Article Text

Original research
Impact of conditional cash transfer programmes on antenatal care service uptake in low and middle-income countries: a systematic review
  1. Ward Jacobs1,
  2. Laura Emily Downey1,2
  1. 1School of Public Health, Imperial College London, London, UK
  2. 2The George Institute for Global Health, Sydney, New South Wales, Australia
  1. Correspondence to Ward Jacobs; wardjacobs{at}icloud.com

Abstract

Objective Antenatal care (ANC) is crucial to protecting the health of pregnant women and their unborn children; however, the uptake of ANC among pregnant women in low and middle-income countries (LMICs) is suboptimal. One popular strategy to increase the uptake of health services, including ANC visits, are conditional cash transfer (CCT) programmes. CCT programmes require beneficiaries to comply with certain conditionalities in order to receive a financial sum. A systematic review was carried out to determine whether CCT programmes have a positive impact on ANC uptake in LMIC populations.

Methods Electronic databases CENTRAL, MEDLINE, Embase, Maternity and Infant Care and Global Health were searched from database inception to 21 January 2022. Reference checking and grey literature searches were also applied. Eligible study designs were randomised controlled trials, controlled before-after studies and interrupted time series analysis. Risk of bias assessments were undertaken for each study by applying the Risk of Bias 2 tool and the Risk of Bias in Non-randomised Studies of Interventions tool.

Results Out of 1534 screened articles, 18 publications were included for analysis. Eight studies reported statistically non-significant results on all reported outcomes. Seven studies demonstrated statistically significant positive effects ranging from 5.5% to 45% increase in ANC service uptake. A further three studies reported small but statistically significant impact of CCT on the use of ANC services in both positive (2.5% increase) and negative (3.7% decrease) directions. Subanalysis of results disaggregated by socioeconomic status (SES) indicated that ANC attendance may be more markedly improved by CCT programmes in low SES populations; however, results were inconclusive.

Conclusion Our evidence synthesis presented here demonstrated a highly heterogeneous evidence base pertaining to the impact of CCTs on ANC attendance. More high-powered studies are required to elucidate the true impact of CCT programmes on ANC uptake, with particular focus on the barriers and enablers of such programmes in achieving intended outcomes.

  • Public health
  • Health economics
  • Health policy

Data availability statement

This study is a systematic review. All included studies can be retrieved through the reference list. More information regarding the review process including title and abstract screening can be obtained by contacting the corresponding author.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

  • To the best of our knowledge, this is the most comprehensive systematic review and synthesis of published evidence on the impact of conditional cash transfer (CCT) programmes on antenatal care (ANC) uptake in low and middle-income country (LMIC) populations to date.

  • Evidence from 18 studies conducted in Africa, Asia and Central and South America was included in this study, representing a diverse sample of LMIC populations.

  • Heterogeneity in study design and implementation prevented a meta-analysis from being conducted to generate macro-impact statistics.

  • The descriptive nature of this study precludes conclusions regarding the causality between CCT programme implementation and ANC attendance.

Introduction

Reduction in maternal mortality is a global commitment outlined by the United Nations in the 2030 Sustainable Development Goals (3.1).1 Despite widespread recognition of the importance of antenatal care (ANC) in reducing maternal mortality2 and enhancing maternal and neonatal health outcomes,3 ANC service uptake remains low in many low and middle-income countries (LMICs).4 WHO recommends that women attend at least eight ANC visits5 during their pregnancy. A substantial proportion of women living in LMICs do not meet this recommendation, and ANC attendance appears to be highly correlated with socioeconomic status (SES) and poverty, reinforcing the notion that the social determinants of health are a strong driving force in influencing health status well before one is even born.6

Numerous reviews have been published that report the effects of demand-side interventions on health service uptake, including ANC attendance.7–10 Cash transfer programmes are one such intervention, and can be an attractive policy lever for increasing positive health-seeking behaviours in certain populations. Cash transfer programmes can be conditional or unconditional. Conditional cash transfer (CCT) programmes require beneficiaries to comply with certain conditionalities (eg, regular health check-ups) while unconditional cash transfer programmes do not set such requirements.11 Substantial resources have been allocated to cash transfer programmes in recent years, with an estimated 718 million people receiving assistance through cash transfer programmes in 2014 alone.12

CCTs may be a viable policy strategy to increase ANC uptake among pregnant women in LMICs. Evidence from several studies on the effectiveness of CCT programmes to increase health-seeking behaviours has shown promising positive results.11 13 However, a recent systematic review drew attention to the heterogenous impacts of cash transfer programmes across a range of health behaviours and outcomes, highlighting the need for further research into the key contexts in which such programmes may lead to success, and the barriers, enablers and opportunities for such programmes to thrive.14

Given the well-established correlation between ANC uptake and improved maternal and neonatal health,2 and the low reported rates of ANC attendance across numerous LMIC settings,4 there is an urgent need for bilateral and multilateral agencies and governments to invest in cost-effective interventions to increase ANC uptake. There is insufficient high-quality consistent evidence to elucidate whether CCTs are one such potentially viable intervention. This review aims to address this important knowledge gap and has two primary objectives: to assess the effectiveness of CCT programmes in improving ANC uptake; and to investigate the impact of poverty in relation to ANC attendance.

Methods

Study design

A systematic review was undertaken, adhering to the guidelines from the Cochrane Handbook for Systematic Reviews of Interventions.15

Eligibility criteria

Eligibility of each article was assessed according to the inclusion and exclusion criteria presented in table 1.

Table 1

Overview of inclusion and exclusion criteria

Participants

Pregnant women and girls residing in LMICs, defined as per World Bank definition, are eligible. Studies focusing on facilities or geographical areas that include service utilisation data were included. All types of healthcare providers were eligible for inclusion.

Intervention

Studies on CCT programmes were considered for inclusion if these constituted direct monetary transfers for the purpose of increasing health service uptake. Studies on unconditional cash transfers and non-cash transfers (eg, vouchers) were excluded. Interventions encompassing multiple components (with CCTs among them) were included, where it was possible to disaggregate cash transfer impacts from other intervention impacts.

Comparator

This review compares pregnant women and girls who took part in CCT programmes against those who did not.

Outcome

The sole outcome of this review is ANC service uptake. ANC utilisation was measured by health facility utilisation data, health service provision data and quantitative survey data.

Time period

We searched for evidence from database inception to 21 January 2022.

Study type

Study designs aligning with the Cochrane Effective Practice and Organisation of Care (EPOC) group criteria were included in this review.16 These encompass:

  • Randomised controlled trials (individual or cluster).

  • Controlled before-after studies, with data for the period before and after the intervention.

  • Interrupted time series analysis, with a clear time indication for the intervention and at least three data points before the intervention, and three data points after the intervention.

Systematic reviews were excluded during the screening process, but their reference lists were checked to possibly identify relevant literature.15

Data availability

In line with the EPOC criteria, studies with incomplete or opaque data were not incorporated in the final selection.16 A good example are studies with missing control variables. Authors were contacted for further inquiry as well. Studies with self-reported data are considered, contrary to the EPOC criteria, as filtering out articles reporting on survey-related data obtained by interviewing people would result in little evidence.

Identification of studies

A search was performed on 21 January 2022 using a sensitive search strategy (see online supplemental appendix A) in the following electronic databases: CENTRAL,17 MEDLINE,18 Embase,19 Maternity and Infant Care20 and Global Health.21 The search results were uploaded to Covidence,22 an online tool to support the selection process. Duplicates were automatically removed by the software and manually checked. Title and abstract screening was undertaken by a single reviewer (WJ) for all records, and a random sample of 20% of identified studies was reviewed by a second reviewer (LD) for quality assurance. Full-text review was undertaken by a single reviewer (WJ) and all records for which there was uncertainty were reviewed by a second author (LD) for final decision regarding inclusion/exclusion.15

Reference searching of included studies and follow-up with authors was carried out by a single reviewer (WJ) to ensure that all relevant articles and data were identified.15 Grey literature was also searched by the primary reviewer.15 The organisations identified for the grey literature search were identified by both reviewers and are listed in online supplemental appendix B.

Data extraction

A standardised Microsoft Excel form was used to assist with qualitative data extraction.15 The obtained information from the various studies contains:

  • Study type (individual or cluster randomised controlled trial, controlled before-after studies and interrupted time series analysis).

  • Study duration.

  • Study setting.

  • Characteristics of participants.

  • Characteristics of the intervention (transfer amounts and conditionalities).

  • Main outcome measures and results.

After extraction, the data were cross-checked against the original studies to avoid human error.23 Authors were contacted in case of data ambiguity.15

Inflation adjustment

Cash transfers were adjusted for inflation by presenting their value for the year 2022. This is to allow comparability across CCT programmes.24

Data analysis

The information extracted from the included studies was analysed by using descriptive thematic analysis.15 The analysis included overall effects demonstrated by the studies with further subanalysis on poverty dynamics.

Risk of bias

The Risk of Bias 2 tool recommended by the Cochrane Collaboration was used to assess the risk of bias for the included randomised controlled trials. The tool describes five domains clarifying the risk of bias by trial. These domains include the randomisation process, deviations from intended interventions, missing outcome data, measurement of the outcome and the selection of the reported result. The Risk of Bias in Non-randomised Studies of Interventions tool was used to assess the risk of bias for the included controlled before-after studies and research applying interrupted time series analysis. This tool uses domains and signalling questions that are tailored to non-randomised study designs, which encompass bias related to confounding, bias due to selection of study participants, bias in classification of interventions, deviations from intended interventions, bias due to missing data, bias in measurement of outcomes and bias in selection of the reported result.15

Patient and public involvement

Patient and public involvement is not applicable as this article is a systematic review of existing evidence. The research question development was informed by the global debate on the effectiveness of CCT programmes.

Results

Search results

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for conducting and reporting systematic reviews were followed.25 The PRISMA flow diagram is presented in figure 1.

Figure 1

Overview of the study selection process25

The database search yielded 2803 records. A total of 1534 records remained for title and abstract screening after duplicate studies were removed. These included three duplicates which were removed by Covidence software but added again to the title and abstract screening pool as abstracts were different. Out of the 1534 records, 308 were short-listed for full-text review against the eligibility criteria.

Eighteen studies were included, of which two were identified through other methods. Triyana was identified by contacting the author after requesting for more information on an excluded study.26 Barber and Gertler was included after a reference check of one of the included studies.27

Included studies

Of the 18 included studies, two were interrupted time series analysis, 10 were controlled before-after studies and the remaining six were randomised controlled trials. Barber and Gertler was the final study out of three reporting against the same randomised controlled trial of the Oportunidades programme.27 The article was selected as it was the most recent publication and covered all the necessary information as per EPOC requirements.16 Another author published two articles28 29 on the same randomised controlled trial. The first publication was selected for inclusion.29

The studies in table 2 are included in this review.

Table 2

Included studies

Included CCT programmes

The selected studies cover 13 CCT programmes presented in table 3. See online supplemental appendix C for more information on the monetary benefits.

Table 3

Conditional cash transfer programmes covered by the included studies

Risk of bias in the included studies

Randomised controlled trials

Among the six included randomised controlled trials, only Vanhuyse et al30 stated if the reported result was in line with a predetermined set of outcome indicators. Okeke and Abubakar,29 Grépin et al31 and Vanhuyse et al30 were rated as having a high risk of bias on randomisation, as each study failed to conceal the allocation sequence until study participants were enrolled and assigned to the CCT or control group (see online supplemental appendix D for comprehensive risk of bias assessment of each study).

Controlled before-after studies and interrupted time series analysis

Of the 12 included non-randomised studies, Joshi and Sivaram32 and Okoli et al33 indicated that reported results were in line with a research protocol. Almost all studies reported difficulties regarding accurate measurement of outcomes as participants were aware of the cash transfers provided to them. Factors lowering this risk were poorly documented in the studies. Edmond et al34 and Okoli et al33 were rated as having a serious risk of bias related to confounding (see online supplemental appendix D).

Effect estimates

The reported effect estimates of CCT programmes on ANC service uptake are presented in table 4.

Table 4

Treatment effects of included studies

Eight studies26 27 31 32 35–38 presented statistically non-significant results on all reported outcomes. Seven studies29 30 34 39–42 reported a statistically significant increase of over 5% in ANC service uptake. Three studies33 43 44 reported limited or negative effects.

A meta-analysis was not performed due to the heterogeneity of the selected studies. There are notable differences regarding the interventions, including the cash amounts and conditionalities. There is also variation in study settings, study population, study methodologies and data reported.15

Poverty dynamics

Out of the 18 included studies in this review, four controlled before-after studies contained in-depth poverty-related information.32 34 39 40 Studies were included if treatment effects could be retrieved for groups with different SES. Studies used different definitions for poverty, thereby impeding potential comparisons across settings. The treatment effects by population group are displayed in table 5.

Table 5

Poverty-related treatment effects from included studies containing information on poverty

Of the four studies that reported on treatment effect disaggregated by SES, two34 39 reported significantly higher ANC attendance in lower SES groups compared with control populations than did higher SES groups. The remaining two studies32 40 did not report statistically significant results in relation to this outcome.

Discussion

There is a pressing need across LMICs to increase the proportion of women who attend ANC, as recommended by WHO, in order to reduce maternal mortality and poor neonatal health outcomes.2 5 CCT programmes are a potentially promising policy lever to increase uptake of ANC across LMIC contexts; however, current evidence for the impact of CCTs on ANC is unclear. In this review, we have built on the evidence generated by previous published reviews7–10 of demand-side interventions on ANC uptake, to elucidate the specific impact of CCTs on this outcome of interest. Our findings are generally consistent with the existing evidence base that indicates that some CCT programmes have a modest positive impact on ANC attendance, but that other programmes fail to generate such impact, indicating high context specificity of such programmes in relation to ANC service uptake.

Of the 18 studies reviewed covering 13 CCT programmes, eight26 27 31 32 35–38 presented statistically non-significant results on all reported treatment effects, three33 43 44 demonstrated statistically significant limited or negative effects on the utilisation of ANC services and seven29 30 34 39–42 demonstrated a statistically significant increase in ANC service uptake ranging from 5.5% to 45%. The studies that did report statistically significant improvement in ANC uptake as a result of CCT programmes were delivered in Peru,41 Nigeria,29 Afghanistan,34 India39 40 42 and Kenya,30 where programme settings and modalities vary greatly. The studies that reported small or negative impacts of CCTs on ANC uptake were delivered in India,44 Nepal43 and Nigeria.33 The fact that divergent associations between CCTs and ANC uptake were reported in programmes implemented in India and Nigeria, coupled with the general heterogeneity of programme impact across the studies reviewed, indicates that programme design and implementation context might be vital factors in determining programme success.

The amount of money transferred has been postulated to play a key role in incentivising behaviour, and may be an important factor in whether or not the CCT programmes included in this review observed a positive impact.45 The study of the ‘Mamata’ scheme in India39 reported a notable positive impact, which could relate to the relatively high transfer amounts (US$70 per pregnancy) provided to women. This positive relationship between transfer amount and positive trends in ANC uptake is also supported by findings from the ‘JUNTOS’ programme in Peru,41 which similarly transferred a relatively high monetary amount (US$343.5 per pregnancy) compared with other studies and reported a statistically significant positive programme impact. However, in this review, we also identified programmes in which CCT using relatively low transfer amounts also reported positive impacts of CCT on ANC uptake. The CCT programmes best illustrating the complex relationship between financial allocation and programme success are those implemented in Nigeria, in which the CCT programme29 reported better results than the Subsidy Reinvestment and Empowerment Programme/Maternal and Child Health (SURE-P/MCH) programme,33 despite it being implemented in the same country with a transfer amount that is more than double of the CCT programme.29

Previous studies have established that conditionalities are crucial for impact across a range of health-seeing behaviours46 and could play a key role in increasing ANC service uptake. The ‘Mamata’ scheme in India39 required incremental ANC attendance, while the Safe Motherhood Programme in India40 42 44 focused on an endpoint of facility-based deliveries, with the former generating more impact overall. The Afya Credits Incentive in Kenya,30 the CCT programme in Nigeria29 and the ‘JUNTOS’ programme in Peru,41 which reported positive impacts, similarly allocated financial payments to ANC attendance conditionality. However, this conditionality of ANC attendance was not uniformly associated with increased ANC uptake across all studies reviewed, for example, the SURE-P/MCH programme in Nigeria33 reported limited programme impact despite ANC conditionality.

The differences in treatment effects among studies examining the same CCT programme warrant further scrutiny. Three included studies40 42 44 reported statistically significant results on the Safe Motherhood Programme in India using different data to analyse programme impact. Reported increase in ANC uptake as a result of the same CCT programme ranged from 2.4%44 to 22.9%.40 Aizawa40 demonstrated the strongest association between CCT and ANC uptake and used data from the National Family Health Survey conducted in 2006 and 2016 comparing various Indian states. Lim et al42 presented a lower positive association (11.1%) and used data from the District-Level Household Survey from 2004 and 2009. Debnath44 reported the smallest impact and used the same survey data as Lim et al,42 but opted for a restricted sample excluding numerous districts in India. Such heterogeneity indicates the complexity of policy evaluation as different results are reported on the same CCT programme.

We found inconclusive results regarding the relationship between poverty and CCT programme impact. The four studies32 34 39 40 that reported comparisons between socioeconomic groups and the impact of CCT on ANC uptake lacked statistical power to formulate robust conclusions due to low-powered sample sizes. Hence, we failed to determine if the level of poverty among people receiving CCTs was an important factor for determining impact on ANC service uptake.

One limitation of the evidence incorporated in this review is the use of survey data by the majority of included studies, opening the potential for data bias. We also note the developments in data capture infrastructure, such as smartphones and tablets, that coincide with the decade covered by the included studies, and the potential impact that this had on later studies in terms of enhanced ability to accurately capture data. The included studies varied in quality, ranging from suboptimal study designs to high levels of bias. Three included randomised controlled trials reported high risk of bias on the randomisation process29–31 and two non-randomised studies presented a serious risk of bias on confounding.33 34 The heterogeneity of study design, population and implementation process among the 18 studies hindered us to perform a meta-analysis to generate overall treatment effects of CCTs on ANC. A number of studies did not clearly present the information required for the summary tables. For example, less than half of all studies reported the actual number of ANC visits attended by programme participant populations, rendering it impossible to compare ANC attendance against the WHO-recommended5 number of visits for the majority of included studies. Together, these factors may contribute to the inconclusiveness of results reported in this review.

Given the high heterogeneity identified in this review in relation to CCT impact on ANC uptake across LMICs, there is substantial scope for future research to explore the most important determinants for CCT programme success, failure and inconclusiveness. Complex process evaluations should be employed alongside the implementation of CCT programmes to elucidate the contextual factors that contribute to programme success, including population characteristics, geographical and environmental factors, conditionalities, cointerventions, baseline ANC service uptake and financial allocations attached to demand-side interventions. Study design is an additional important consideration for future CCT programmes, whereby more high-powered randomised controlled trials are required to strengthen the evidence base for whether such programmes are truly impactful from a health perspective.

Conclusion

This systematic review investigated the relationship between CCT programmes and ANC service uptake. These programmes are an alluring instrument for policy makers in LMICs to expand ANC coverage. Our review demonstrated divergent effects of CCTs among the included studies, indicating high context specificity for these programmes to achieve the desired impact of increased ANC service uptake. The global health community, most notably multilateral organisations and donor community, has invested substantially in CCTs during the past few decades. This review highlights that further high-quality, high-powered evidence is required in order to elucidate the true impact of CCT programmes on ANC uptake, with special focus on process evaluation of the barriers, enablers and opportunities for programmatic success.

Data availability statement

This study is a systematic review. All included studies can be retrieved through the reference list. More information regarding the review process including title and abstract screening can be obtained by contacting the corresponding author.

Ethics statements

Patient consent for publication

Ethics approval

This study is a systematic review of already published literature.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Twitter @Downey1L

  • Contributors WJ: project administration, research protocol, conceptualisation, title and abstract screening, data extraction, data analysis and synthesis, methodology, grey literature search, background reading, risk of bias assessment, drafting the first manuscript, editing and overall review. LD: research protocol, title and abstract screening, editing of the draft manuscript, overall review, provision of guidance and direction. LED acts as guarantor and is responsible for the overall content in this manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient and public involvement Patient and public involvement is not applicable as this article is a systematic review of existing evidence. The research question development was informed by the global debate on the effectiveness of CCT programmes.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.