Incentive-based interventions, including Conditional Cash Transfers (CCTs) and vouchers, have been implemented globally with some success in various maternal and neonatal health programs. ^{11 , 12} These initiatives have been implemented across various settings to overcome healthcare access barriers, contributing to enhanced maternal health practices and positive birth outcomes. ^{13 – 19}

A systematic review on psychosocial interventions among pregnant women for smoking cessation found that incentives influenced smoking reduction but the effect on outcome measures pertaining to low birth weight, preterm births, and mean birth weight was not clear owing to small sample sizes. ²⁰

A previously conducted systematic review has shown only limited evidence that incentives may improve the frequency of prenatal care. This evidence is based on 5 trials and participants were majorly drawn only from low-income communities in Central America and North America. ¹⁷ A review of literature by Morgan et al., in 2013 showed that various types of incentives increased the number of antenatal care visits based on evidence from various studies conducted in LMICs. This study examined the allocation of incentives, considering both the demand side involving patients and the supply side involving healthcare providers, government agencies, and other stakeholders. ²¹

In a systematic review published in 2014, the evidence regarding five modes of demand-side financing was evaluated, focusing on their impact on out-of-pocket expenditure, utilization of health services, and cost-effectiveness. In addition, the study assessed maternal, perinatal, infant mortality, and morbidity among the poor and socially excluded pregnant women from LMICs. The study found that financing modes like cash transfers, conditional cash transfers and vouchers targeted for pregnant women increased the institutional deliveries and utilisation of antenatal and postnatal services. Evidence concerning maternal and neonatal mortality was sparse owing to the small sample size and shorter follow-up periods. ²² This systematic review has considered quantitative studies and qualitative studies. The study fails to clarify the design of intervention studies involved in the reviews and makes it difficult to relate the effect size to the various types of interventions.

The evidence generated in some of these reviews regarding the effectiveness of the provision of incentives for pregnant women focuses on a limited number of maternal and neonatal health outcomes like low birth weight, frequency of prenatal care, maternal and neonatal mortality. These reviews do not investigate any of the incentive-based behaviour change interventions that are also known to play a role in the prevention of pregnancy-related complications. This could include incentive-based tobacco and alcohol cessation, diet control intervention studies among pregnant women, etc.

For an effective continuum of care, it is essential to strengthen the link between the home, the primary health care facility and the referral centres. ⁸ Studies from various countries have shown that there is a higher risk of poor pregnancy-related outcomes when the antenatal care received is inadequate. ^{23– 25} In accordance with the revised antenatal care guidelines 2016, the WHO recommends a minimum of eight contacts to reduce perinatal mortality and pregnancy complications. ²⁶ There are various SBCE (Social Behavioural Community Engagement) interventions that facilitate the mobilisation of individuals, their family members, and communities for improving maternal and newborn health. ²⁷ Demand-side financing is one of the SBCE interventions to promote the utilisation of health services for better health outcomes. ²⁸ These include vouchers, Cash transfers, temporary payments, etc.

As per our knowledge, there are no comprehensive reviews focussing on the effectiveness of various incentive-based interventions that are targeted only for pregnant women during the antenatal period regarding broader outcome measures like the proportion of pregnancy, childbirth-related complications, perinatal deaths and proportion of preterm deliveries in resource-constrained settings that follows rigorous methodology of Systematic reviews. We aim to fill in the gap based on the work undertaken in previous reviews regarding the outcomes and the methodology. It is known that behavioural economics has potential in the formulation of effective health policies. Designing the incentives accordingly is one of the tools in the behavioural approach. ²⁹ This review of trials is thus essential to inform the effectiveness of incentive-based programmes targeted for pregnant women in LMICs. It will help the policy makers to utilise the resources more effectively and to integrate the evidence-based public health initiatives into the health system.

Our research questions are as follows:

1. Does the provision of incentives to pregnant mothers during the antenatal care period achieve better maternal and neonatal health outcomes than the absence of such services for pregnant women?

1a. What are the effects of incentives (any type) on the uptake of antenatal care services/utilisation of antenatal health care services (frequency of antenatal care, proportion of institutional delivery)?

1b. What are the effects of incentives (any type) on maternal and neonatal morbidity (proportion of preterm deliveries, low birthweight (less than 2500 g), proportion of antenatal and postnatal complications, compliance to Iron and Folic Acid (IFA) tablets intake and Tetanus Toxoid Injection (Inj TT) coverage, coverage and utilisation of maternal incentive based nutritional interventions, cessation of smoking, alcohol, tobacco, or any other unhealthy behavioural practises)?

1c. What are the effects of incentives (any type) on maternal and neonatal mortality (neonatal deaths, maternal deaths)?

•

To determine if any of the incentive-based interventions had an effect on maternal outcomes (proportion of antenatal and postnatal complications, proportion of institutional delivery, frequency of antenatal care, maternal deaths, compliance to IFA tablets intake and Inj TT coverage, coverage and utilisation of maternal incentive based nutritional interventions, cessation of smoking, alcohol, tobacco or any other unhealthy behavioural practises)

We used the following Population, intervention, comparator, and outcome (PICO) format to assess and select articles. 1.

Population: We have included all incentive-based interventional studies conducted on pregnant women in LMICs. We have adhered to the definition of low-middle-income economies as per World Bank 2019, “lower-middle-income economies are those with a Gross National Income (GNI) per capita between $1,026 and $3,995”. ³² We excluded studies conducted among pregnant women who belong to low- and middle-income countries but reside in high-income countries during the antenatal and postnatal period. We have excluded studies that involve pregnant women living in developed countries but belonging to low-income families/communities. We define immigrants as people who migrate from one country to another country. So, women who hold citizenship in a low-middle-income country but reside in high, upper-middle and low-income countries during pregnancy will be excluded.

Maternal outcomes are as follows: •

Frequency of antenatal care (number of visits and the content of care)

We have excluded studies that did not measure at least one of the above outcomes.

The search strategy is added in the repository and is cited in the data availability statement. Searches were finalised and conducted on different databases from inception until 2021. We searched Medline, CINAHL, SCOPUS, Web of Science, and Embase for relevant records. The PICO acronym was used to describe the search strategy. Additionally, hand searching was carried out on the Cochrane Central database. Records were additionally retrieved from Google Scholar, and the reference lists of included studies were examined.

The screening process was conducted using the Covidence software ( https://www.covidence.org/). Initially, two reviewers screened titles, followed by abstract screening according to specified eligibility criteria. The full-text screening was then independently carried out by two reviewers. Ineligible records were excluded with specific reasons documented in flow chart ( Figure 1). Any discrepancies were resolved through discussions between the two reviewers. Additionally, we have documented ongoing trial protocols are added in the repository and is cited in the data availability statement. ^{37– 39}

The data was extracted by one reviewer in the data extraction form under the following headings: study ID, author details, publication year, country details, participant details, intervention details, comparators, outcomes, type of incentive, funding details and type of setting (rural or urban). A second author verified the extracted information. The data was entered into Review Manager (RevMan) version 5 ( https://revman.cochrane.org/) (Cochrane Collaboration) by one author and verified by another author. Predefined outcomes that contained analysable data were as follows: Frequency of antenatal care, compliance to Inj TT coverage, the proportion of institutional delivery, proportion of women who received iron tablets and low birthweight.

The risk of bias was assessed using the tools mentioned in the Cochrane Handbook for Systematic Reviews of Interventions. ⁴⁰ This assessment was independently conducted by two investigators. In case of disagreements, a third investigator was involved for the resolution of the same. The ROB-2 tool (version 2019) ( https://methods.cochrane.org/bias/resources/rob-2-revised-cochrane-risk-bias-tool-randomized-trials was used for the studies with randomised study designs. The variant of the ROB 2 tool adapted for) cluster randomised trials (2021 version) was used for cluster randomised trials. We evaluated and reported the results in ‘Risk of Bias’ tables. (The data on ‘Risk of Bias” tables has been added in the repository and is cited in the data availability statement)

We synthesized the evidence both narratively and graphically, using forest plots. Additionally, a meta-analysis was conducted for certain outcomes with available data. We calculated the risk ratio (RR) with a 95% confidence interval (CI) to compare outcomes between the two groups and performed a random-effects meta-analysis. Statistical heterogeneity was evaluated using the I ² statistic, Tau ², and the Chi ² test.

A total of 3824 records were imported into Covidence software ( https://www.covidence.org/) for screening, sourced from various databases: Embase (1350), CINAHL (217), Web of Science (637), Scopus (460), Medline (960), and Google Scholar (200). After removing 721 duplicates, 3103 records underwent screening based on their titles and abstracts using Covidence software. We could not export records from Cochrane (CENTRAL) to Covidence. Hence those records (73) were screened separately on Excel. We assessed 212 full texts for eligibility. We could retrieve 14 records (10 studies) on Covidence. In addition, 8 records (1 study) were found through backward and forward citation searches. Altogether we have included 11 studies for data extraction ( Figure 1). We did not find any studies that focussed on incentive-based behaviour change intervention conducted in LMICs during the antenatal care period among pregnant women for e.g. incentive-based smoking/tobacco cessation interventions, incentivised nutrition-focussed interventions for appropriate dietary behaviour during pregnancy, etc.

The detailed characteristics of the included studies are mentioned in the Table 1.

Total number of pregnancies were 19565 from 10 studies. One study by Kusuma et al., ⁴¹ did not have data pertaining to pregnant women alone. However, we have included the study since the program targeted a mixed population that showed improved healthcare utilisation among pregnant women. The study duration ranged from 3 months to 6 years. Out of the 11 included trials, 8 studies were cluster RCTs ^{41– 48} and the remaining 3 studies were individually randomised parallel controlled trials. ^{49– 51}

Participants in seven of the studies lived in rural areas ^{42, 43, 46– 48, 50, 51} while three had residents from rural, semi-rural or urban areas ^{44, 45, 49} One study had an urban population. ⁴¹ Four studies ^{41, 43, 44, 47} had a population comprising both pregnant women along with mothers and lactating women. These studies were included since they had done subgroup analysis for pregnant women participants. While seven studies had only pregnant women as participants. ^{42, 45, 46, 48– 51}

All seven cluster RCTs have followed adequate accounting of cluster unit randomization in their analysis except Kusuma et al., 2016. ⁴¹ The studies included in the review were conducted in the following low- and middle-income countries: Uganda, Honduras, China, Kenya, Togo, Mexico, and Indonesia. Two of the studies were conducted in Zambia ^{44, 48} and two more in Nigeria. ^{45, 49} Eight studies were government-based interventions for the community. ^{41, 43– 48, 50} One of the studies was carried out by local Non-Governmental Organisations (NGOs) ⁴⁹ while the other one was funded by an international agency. ⁴²

We performed a meta-analysis for the three outcomes listed below, with the effects of the intervention shown in Table 2. Five RCTs did not have data for the outcome suitable for meta-analysis. 1.

Delivered at the facility: There was no clear difference between the two groups [Risk ratio (RR) 1.13, 95% confidence interval (CI) 0.86 to 1.47, 2676 women, two studies]. There was a high level of statistical heterogeneity. (I ² = 76%, Tau ² = 0.03 and Chi ² test for heterogeneity P = 0.04) ( Figure 2).

Whereas, for the compliance to IFA tablet intake only one study has reported data for iron tablet intake by pregnant women, hence no meta-analysis could be performed. Also, no data was available for the outcomes such as maternal death, coverage and utilisation of maternal incentive-based nutritional interventions, proportion of antenatal and postnatal complications, and cessation of smoking, alcohol, tobacco, or any other unhealthy behavioural practices.

1.

Low birthweight: Only one study has given details about the same. Another study did not assess birthweight as its primary outcome. ⁴⁶ The program showed a 127.3 g increase in birthweight among the intervention group and a reduction of 4.6 percentage in the incidence of low birth weight. ⁴⁶

The current systematic review has followed the methodology adopted in the Cochrane Handbook for Systematic Reviews of Interventions. ⁵⁵ The results generated out of this systematic review will inform or guide in implementing public health interventions and policies about incentive-based initiatives for pregnant women.

No restrictions concerning the publication year were followed for the search strategy.

We included all the financing modes implemented for pregnant women without any restrictions, including vouchers, transport incentives, in-kind goods, cash transfers, etc. However, the available evidence is limited to LMICs. The classification of LMICs has undergone significant changes in the past two years, particularly during the pandemic. ^{56, 57} We did not have enough resources to include studies conducted in languages other than English in LMICs. We could have missed out on the effects of studies conducted in LMICs in regional languages. Additionally, we did not search registries or explore grey literature sources.

The inclusion of studies was restricted to English and those published in scientific journals.

The current review was limited to studies that were primarily concerned about demand-side incentives. In the future, studies should involve a combination of both demand and supply-side incentives. In addition, focus of trials should be on outcomes like neonatal and maternal mortality, postpartum complications, and skilled attendance at homes alongside institutional delivery.

It is likely that the very low number of studies in LMICs could be attributed to the very low incentives, investments, and infrastructure for research activities in the regions compared to high-income countries that leads to substantial disparities in the conduct of research studies. Studies on tobacco and smoking cessation were excluded since they were all conducted in high-income countries. ^{58– 62} Most of the studies in the current review were funded by various external sources and some were conducted by the government. While a few were done in collaboration with the national government with the support of external funding. To promote more studies in LMICs in the future, national governments should prioritize extensive primary research. ⁶³ Communities need to be empowered to conduct community-based research activities that enable them to identify issues and develop solutions for the same. ⁶⁴

The outcomes examined in this review did not have any significant differences in the intervention group when compared to the control group.

The interventions to address maternal health concerns need to follow a multifactorial approach. Poverty is not the sole barrier for pregnant women to access healthcare services. Reports of obstetric violence/labour room violence have been commonly seen across India and the globe. ^{65– 67} Efforts to address the same needs to be taken. In addition, one needs to address the challenges faced by health care workers, ⁶⁸ including high patient numbers, resource limitations, and inadequate training in respectful maternity care, which can contribute to negative experiences for pregnant women. ⁶⁹ Maternal and neonatal health is also influenced by socioeconomic and environmental factors. ^{70– 72} In societies dominated by patriarchal norms, women often face lack of autonomy and decision-making power. This limitation frequently hinders their access to timely prenatal care and childbirth services. ⁷³ Moreover, deeply ingrained cultural practices such as inclination towards giving birth at home or the dependence on traditional midwives can further deter facility-based deliveries. ⁷⁴ Additionally, long travel distances, inadequate transport, and poor healthcare infrastructure create additional barriers, especially in rural and remote areas. ^{75, 76} Addressing these multifaceted challenges requires community-based awareness programs, policy reforms, and strengthening of healthcare systems to ensure equitable and respectful maternal healthcare access.