified 406 articles and abstracts for screening; 12 studies met the inclusion and exclusion criteria.

Results

A majority (83 %) of studies revealed at least one positive association of measures for women’s agency with immunization coverage. These relationships varied by geographic location, and most studies focused on women’s decision making rather than freedom of movement. No included study came from Latin America or the Middle East.

Conclusions

Overall, women’s agency, typically measured by decision-making, was positively associated with the odds of complete childhood immunizations. Yet, the concept of agency was inconsistently defined and operationalized. Future research should address these inconsistencies and focus on under-represented geographic regions including Latin America and the Middle East.

Keywords

Child immunization Vaccine coverage Vaccine preventable diseases Women’s agency Women’s empowerment 

Significance

As researchers and practitioners explore the relationship between gender-based approaches to improving child health outcomes including, vaccine preventable disease, systematic evidence is required to identify feasible pathways of enhancing health.

Introduction

Despite the known cost effectiveness and health benefits of immunization for children under five, mortality and morbidity due to vaccine preventable diseases (VPDs) remains a concern [1]. As governments aim to reach Millennium Development Goal 4, to reduce the risk of child mortality by two-thirds, understanding why an estimated 19.3 million children under five remain unvaccinated is crucial, especially because vaccines can avert 2.5 million deaths in this age group [2, 3, 4]. Since the World Health Organization’s (WHO) Expanded Program on Immunization (EPI) began in 1974 to recommend routine vaccinations for tuberculosis, diphtheria, pertussis, tetanus, poliomyelitis, measles, and polio, a growing share of children have gained access to these life-saving vaccines [2, 3, 4]. The number of vaccine preventable deaths among children under five declined from approximately 9.6 million in 2000–7.6 million in 2010 [3]. Still, disparities in coverage, particularly in sub-Saharan Africa and South Asia, lead to persistent under five mortality and morbidity due to VPDs [1, 3].

National, household-level, and individual-level determinants may explain persistent gaps in vaccination coverage [3, 4, 5]. Lower-income countries are less likely than higher-income countries to have the political stability, infrastructure, and services required for effective vaccine delivery [3]. The focus on geographic differences and barriers is shifting. Rapid urbanization increases the number of individuals living in slum dwellings, which elevates a child’s risk of exposure to under-immunized populations, resulting in higher rates of under-five mortality [5].

Parental biases in the decision to vaccinate sons and daughters are reported in lower-income countries, including in India, Bangladesh, Nigeria and Ethiopia [6, 7, 8, 9]. Household characteristics, such as assets and expenditures, also are cited as predictors of child immunization, as are maternal characteristics like age, use of antenatal services, location of delivery, schooling, and marital status [5, 6, 10]. Mothers with more schooling, lower parity, and higher household wealth are more likely to have timely vaccinated children [1, 11, 12, 13]. Yet, many other attributes of the mother–child relationship remain unaddressed theoretically and empirically in this literature.

Increasingly, researchers are adopting gender-based approaches to understand health inequities, such as incomplete vaccination coverage in children [14, 15]. These researchers have focused on explanatory variables related to the underlying enabling resources of mothers, such as their schooling, employment, and income [16, 17]. Based on the findings, researchers have advised that empowering mothers is a means to increase vaccination coverage in children; yet, the empowerment pathways of the “maternal resource-child vaccination” relationship are under-studied, including pathways capturing a mother’s “agency” or capacity to influence and enact decisions that may enhance the vaccination coverage of children. To address this gap, this systematic review examines (1) how researchers have defined and measured women’s “agency,” (2) how such measures are used in studies of children’s vaccination coverage, and (3) empirical patterns of association. This review informs researchers and practitioners about the state of the evidence and next steps for research and practice regarding women’s empowerment as a potential pathway to enhance children’s vaccination coverage [18].

Conceptual Framework

Globally, the term women’s empowerment has been used interchangeably with terms like women’s status, gender equality, and women’s autonomy [17]. Following Kabeer [18], we define women’s empowerment as a “process by which those who have been denied the ability to make strategic life choices acquire such an ability” [17]. Our adapted framework describes women’s empowerment as the process by which women acquire enabling resources, exercise agency, and attain life achievements (Fig. 1) [17]. Agency, therefore, as a component of women’s empowerment, refers to a woman’s ability to state her goals and to act upon them with motivation and purpose [17]. Thus, because enabling resources like schooling are necessary but insufficient conditions for women to exercise agency, our literature review focuses on agency [17] to understand it as a potential pathway to improved children’s immunization status.
Fig. 1

Conceptual framework for women’s empowerment [18]

Methods

Search Strategy and Selection of Studies

In 2014, we searched Socindex, Pubmed, Web of Science, and Women’s Studies International databases with defined search terms (Table 1) and retrieved the titles and abstracts of peer-reviewed, quantitative studies published in English between January 1, 1970 and September 1, 2013. This search yielded 409 unique titles and/or abstracts that met the initial screening criteria (Fig. 2). One researcher (ST) screened these titles/abstracts, identifying 89 articles for full-text review and application of a priori inclusion and exclusion criteria (Table 2). Two researchers (ST and LB) pilot tested the inclusion and exclusion criteria on a subset of articles, independently read and reviewed the 89 articles, and resolved by consensus any discrepancies in opinion regarding the relevance of each article for review. Fourteen articles met the inclusion and exclusion criteria. Before data extraction, one researcher (ST) conducted a key author and reference list search, yielding another three articles to which the inclusion and exclusion criteria were applied and of which one article was retained, yielding a total of 15 articles. An adapted Cochrane Review data collection form was applied to capture general information on the study and detailed information on outcome and exposure variables [17]. During the data extraction, three articles were excluded based on the inclusion and exclusion criteria (Table 2), for a total of 12 studies included in the data analyses. This systematic review follows the PRISMA guidelines, including adherence to all items listed in the PRISMA checklist.
Table 1

Search terms for identifying the associations between women’s empowerment and childhood immunization

Women’s empowermenta

Search terms

Child immunization

Women’s agency OR

 

Child vaccine preventable disease OR

Women’s mobility OR

 

Child immunization coverage OR

Women’s empowerment OR

 

Child passive vaccination OR

Women’s autonomy OR

AND

Child immunization OR

Women’s decision making OR

 

Child routine vaccination OR

Women’s freedom of movement OR

 

Child immunization uptake

Gender equality OR

  

Women’s status

  
 

NOTb

 
  

HPV OR

  

HIV OR

  

Cervical cancer

aFollowing Kabeer’s framework [18], women’s empowerment includes resources, agency and achievements. Because much of the work on empowerment is conceptual, we included terms, such as autonomy, gender equality, and women’s status that have been used to describe analogous constructs [15]

bAs our focus was not on HPV, HIV, or cervical cancer, we excluded studies focused on these outcomes

Fig. 2

Article identification procedure

Table 2

Inclusion and exclusion criteria

Criteria

Included

Excluded

Rationale

Sampling method

Population-based

Clinic-based, convenience-based etc.

The study aims to understand population-level health outcomes

Analysis

At least bivariate

Anything less than bivariate

Included as a minimum in order to capture the various ways that empowerment and vaccinations are measured and operationalized

Date

January 1, 1970–September 1, 2013

Anything below or above range

This date range covers the period during which WHO established the expanded program on immunization (in 1974)

Geographic (based on World Bank definitions)

Low income; lower-middle income; upper-middle income

High-income

Focus of this review is on geographic locations that bear a the high burden of vaccine preventable diseases (Black 2003)

Outcome variable

Complete vaccination or at least one of the recommended vaccines (BCG, Hepatitis B, Polio, DTP, haemophilus influenza Type B, pneumococcal, rotavirus, measles, and rubella)

Influenza and/or HPV

This review followed the WHO guidelines for recommended vaccines for children under five

Exposure variable

Decision making and/or freedom of movement

Items that fall within the resource category according to the conceptual framework (Fig. 1)

The term empowerment serves as the umbrella term under which agency is conceptualized. We defined women’s decision making and freedom of movement as domains of agency as expressed in the conceptual framework (Fig. 1)

Language

English

All other languages unless translation was provided

Majority of research in this area published or translated into English; linguistic limitations of the authors

Peer reviewed

Peer reviewed

Non-peer reviewed

This criterion is reflective of the focus on the highest-quality research examining the association between women’s agency and child vaccination

Population of interest

Women with children less than the age of five

Women with adolescents, teenagers

This serves as the topic of interest for this review and captures the time period of interest between the exposure and outcome

Results

Characteristics of Included Studies

Of the included studies, all were published on or after 2005 and were based on data collected between 1991 and 2009 (Table 3). Most studies were secondary analyses of Demographic and Health Surveys (DHS) [9, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28] conducted in urban and rural settings in lower-middle income countries [9, 15, 19, 21, 23, 24, 25, 28] and representative of two major geographic regions: Africa [9, 23, 24, 27] and South Asia [15, 19, 20, 21, 25, 26, 28]. One African multi-country study was included [28]. India [15, 19, 21, 24, 25] and Nigeria [9, 12, 22] were the most frequently represented countries in the review. The sample sizes ranged from around 1000 to around 15,000 respondents. The majority of authors used multivariate (MV) logistic regression in their analyses [19, 20, 21, 26], with the remainder using multilevel (ML), stepwise (SW), or bivariate regression analysis. Only two studies included a theoretical framework that addressed both empowerment and immunization coverage. The most common age range for reporting immunization outcomes was 12–23 months [9, 21, 23, 24, 25, 28].
Table 3

Characteristics of included studies (N = 12)

Characteristic

%a

Author reference number

Year published

 1970–2000

0

N/A

 2000–2005

8

[15]

 2006–2010

25

[19, 22, 23]

 2011–September 2013

67

[9, 20, 21, 23, 25, 26, 27, 28]

World Bank classification

 Low income

17

[19, 27]

 Lower-middle income

83

[9, 15, 20, 21, 22, 23, 24, 25, 26, 28]

 Upper-middle income

0

N/A

Data source

 DHS

92

[9, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28]

 Human Development Profile Index

8

[15]

Empowerment framework

 Empowerment only

33

[19, 20, 23, 25]

 Immunization coverage only

8

[21]

 Empowerment and immunization coverage

17

[9, 27]

 No framework

42

[15, 22, 24, 26, 28]

Type of analysis

 Bivariate logistic regression

17

[21, 24]

 ML logistic regression

25

[9, 22, 28]

 MV logistic regression

42

[19, 20, 23, 26, 27]

 SW logistic regression

8

[25]

 HLM

8

[15]

Country

 Nigeria

25

[9, 22, 23]

 India

42

[15, 20, 21, 24, 25]

 Bangladesh

8

[19]

 Nepal

8

[27]

 Kenya

8

[26]

 Multiple countries

8

[28]

Temporality of reporting

 12–23 months

50

[9, 21, 23, 24, 27, 28]

 12–35 months

25

[15, 22, 26]

 12–59 months

17

[19, 25]

 Not reported

8

[20]

Study location

 Urban

8

[24]

 Rural

17

[15, 19]

 Urban and rural

67

[9, 20, 21, 23, 25, 26, 27, 28]

 Not indicated in study

8

[22]

Empowerment measure

 Decision-making (DM items used)

58

[9, 22, 23, 25, 26, 27, 28]

 Freedom of movement (FM items used)

17

[15, 19]

 Other (OT items used)

8

[24]

 Multiple domains (OT, MB and DM items used)

17

[20, 21]

Vaccine measure

 Full immunization (VAC items used)

83

[9, 15, 19, 20, 21, 23, 24, 25, 27, 28]

 Full immunization and individual immunizations (VAC and VACI items used)

8

[27]

 Individual immunization (VACI)

8

[22]

Vaccine measures follow WHO guidelines

 Yes

83

[9, 15, 19, 20, 21, 23, 24, 26, 27, 28]

 No

17

[22, 25]

Vaccine measure accounts for timing and spacing of vaccine in definition

 Yes

17

[9, 15]

 No

83

[19, 20, 21, 22, 23, 24, 25, 26, 27, 29]

Vaccine measurement tool

 Household survey ONLY

8

[20]

 Immunization card or mother’s recall

50

[9, 15, 22, 25, 26, 27]

 Does not report method

42

[19, 21, 23, 24, 28]

(Adjusted) association between measure of agency and all immunization measures (N = 10)

 Consistently positive relationship(s)

50

[9, 15, 19, 22, 25, 28]

 Inconsistent relationship (s)

17

[20, 23]

 No association

17

[26, 27]

(Unadjusted) association between measure of agency and all immunization measures (N = 2)

 Consistently positive relationship(s)

50

[24]b

 Inconsistent relationship(s)

0

N/A

 No association

50

[21]

aDue to rounding, percentages may not sum to 100 %

bDue to referencing of the exposure and/or outcome these are positive relationship(s)

Measurement of Agency

In general, the authors varied in their definition of agency. The most common terms were “autonomy” [9, 20, 21, 23, 24, 26] and “empowerment” [25, 27] instead of “agency,” as defined by Kabeer [18]. For clarity, we categorized the items used to measure agency into three domains: decision making (DM), freedom of movement (FM), and other (OT) (Table 4), with the latter category including items that did not reflect only the DM or FM categories. Most often, agency items pertained to: health decisions of the woman; financial control; permission to go to/visit health centers, friends and family; as well as both large and daily household purchases.
Table 4

Agency measures and response categories

Agency measurement items

Response itemsa

Decision making (DM)

DM 1. The main decision maker when the child is ill

RC 1. Dichotomous (Yes/No)

DM 2. Can you go to the local health center without seeking permission

RC 2. Index for contribution to household decision making (0–12)

DM 3. Who usually makes decisions about health care for you/decisions regarding healthcare

RC 3. Decisions alone or jointly = high No participation = low

DM 4. Decisions on major/large household purchases/goods

RC 4. Who makes the decision, Index ranging from (0–2)

DM 5. Decisions on daily household needs

RC 5. Whether mother has money for own use (reference: has money for own use)

DM 6. Decisions on visiting friends, family or relatives

RC 6. Index categorized into low and high autonomy

DM 7. 12 Indicator index of contribution to decision-making

RC 7. If woman made decisions or her opinion was included in decision in any one item = 1 otherwise coded as 0

DM 8. Decisions to go to a health facility

RC 8. Women that made all decisions either alone or jointly = high; not involved in all four items = low

DM 9. Involvement in/decision on daily and major household decisions

RC 9. Self; jointly; others

DM 10. Decisions on how to spend husbands money/what to do with husbands money

RC 10. Unrestricted or restricted

DM 11. Purchasing of daily household goods

 

DM 12. How to spend money in the household

 

DM 13. Decisions on contraception

 

Freedom of movement

FM 1. Any HH female mobility

 

FM 2. Allowed to go market, health facilities and outside the home/village/community

 

FM 3. Permission to go to the health center alone

 

Other

 

OT 1. Whether mother has money for own use or not

 

OT 2. Control over spending

 

aResponse categories reflect author ST’s categorization of the responses

Most authors used one of 13 DM items to measure that domain of women’s agency (Table 3) [9, 22, 23, 25, 26, 28]. Two items (DM2 and DM8) captured a woman’s ability to decide to go to a health facility, and one item (DM1) asked if the mother was “the main decision-maker when the child is ill” (Table 4). Otherwise, the remaining DM items pertained to more general decisions about household purchases, health, finances, and ability to travel to see friends, family, or relatives (Table 3). The most common DM items were DM4 regarding major/large household purchases/goods [9, 23, 27, 28], DM6 regarding visits to friends, family, or relatives [9, 20, 21, 22, 27, 28], and DM3 regarding healthcare in general or for her [20, 27, 28], either alone [20, 22] or in combination with other items in other domains [9, 23, 27, 28].

The authors rarely measured agency using only freedom-of-movement items (n = 2, Table 4) [15, 19]. The authors of the two studies that included freedom-of-movement items analyzed DHS data from India [15, 19]. As such, we cannot generalize findings for FM because too few studies systematically included FM studies. Otherwise, authors operationalized freedom of movement with items reflecting decisions about movement rather than the action itself [9, 22, 25, 28]. The remaining authors combined items across all categories [20, 21] or used only OT items [24].

Measurement of Immunization Coverage

Definitions of immunization coverage varied, even though 83 % of the authors indicated that they followed WHO guidelines at the time of the study for recommended vaccines for children under five (Table 3) [9, 15, 19, 20, 21, 23, 24, 26, 27]. The WHO recommends vaccines for tuberculosis, diphtheria, pertussis, tetanus, poliomyelitis, measles, and polio [3]. For clarity, we organized immunization measures into two categories: full immunization (VAC) and individual immunization (VACI). The authors often used the terms “full” and “complete” interchangeably to describe immunization, as defined by the WHO VAC Items 1–6 (Table 5). One author used DPT3 as a proxy for full immunization coverage [23]. We assume that VAC 6 (complete vaccination) incorporates the aforementioned recommended vaccines. The most common measures for full immunization included a range of immunizations covering DPT, polio, BCG and measles [9, 15, 19, 20, 23, 25, 26, 28]. The authors’ use of vaccine measures varied within countries and across studies that used the same DHS data sources [9, 21, 24, 26]. For example, among the five studies in India, all studies ostensibly used full immunization measures, but the specific immunizations included varied (Table 5) [15, 20, 21, 25, 26]. The most commonly measured vaccines in the VAC and VACI categories were DPT [9, 15, 19, 20, 22, 23, 25, 26, 27, 28] and measles [9, 15, 19, 20, 23, 24, 25, 26, 27, 28].
Table 5

Immunization measures and response categories

Measurement item

Response item

Full immunization measures

VAC 1. 3 doses of diphtheria, pertussis and tetanus (DPT), 1 dose each of BCG and measles vaccine before 12 months of age

RC 1. Dichotomous (Yes/No)

VAC 2. 1 dose of BCG, 3 doses of DPT,3 doses of Polio vaccine, and 1 dose of measles vaccine

RC 2. Dichotomous (Full = 1/otherwise = 0)

VAC 3. Children aged 12–23 months who received 1 dose each of BCG and measles, and 3 doses each of DPT and polio vaccine

RC 3. Dichotomous (1 dose = 1/otherwise = 0)

VAC 4. 3 doses of oral polio vaccine, 3 doses of diphtheria, pertussis, and tetanus, one dose each of BCG and measles vaccine before 12 months of age

RC 4. Dichotomous (3 doses = 1/otherwise = 0)

VAC 5. 1 BCG vaccine at birth, 3 doses each of DPT and oral polio at 6, 10, and 14 weeks of age, and finally, 1 measles vaccine at 9 months or soon thereafter

RC 5. Dichotomous (1 = received DPT3/otherwise = 0)

VAC 6. Complete Vaccination

RC 6. None = if the child did not receive any immunizations at the time of the survey; Some = if the child received at least one but not all eight immunizations; All = if the child received all eight immunizations −1 BCG, 3 DPT, 3 polio, 1 measles

Individual immunization

VACI 7. Measles immunization

 

VACI 8. Full series DPT3

 

VACI 9. DPT immunization

 

VACI 10. Polio immunization

 

The measurement of “full” or “complete” immunization, specifically for VAC items 1–6, also varied in their consideration of timing and spacing metrics, which are important factors for vaccine effectiveness [14]. A majority of the authors ostensibly met the WHO guidelines, but in only two studies using data from Nigeria and India, respectively [9, 15], were timing and spacing of the vaccines considered in their measurements, and only studies that used VAC1 and 3–5 included age matrices for receipt of immunizations, ranging from birth to 23 months (Table 3). The sources of data on immunization varied, with 50 % of authors reporting to use vaccines cards, and in their absence, maternal recall (Table 3).

Associations between Measures of Women’s Agency and Childhood Immunization

Overall Patterns

Among included studies, a majority (58 %) documented consistent relationships [9, 15, 19, 23, 25, 26, 27] and 42 % documented inconsistent or no relationship [20, 21, 23, 26, 27] between measures of agency and immunization coverage. However, among the studies with inconsistent findings, the authors typically found at least one positive relationship between a measure for maternal agency and immunization. So, overall, the majority of authors showed at least one positive relationship between a measure for women’s agency and child immunization (n = 10). Two of these authors found inconsistent relationships, one of which was not significant [20, 23]. Here, we define significance as P ≤ .05 and marginal significance as P ≤ .10. In general, the larger the sample (n > 3000), the more often an association was seen between at least one measure of agency, particularly decision making, and child immunization (n = 5) (Table 6) [15, 20, 23, 24, 26].
Table 6

The associations between women’s agency and child immunization (n = 12)

Article

Data source

Sample size

Agency definition

Agency measure and response items

Agency measurement instrument

Vaccine measure and response items

Temporality of vaccine reporting

Type of analysis

Outcomes

Summary of relationship

Decision making (DM 113)

22

Nigeria DHS (2003)

1472

Mother conjugal power

DM 6 RC2

Summative continuous index (0–12, vague)

VACI 8 RC 5

Any child born in past 5 years

MV logistic regression

Conjugal DM power range aOR 1.22*

Positive

20

Nigeria DHS (2008) 

3725

Decision-making autonomy

DM 10, 11, 6, 4; RC1

Dichotomous: If responded alone or with husband to one or several = yes if responded with other person to all = no

VAC 5 RC 1

All children born since 2003 (5 years)

ML logistic regression

Square score for conjugal DM power range aOR .97*

Positive

26

Kenya DHS (2003)

2169

Autonomy

DM 12 and 13 RC 4

Summative continuous index 0 = no autonomy 2 = high autonomy

VAC 2 RC 2

Women who gave birth in past 35 months

MV logistic regression

DM Autonomy aOR.76 SE (.11)

Consistent

23

Nigeria DHS (2008)

3250

Autonomy

DM 3, 10, 11, 4 RC 3

Categorical autonomy index

VAC 1 RC 1

Any child born in past 5 years

MV logistic regression

Household DM (ref: low) aOR 1.64 CI (1.25–2.14)*

Inconsistent

         

Financial DM (ref: low) aOR .98 CI (.76–1.27)

 

25

India Human Development Survey (2004–2005)

5287

Empowerment

DM 1 and 2 RC 1

Dichotomous

VAC 1 RC 1

Most recent birth

SW Logistic Regression

Visit Health Center aOR 1.25 SE (.103)**d

Positive

         

Decision making aOR 1.21 SE (.090)f**

 

27

Nepal DHS (2006)

1056

Empowerment

DM 3, 4, 5, 6 RC 7

Unidimensional dichotomous

VAC 2 RC 2

VACI7 RC3

First child born

MV Logistic Regression

Financial, Health, and Mobility Decision making

Inconsistent

      

VACI 9 RC 4

VAC I10 RC 4

  

All 8 Vaccines aOR .78

 
         

3 Doses DPT aOR .76

 
         

3 Doses Polio aOR .76

 
         

1 Dose Measles aOR .91

 

28

DHS; Democratic Republic of Congo (2008); Ghana (2008); Liberia (2007); Mali (2006); Nigeria (2009);Uganda (2006); Zambia (2007)

14,150 (Pooled Data)

Gender equality

DM 3, 11, 6, 4 RC 8

Decision making summative categorical household decision making (High/Low)

VAC 4 RC 1

Any child born in past 5 years

ML Logistic Regression

Household DM (ref: low) aOR 1.31 CI (.92, 1.87)

Consistent

Freedom of Movement (FM 13)

19

Bangladesh DHS (2004)

3530

Mobility characteristics

FM 3 RC 10

Unidimensional Dichotomous

VAC 2 RC 1

Last child born

MV Logistic Regression

Permission to go to Health Center Alone (ref: unrestricted) aOR .921 CI (.736–1.51)

Positive

15

India Human Development Profile Index (1994) and Indian Census (1991)

5623

Mobility of household women

FM 1 RC 1

Dichotomous freedom to move outside the home (no permission needed = 1; not allowed to go outside = 0)

VAC 5 RC 6

Most recent child who is alive

HLM

Any female HH mobility aOR .17 SE (.06)* e

Positive

Other (OT 12)

24

India DHS-NFHS-3 (2005–2006)

1527

Mother’s autonomy

OT 1 RC 1

Dichotomous: mother’s lack of autonomy reverse coding

VAC 6 RC 1

Women who gave birth in the past 5 years

Bivariate logistic regression

Mother’s autonomy (ref: has money for own use) OR .6267 (.4084–.9615) (ref: ever vaccinated)c

Positive

Combination DM, FM, and OT items

21

India DHS-NFHS-3 Survey (2005–2006)

1607

Autonomy

DM6,8,9,10; FM2 RC 6

Categorical autonomy index (high/low)

VAC 3 RC 1

Bivariate Analysis Most recent birth of women who had teen pregnancies

Autonomy: Chi-square (1.65) Low: 40.56 High: 43.81 b

Inconsistent

20

India DHS NFHS—3f (not reported)

Only entire NFHS-3 sample reported

Autonomy

OT 2 RC 9; DM 3 RC 9; DM 13 RC 9; DM 6 RC 9; DM 9 RC 9

Principal component analysis

VAC 2 RC 2

MV logistic regression Does not report

OT control over Spending (ref: Self) aOR Jointly .974, aOR.764 Others

Inconsistent

         

DM on own healthcare (ref: Self) aOR .795 Jointly, aOR 1.013 Others;

 
         

DM on large household purchases (ref: Self) aOR .887 Jointly, .754 Others

 
         

DM on daily household purchases (ref: Self) aOR .682 Jointly, aOR.972 Other

 
         

Mobility DM (ref: self) aOR1.695* Jointly, aOR 1.261 Others

 

NS not significant

aSame data source

bAll the variables identified as significant in the bivariate analyses using the Chi-square test were included in the binary logistic regression model

cPositive association due to reference coding even though alpha not reported

dTransformed beta coefficients provided by author

eNon-transformed

fDate not reported in study, secondary source indicates data was collected from 2005 to 2006

P < .01

** P < .0324

Associations with Decision-Making

The most commonly analyzed item was DM6 (n = 3, decisions on visiting friends, family or relatives) [9, 22, 28]. Of the seven [10, 16, 20, 23, 25, 26, 29] studies that reported consistent relationships between agency and immunization alone and not in combination with other exposure measures of women’s agency, most analyzed items related to DM (n = 4) [10, 23, 26, 29]. Of studies showing positive relationship(s) for at least one measure of agency and immunization, four were based in South Asia [15, 19, 24, 25] and three in Africa [9, 22, 28]. Four authors exclusively found a positive relationship between DM and immunization alone and not in combination with other measures of agency [9, 22, 25, 28]. In sum, the results suggest that low immunization often is associated with a lack of decision-making agency among mothers, mainly in India and Nigeria.

Two studies included reports on vaccine completion before age 9 months and found significant positive relationships with women’s agency [9, 15]. Five authors reported vaccination on any child born in the past 5 years and reported significant positive relationships between at least one measure of agency and immunization [9, 22, 23, 24, 25, 28]. Three of these studies were based in Nigeria [9, 22, 23].

The authors of the remaining studies, who found either inconsistent relationships or no association between DM and immunization outcomes, used different DM items than those reported above [20, 21, 23, 24, 27]. These authors included items pertaining to the woman’s role in household finances, which led to inconsistent findings or no association [20, 23, 27]. One Nepali study combined DM items related to finances, health, and freedom of movement and reported a negative relationship between agency and all 8 vaccines [27]; whereas, one Nigerian study found significant associations of items related to household DM (DM 3,4,11, Table 4) and vaccination but not financial DM (DM 10, decisions on how to spend husband’s money/what to do with husband’s money) and vaccination [23].

Associations with Freedom of Movement

The authors of all studies measuring FM, whether alone or with other agency exposures, did not find consistent relationships between FM and immunization. The relationships varied by the venue to which a woman traveled and the measurement scale for freedom of movement. Across the two studies whose authors analyzed freedom of movement alone, one found a significant relationship between “any HH female mobility” (FM1, Table 4) with immunization, and the other found a significant relationship with “permission to go to the health center alone” (FM3, Table 4) with immunization [15, 21]. One of the studies in India did not show a bivariate (unadjusted) association, so the authors did not apply a regression analysis [21]. This study characterized autonomy through freedom of movement, being “allowed to go to market, health facilities and outside the home/village/community” (FM 2, Table 4) and immunization, along with DM 6, 8 and 9 [21]. Two authors operationalized freedom of movement with items that other authors used to measure decision making and found a significant positive relationship between FM and immunization [20, 25]. In our analysis, these studies were grouped with those related to decision making. In general, the studies showing a positive relationship between freedom of movement and immunization were based in India [15, 20, 25]. However, only two of these studies actually found a positive relationship between DM and immunization due to the authors’ definition of freedom of movement. The inconsistent findings outside of India suggest a need for more research on the relevance of measures for maternal freedom-of-movement and child vaccination.

Findings by Measure of Immunization

Among the authors that reported consistently positive relationships between agency and immunization, five used full immunization items, VAC1 [24], VAC4 [27] VAC6 [24], and VAC5 [9, 15] and one used the full series of DPT3 VAC8 as a proxy for full immunization [22]. The only study that included multiple vaccine measurements found no association between agency and any immunization measure [28]. The authors that analyzed vaccine measures incorporating timing and spacing of immunizations found a positive relationship with agency in terms of either DM or FM [9, 15]. Thus, despite some inconsistencies within and across countries in definitions of, measurement of, and relationship with women’s agency, we generally found that women’s agency was positively associated with complete immunization of children in the countries represented.

Data Quality

In general, the authors of included articles adequately and appropriately reported their results. Often, the authors did not fully explain the study design and methods. Moreover, the authors often were vague and brief in their descriptions of the measures for women’s agency and the sources of data on immunization, which may affect the estimated associations between these measures. As a methodological strength, most included studies (n = 11) captured the temporality of the mother’s reporting on the immunization of her child, which helps to explain recall bias [9, 15, 19, 21, 22, 23, 24, 25, 26, 27, 28].

Discussion and Recommendations

This systematic review is the first to investigate the relationship between women’s agency and childhood immunization in lower-income settings. Based on the included studies, which represented countries in South Asia and Africa, we observed the general pattern that higher agency among mothers was associated with higher odds of childhood immunizations. This pattern of association was most apparent when women’s agency was measured by items capturing their ability to make decisions and immunization was measured by items reflecting complete immunization. This pattern corroborates one from eight African countries showing a significant, positive relationship between women’s decision-making agency and the adjusted odds that a child under two with acute respiratory infection visits a health facility (aOR 1.31 CI 1.12, 1.54) [29]. In general, fewer studies included measures for women’s freedom of movement, and when such measures were included, their associations with child immunization were inconsistent. Thus, our systematic review suggests, for lower-income settings, that specific dimensions of women’s agency may enhance vaccination coverage for children, and that empowering women in such settings shows promise as a means to improve child health.

Our review also highlights important limitations in the literature reviewed that inform key recommendations for future research. First, a majority of included studies analyzed DHS data, revealing a heavy reliance on this source for information on women’s empowerment. The DHS is considered a gold standard for cross-national comparison, yet measures of agency still differ in the DHS, both across countries and within countries over time [29]. Context-specific measures of agency have the advantage of capturing locally relevant domains and manifestations of agency [30], but they complicate our ability to generalize more broadly about the relationship of women’s agency to child immunizations, and ultimately child health. Efforts to develop scales for women’s agency that include comparable and context-specific items in multiple domains would advance both comparative and context-specific studies of these relationships.

Second, the limited use of measures to capture women’s freedom of movement as a domain of their agency also informs our recommendations. The authors who measured FM alone showed a positive relationship [15, 20], and those who measured women’s freedom of movement as a component of decision making found positive relationships [20, 25]. While both sets of findings are suggestive, the studies had limited geographic scope, cautioning against broader generalizations at this time. Contemporary research on women’s agency suggests that women’s freedom of movement remains a salient dimension of agency in particular contexts, such as the Middle East [30] and South Asia [31]. In such settings, researchers should systematically measure and include women’s freedom of movement as one of multiple domains of agency in studies of child health. Such efforts may mitigate potential bias in the estimated associations of women’s decision-making with child immunization. In addition, these efforts would improve an understanding of the pathways by which women’s enabling resources may enhance child health through the multiple domains of women’s agency. Successful efforts like those outlined above would require clarity regarding the definition of freedom of movement, its comparable and context-specific component items, and scales derived from these items.

A third limitation, which highlights a key area for research, involves the content areas of DM and FM items. In theory [17], the concept of women’s agency involves decisions in an array of family domains, including those historically reserved for men, and also involves the freedom to visit or travel to a range of public venues. Yet, in the studies included in our review, only one DM item captured decisions specifically related to children’s health, which may be more strongly associated with the outcome of interest. Indeed, the authors of one study in India included “The main decision maker when the child is ill” in there measure of DM1 and found a significantly positive relationship between decision making and immunization (Table 4) [25]. Moreover, the DM items used to measure women’s agency in this review captured decisions that historically have been relegated to women, such as cooking and small household purchases. Such items may not reflect agency, as Kabeer has defined, in that they do not capture decisions historically reserved for men. Thus, measures of women’s agency that may be more relevant for children’s health outcomes may include women’s capacity to make health-related and major financial decisions and to travel unaccompanied to associated venues, such as clinics, hospitals, pharmacies, and local markets for food.

The measurement of vaccination coverage is a fourth limitation. In half of the studies, immunizations were measured first by vaccine cards, and in the absence of a complete vaccine card, mother’s recall, which sometimes spanned 5 years. A heavy reliance on maternal reports of child vaccinations over several years may lead to systematic error in the reported occurrence and timing of vaccinations. The two studies showing statistically significant relationships between included measures of agency and immunization before the age of 9-month could reveal the utility of shorter recall windows in limiting recall bias [9, 15]. Only two authors measured the timing and frequency of vaccinations [9, 15], which is important because as evidence suggests, repeated vaccinations decrease the effectiveness of vaccines [9, 15, 28]. Improvements in such data may better inform global strategies and efforts that aim to address gaps in coverage such as the Global Vaccine Action Plan and EPI [2, 3, 4].

A fifth limitation concerns the cross-sectional design of most of the included studies. In such studies, agency is measured with respect to the date of interview; whereas, information on vaccinations refers to events that occurred some months or years before the interview. Thus, the estimates from such studies are purely correlational, and appropriate temporal ordering is needed to establish the causal direction from women’s agency to child immunization. Thus, longitudinal studies—ideally embedded in randomized women’s empowerment interventions—are needed to test the “impact” of increases in women’s agency on child vaccination coverage.

Finally, because all included studies in this review came from South Asian and African countries, future research should include under-researched settings where vaccination in childhood remains problematically low. This recommendation is especially important, as 70 % of the 22.3 million children who did not receive DTP3 in 2011 lived in the Democratic Republic of Congo, Ethiopia, India, Indonesia, Iraq, Nigeria, Pakistan, Philippines, Uganda, and South Africa at the time of survey, and only half (n = 6) of these countries were represented in this review [1].

In sum, this review assessed the strength of existing evidence on the association of two dimensions of women’s agency—freedom of movement and decision making—with child immunization. With growing interest among funders—such as the Gates Foundation and UK Department for International Development—on the impact of women’s empowerment on maternal and child health, this review is timely and offers a baseline of evidence on which researchers may build to improve an understanding of these relationships. Our findings suggest that women’s capacity to make family decisions is positively related to child immunization in Nigeria and India, the most populous countries in Africa and South Asia, respectively, and where child immunization remains low. Longitudinal research across diverse contexts would confirm the external validity of our findings. As countries strive to reduce under-five mortality and morbidity and to meet MDG 4, our findings suggest that programs to empower women should not be ruled out as a strategy to improve child health.

Notes

Acknowledgments

This paper was developed in collaboration with a group of mentors and colleagues who the principal researcher would like to acknowledge and thank for their contributions. The authors acknowledge Emory University, Rollins School of Public Health, and the Hubert Department of Global Health for its continuous support in the advancement of global health research and public health professionals. The views expressed in this article do not necessarily reflect those of Emory University Rollins School of Public Health.

References

  1. 1.
    Harris, J., Gacic-Dobo, M., Eggers, R., Brown, D., & Sodha, S. (2014). Global immunization data (pp. 1–4). Geneva: WHO.Google Scholar
  2. 2.
    Harahap, J. (2000). Factors affecting childhood immunization in North Sumatra Province, Indonesia (pp. 1–63). Master of Arts Thesis, Mahidol University.Google Scholar
  3. 3.
    World Health Organization (WHO). (2011). Global Vaccine Action Plan 20112020 (NLM classification: WA 115). Geneva: WHO Press. http://cdrwww.who.int/immunization/global_vaccine_action_plan/GVAP_doc_2011_2020/en/.
  4. 4.
    Burton, A., Monasch, R., Lautenbach, B., Gacic-Dobo, M., et al. (2009). WHO and UNICEF estimates of national infant immunization coverage: Methods and processes. Bulletin of the World Health Organization, 87, 535–541.PubMedPubMedCentralCrossRefGoogle Scholar
  5. 5.
    Kamau, N., & Esamai, F. O. (2001). Determinants of immunization coverage among children in Mathare Valley, Nairobi. East African Medical Journal, 78(11), 590–594.PubMedCrossRefGoogle Scholar
  6. 6.
    Mutua, M. K., Kimani-Murage, E., & Ettarh, R. R. (2011). Childhood vaccination in informal urban settlements in Nairobi, Kenya: Who gets vaccinated? BMC Public Health, 11, 6.PubMedPubMedCentralCrossRefGoogle Scholar
  7. 7.
    Corsi, D. J., Bassani, D. G., Kumar, R., Awasthi, S., et al. (2009). Gender inequity and age-appropriate immunization coverage in India from 1992 to 2006. BMC International Health and Human Rights, 9(1), S3.PubMedPubMedCentralCrossRefGoogle Scholar
  8. 8.
    Chowdhury, A. M., Bhuiya, A., Mahmud, S., Abdus Salam, A. K., et al. (2003). Immunization divide: Who do get vaccinated in Bangladesh? Journal of Health, Population and Nutrition, 21(3), 193–204.Google Scholar
  9. 9.
    Antai, D. (2012). Gender inequities, relationship power, and childhood immunization uptake in Nigeria: A population-based cross-sectional study. International Journal of Infectious Diseases, 16(2), e136–e145.PubMedCrossRefGoogle Scholar
  10. 10.
    Fatiregun, A. A., & Okoro, A. O. (2012). Maternal determinants of complete child immunization among children aged 12–23 months in a southern district of Nigeria. Vaccine, 30(4), 730–736.PubMedCrossRefGoogle Scholar
  11. 11.
    Babirye, J. N., Rutebemberwa, E., Kiguli, J., Wamani, H., et al. (2011). More support for mothers: A qualitative study on factors affecting immunisation behaviour in Kampala, Uganda. BMC Public Health, 11, 723.PubMedPubMedCentralCrossRefGoogle Scholar
  12. 12.
    Fadnes, L. T., Nankabirwa, V., Sommerfelt, H., Tylleskär, T., et al. (2011). Is vaccination coverage a good indicator of age-appropriate vaccination? A prospective study from Uganda. Vaccine, 29(19), 3564–3570.PubMedCrossRefGoogle Scholar
  13. 13.
    Pande, R. P., & Yazbeck, A. S. (2003). What’s in a country average? Wealth, gender, and regional inequalities in immunization in India. Social Science and Medicine, 57(11), 2075–2088.PubMedCrossRefGoogle Scholar
  14. 14.
    Jamil, K., Buhiya, A., Streatfield, K., & Chakrabarty, N. (2005). The immunization program in Bangladesh: Impressive gains in coverage, but gaps remain. Health Policy and Planning, 1(14), 49–58.Google Scholar
  15. 15.
    Parashar, S. (2005). Moving beyond the mother–child dyad: Women’s education, child immunization, and the importance of context in rural India. Social Science and Medicine, 61(5), 989–1000.PubMedCrossRefGoogle Scholar
  16. 16.
    Topuzoğlu, A., Ay, P., Hidiroglu, S., & Gurbuz, Y. (2007). The barriers against childhood immunizations: A qualitative research among socio-economically disadvantaged mothers. European Journal of Public Health, 17(4), 348–352.PubMedCrossRefGoogle Scholar
  17. 17.
    Kabeer, N. (1999). Resources, agency, achievements: Reflections on the measurement of women’s empowerment. Development and Change, 20(3), 435–464.CrossRefGoogle Scholar
  18. 18.
    Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Medicine, 6(6), e1000097. doi:10.1371/journal.pmed1000097.Google Scholar
  19. 19.
    Rahman, M., & Obaida-Nasrin, S. (2010). Factors affecting acceptance of complete immunization coverage of children under five years in rural Bangladesh. Salud Publica de Mexico, 52(2), 134–140.PubMedCrossRefGoogle Scholar
  20. 20.
    Mahapatro, S. R. (2012). Utilization of Maternal and child health care services in India: Does women’s autonomy matter? The Journal of Family Welfare, 58(1), 22–33.Google Scholar
  21. 21.
    Singh, L., Rai, R. K., & Singh, P. K. (2011). Assessing the utilization of maternal and child health care among married adolescent women: Evidence from India. Journal of Biosocial Science, 44(1), 1–26.PubMedCrossRefGoogle Scholar
  22. 22.
    Babalola, S. (2009). Determinants of the uptake of the full dose of diphtheria–pertussis–tetanus vaccines (DPT3) in Northern Nigeria: A multilevel analysis. Maternal and Child Health Journal, 13(4), 550–558.PubMedCrossRefGoogle Scholar
  23. 23.
    Singh, K., Haney, E., & Olorunsaiye, C. (2012). Maternal autonomy and attitudes towards gender norms: Associations with Childhood immunization in Nigeria. Maternal and Child Health Journal, 17(5), 837–841.CrossRefGoogle Scholar
  24. 24.
    Agarwal, S., & Srivastava, A. (2009). Social determinants of children’s health in urban areas in India. Journal of Health Care for the Poor and Underserved, 20(4 Suppl), 68–89.PubMedCrossRefGoogle Scholar
  25. 25.
    Vikram, K., Vanneman, R., & Desai, S. (2012). Linkages between maternal education and childhood immunization in India. Social Science and Medicine, 75(2), 331–339.PubMedPubMedCentralCrossRefGoogle Scholar
  26. 26.
    Abuya, B. A., Onsomu, E. O., Kimani, J. K., & Moore, D. (2011). Influence of maternal education on child immunization and stunting in Kenya. Maternal and Child Health Journal, 15(8), 1389–1399.PubMedCrossRefGoogle Scholar
  27. 27.
    Pandey, S., & Lee, H. N. (2012). Determinants of child immunization in Nepal: The role of women’s empowerment. Health Education Journal, 71(6), 642–653.CrossRefGoogle Scholar
  28. 28.
    Singh, K., Bloom, S., & Brodish, P. (2013). Gender equality as a means to improve maternal and child health in Africa. Health Care for Women International, 36(1), 57–69.PubMedPubMedCentralCrossRefGoogle Scholar
  29. 29.
    Upadhyay, U. D. & Karasek, D. (2010). Women‘s empowerment and achievement of desired fertility in Sub-Saharan Africa (DHS Working Papers No. 80). Calverton, Maryland, USA: ICF Macro.Google Scholar
  30. 30.
    Yount, K. M., VanderEnde, K., Dodell, S., & Cheong, Y. F. (2015). Measurement of women’s agency in Egypt: A National Validation Study. Social Indicators Research. doi:10.1007/s11205-015-1074-7.Google Scholar
  31. 31.
    Agarwala, Rina, & Lynch, Scott M. (2006). Refining the measurement of women’s autonomy: An international application of a multi-dimensional construct. Social Forces, 84(4), 2077–2098.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2015

Authors and Affiliations

  • Sara Thorpe
    • 1
  • Kristin VanderEnde
    • 1
  • Courtney Peters
    • 1
  • Lauren Bardin
    • 1
  • Kathryn M. Yount
    • 1
  1. 1.Emory UniversityAtlantaUSA

Personalised recommendations