Identifying the association of women’s empowerment with reproductive and maternal health practices using a cross- sectional study in the context of self-help groups in rural India [version 1; peer review: awaiting peer review]

Background: Globally, women’s empowerment is captured through different combinations of indicators related to their collective and individual empowerment. In this paper, the association of diverse domains related to women’s collective and individual empowerment on their reproductive and maternal health practices were assessed. Collective empowerment referred to a cluster of indicators measuring mutual support, rights, and access to services women received, as a result of membership in self-help groups (SHGs) focused on microfinance activities. Women’s individual empowerment was measured through their perception to make decisions on their own, be mobile, financially self-reliant, self-confident in expressing themselves, having self-esteem and being free from spousal violence. Methods: Cross-sectional data were collected in 2017 from 2,197 SHG members across 57 administrative blocks of Uttar Pradesh, India. Using exploratory and confirmatory factor analysis, the sub-domains of individual and collective empowerment were developed. The reproductive and maternal health indicators included antenatal care, delivery preparedness, postnatal care and current contraceptive use. Bivariate and multivariate analyses, using logistic regression, were carried out to measure the association of sub-domains of women’s empowerment with health practices. Results: Collective and individual empowerment were independently and jointly associated with correct health practices. Women with greater empowerment through independent mobility, high selfesteem, access to financial resources, and confidence in interacting with a frontline worker, were more likely to access antenatal care. Similarly, delivery preparedness—which entailed deciding on the place of delivery, arranging for transportation during labor, and managing Open Peer Review Approval Status AWAITING PEER REVIEW Any reports and responses or comments on the article can be found at the end of the article. Gates Open Research Page 1 of 13 Gates Open Research 2022, 6:72 Last updated: 14 JUN 2022


Introduction
The United Nation's Sustainable Development Goals identify gender equality as a key development indicator achieved if men and women enjoy equal rights, opportunities, and freedoms (European Institute for Gender Equality. Concepts and definitions. Vilnius, Lithuania: European Institute for Gender Equality; 2019). However, women continue to be disproportionately disempowered in terms of freedoms, power dynamics, and autonomy 1 . A growing body of literature acknowledges the nuances and complexities in measuring women's empowerment as a key measure of gender equality 2 . Dimensions of equality vary in significance according to local contexts, which evolve at their own pace 3 . Research strongly indicates the need to bridge gaps between theoretical constructs and empirical research to better understand the association of women's empowerment with health and other development outcomes 3,4 . This paper identifies key domains of women's empowerment and measures their associations with reproductive and maternal health practices in a rural Indian context and thereby contributes to bridging those gaps.
Recent literature synthesizes existing evidence on the various measures of women's empowerment [3][4][5][6][7] . Mandal et al., acknowledge that women's empowerment can be measured at household, service delivery, and community levels 7 . Huis et al., categorize women's empowerment within three dimensions: micro, referring to personal empowerment; meso, capturing relational empowerment; and macro, measuring empowerment within societies 3 . A review of empirical evidence reveals over nineteen domains that have measured women's empowerment in sixty studies 4 , ranging from individual characteristics such as age, education, household wealth, and employment, to constructs such as decision-making, freedom of movement and mobility, financial autonomy, gender attitudes, and self-efficacy; the two most common constructs of women's empowerment being women's participation in household decision-making and their mobility. Different combinations of questions in each of these studies measured women's empowerment.
Literature also documents the association of empowerment with the health and wellbeing of women and their children. An analysis of India's 1998-1999 National Family Health Survey-2 shows that increased mobility and decision-making, opposition to spousal violence, and financial autonomy were associated with a higher number of antenatal care visits, utilization of postnatal care, and institutional deliveries; the results were further substantiated by more recent studies in Ethiopia and Nepal [8][9][10][11] . Another study in Nepal shows that increased household decision-making is associated with improved postnatal care but not antenatal care 12 . Studies have also established associations of various measures of women's empowerment with family planning method use [13][14][15][16][17] . Generally, women's empowerment in these studies is linked to increased method use. Women are more likely to use family planning services if they enjoy increased mobility, greater financial autonomy, freedom from violence, better spousal communication, and increased decision-making autonomy for various aspects of their individual and familial lives 13,[15][16][17][18][19][20][21][22] . In addition to these studies, three recent extensive literature reviews examining the relationship between women's empowerment, and maternal and child health practice show that increased household decision-making, lack of exposure to spousal violence, financial autonomy, and increased mobility are associated with correct healthy behaviors such as antenatal care, skilled attendance at birth, contraceptive use, and reduced maternal mortality 17,23,24 .
Women's empowerment and its impact on health practices has been captured by two measures at the aggregate/community level. First, some studies aggregate individual characteristics of women within communities, showing that communities with higher-than-average education or employment among women are more likely to exhibit healthy behaviors than communities with lower-than-average education or employment among women 10,25 . Second, few other studies acknowledge the role of women's microfinance collectives, called self-help groups (SHGs), in increasing women's individual empowerment or improving health practices [26][27][28][29][30][31][32] . SHGs organize 10 to 12 women within a community to attend regular meetings and save small amounts of funds every week; the funds are used for mutual lending. Groups also facilitate bank loans at small interest rates for family emergencies and entrepreneurial activities. SHGs also engender social cohesion among members, encouraging them to demand their rights and services while developing a support structure that women and their families can rely upon in times of need. SHGs are also considered a means through which various health-related interventions or activities can be integrated to improve the practice of healthy behaviors among member households 26,28,31,32 .
Evidence on the effect of women's collectives or SHGs in improving women's authority and resources in their own lives or supporting their individual practice of healthy behaviors is limited, with variation among studies 27,29,30,33 . A systematic review showing that SHG membership increased women's individual empowerment for financial autonomy and mobility did not show consistent good health practices 30 . A recent Ethiopian study examining the effect of SHGs on women's health and wellbeing showed that SHGs helped increase women's awareness of their rights, as a collective, but with no increased access to services or greater financial autonomy within their households 27 . In other studies, interventions through SHGs increased antenatal care coverage and institutional delivery but without other health benefits or increased individual empowerment 26,[34][35][36] .
Aggregating the individual characteristics of women in communities and examining the characteristics of SHGs may need to be performed separately. Aggregated values of individual characteristics represent the general condition of women in that community, while the characteristics of women's SHGs represent the support structures and social networks women create among themselves that they can utilize in times of need. There remains a great deal to learn about the interplay of women's collective empowerment through SHGs and women's individual empowerment within their own homes and personal lives, and particularly how these expressions of empowerment are affected by SHG membership, further associated with access to health services, and improved maternal and child health practices. This paper examines the association of women's empowerment with their reproductive and maternal health practices, in relation to women's SHGs. Women's empowerment is captured in two broad dimensions: 1) collective empowerment, measured as women's empowerment within SHGs for access to services, exercising of rights, and mutual support in times of need, and 2) individual empowerment, measured through women's personal experiences and perceptions of support within their households and relationships. It is critical to consider both collective and individual empowerment simultaneously, as they do not change in isolation but represent a general trend of increased awareness of one's rights and entitlements through increased access and utilization of opportunities, such as health services. We hypothesize that these two domains of empowerment -collective and individual -are correlated, and that greater levels of women's empowerment are associated with improved health practices. This paper also examines the relationship between collective and individual empowerment and their association, independently as well as jointly, on selected health practices.

Study settings
This study is part of an ongoing evaluation of 'Uttar Pradesh Community Mobilization Project' implemented by Rajiv Gandhi Mahila Vikas Pariyojana, a non-governmental organization based in Uttar Pradesh, India that supports a network of SHGs across the state (rgmvp.org). Since 2002, these SHGs have been established primarily among the most marginalized women in the community to address gender inequality, improve women's access to microfinance credits, and share livelihood information and opportunities. Each SHG meets weekly to address savings, microfinance, and livelihood matters. With external support since 2011, Rajiv Gandhi Mahila Vikas Pariyojana expanded its scope and established SHGs for over 1.7 million poor women in over 49 districts of Uttar Pradesh. In these groups, the organization integrates discussions of reproductive and maternal health in the weekly meetings in addition to community outreach events, strengthening links between local women and their frontline workers.

Evaluation approach
This study uses data from a cross-sectional survey conducted during September 2017 to January 2018 of eligible women who were currently married, aged 15 to 49 years, and had a live birth in the 12 months preceding the survey. If a woman under eighteen was interviewed, she was considered an emancipated minor, as per Indian guidelines, and hence parental permission was not sought. The study sample comprises SHG members from 57 sampled blocks in 20 districts of Uttar Pradesh. In order to address the potential selection bias of the sampled respondents, a two-stage sampling design was applied to select the study participants: blocks sampled in the first stage, and gram panchayats in the second. Blocks were first arranged in ascending order by percent of scheduled caste or tribe populations and following a systematic random sampling technique the required number of blocks were selected. Gram panchayats with varying proportions of their populations covered by SHGs were included by dividing them into three strata. The required number of panchayats were drawn randomly and equally from each stratum. Within each selected gram panchayat, all households with an SHG member were mapped and listed. All eligible women were identified and approached for a face-to-face interview. If more than one eligible woman was in a household, one woman was randomly selected for the interview. Being a cross-sectional survey, eligible women were not followed up after the initial interview.
The study sample comprises of 2,197 eligible women, which is a sub-sample of the original evaluation and has more than 90% power and 95% confidence interval to examine the relationship between the reproductive and maternal health outcomes of interest and the key independent variables related to collective and individual empowerment. Information was collected from three sources for this study -the eligible woman, the head of her household, and her SHG leader. Information on housing characteristics was collected from the heads of the household, eligible women provided information on reproductive and maternal health practices and empowerment, and SHG leaders were asked about their groups' characteristics and functioning. Written informed consent in local language Hindi was obtained from all study participants to participate in the study and share the findings from this study. A hard copy of the consent form was left with the participant for their records. The tools were also administered in Hindi.
The study was reviewed and approved by the institutional review board of the Population Council, the organization conducting the evaluation, and recorded as protocol number 764.

Measurements
The health practices examined in this paper are: 'four or more antenatal care visits during last pregnancy', 'at least one activity to prepare for delivery, such as deciding on a place of delivery, identifying mode of transportation, and saving adequate funds for delivery', 'postnatal care visit from a frontline worker or SHG member within seven days of delivery' and 'using a family planning method' at the time of the survey. These indicators were based on the World Health Organization guidelines and were asked of each eligible woman during her individual interview 37,38 . Delivery preparedness includes accomplishing at least one of three activities: deciding on a place of delivery (home or health facility), identifying and arranging the mode of transportation to a facility for delivery, and saving or arranging money for delivery expenses or emergency. A woman is said to have received postnatal care if a frontline worker or SHG member conducted a home visit, or if the woman herself visited a frontline worker or health facility within seven days of delivery. Each of these variables were dichotomized for analysis.
The primary independent variables of interest are collective empowerment and individual empowerment, which were collected during the interview with the eligible woman. Collective empowerment was measured through four key sub-domains: social cohesion, referring to a woman's belief that her SHG will support her in times of need (captured through 12 questions/items in the tool); efficacy, referring to the belief that women work together for positive changes in health (measured through seven questions in the tool); agency, referring to women assisting other members for local health and administrative services (measured through five questions in the tool); and action, capturing the respondent's own experiences in creating social change within the past year (captured though 12 questions in the tool) (refer to the Extended data 39 for the full list of questions and responses). Formative research validated these sub-domains of collective empowerment in previous studies of SHGs in similar settings of Bihar and elsewhere 36,40,41 .
Individual empowerment was measured through six key sub-domains: confidence, referring to a woman speaking in public and recognizing a health emergency (captured through nine pre-tested questions in the tool); mobility, referring to the ability to leave the house for various chores and activities (captured through 12 questions in the tool); decision-making, referring to engagement in major decision-making within the household for health services, making purchases, determining major life decisions, etc. (captured through 26 questions in the tool); self-esteem, capturing a respondent's perception of her own worth (captured through seven questions in the tool); financial inclusion, capturing a woman's ability to obtain a loan from the SHG, own her own assets and resources, and make basic financial decisions independently (captured through 12 questions); and freedom from violence, capturing a respondent's reported experience of any kind of physical, emotional, or sexual violence from her spouse within the past 12 months (captured through 14 questions) (refer to the Extended data for the full list of questions and responses). This holistic approach to understanding women's empowerment enabled a determination of which characteristics influence the practice of reproductive and maternal health behaviors. Formative research refined questions related to mobility from the recent National Family Health Survey 42 , while other domains were taken from tools administered in multi-state SHG evaluations over the past five years 31,36 .
Analytical approaches A reliability test for each sub-domain tested for internal consistency within the dataset. Methodologists recommend a minimum alpha coefficient between 0.65 and 0.8 (or higher in many cases) 43,44 . The Cronbach alpha score of more than 0.7 for the items in each scale, except for collective agency, indicates good internal consistency (see Table 1 for details on scores) [45][46][47][48] . We identified potential measures using exploratory factor analysis with factor loading (0.5) followed by a priori confirmatory factor analysis. We then tested these models using goodness of fit tests, which help identify an acceptable model that is the best fit. The goodness of fit tests included the Root Mean Square Error of Approximation (<0.8) 49 , Comparative Fit Index (>0.9) 50 , Tucker Lewis Index (>0.9) 51 , and Standardized Root Mean Square Residual (<0.8) test 52 . In instances where models did not fit the data well, confirmatory factor analysis was repeated including additional model parameters (allowing measurement errors to covary among specific items) to improve the model fit. We identified the best model using Root Mean Square Error of Approximation, Comparative Fit Index, Tucker Lewis Index, and Standardized Root Mean Square Residual in consonance with relevant expert literature 53,54 (see Table 2 for goodness of fit statistics obtained by confirmatory factor analysis). Following these tests, specific items with lower factor loadings (allowing measurement errors to covary) were dropped while constructing the sub-domains (16 items were dropped out of 116) 55 . Internal consistency of each sub-domain and domain was measured again after the model fit. Although the Cronbach alpha did not improve for collective agency, we accepted the confirmatory factor analysis model as a 'best fit'. This is corroborated by Nunnally 56,57 , who suggested that newly developed measures can be accepted with an alpha value of 0.60 compared to the norm of 0.70 and above. The Cronbach alpha for financial inclusion dropped below 0.7 after the best fit model was generated (Table 1). Similar to collective agency, we included the confirmatory factor analysis model of financial inclusion as the best fit, in accordance with Nunnally's suggestion 56,57 . We constructed an additive index to include all items within a sub-domain to form a continuous variable for collective as well as individual empowerment, which subsequently assessed the correlation of various domains of empowerment and were used for multivariate logistic regression analyses.
We also generated an overall combined score of women's empowerment using 36 items of collective empowerment and 80 items of individual empowerment. We then tested the model for the best fit using confirmatory factor analysis tests and identified the best model with 30 items of collective empowerment and 70 items of individual empowerment. We also ran a reliability test for the best model indicating a Cronbach alpha of greater than 0.7, hence ensuring high internal consistency. All the items derived after the model fit were summed to form a continuous variable, which was used for multivariate logistic regression analysis.
While examining the association between women's empowerment and reproductive and maternal health practice, we controlled for respondents' individual characteristics associated with empowerment and health practices in the literature 4,11,25 . Characteristics controlled for in regression analyses were based on single item questions in the women's questionnaire: respondent age (captured as a continuous variable), education as a categorical variable (no education, one to seven years, eight to 11 years and 12 or more years of formal schooling), caste (scheduled caste/tribe, other underprivileged castes, others), employment (employed for income/not), in addition to multiple questions assessing household wealth. The household wealth index was developed through principal component analysis using information on 26 household amenities measured for six categories, five housing characteristics and one on asset ownership, taken from a nationally representative tool and administered to the head of the household 41 . Using these data, wealth terciles were developed based on equal proportion of the population being divided into three categories: poor as low, middle as medium, and rich as high.
Because Rajiv Gandhi Mahila Vikas Pariyojana was implementing a program to increase women's knowledge of healthy reproductive and maternal health practices, SHG women were exposed to various program activities. Some blocks received more active interventions, while others received ad hoc information and support from program staff. The survey captured and controlled for program exposure to measure the effect of empowerment on women's reproductive and maternal health practices as a dichotomous variable of exposure to SHG activities related to health or no exposure. The 11 SHG activities reported by SHG eligible women were: 1) accompaniment by a SHG member during antenatal care or delivery, or met SHG member during pregnancy or met SHG member at least two times after delivery; 2) advice from a SHG member on delivery preparation, pregnancy and neonatal complications, place for treatment during complication, cord care, kangaroo mother care, breastfeeding, family planning, and sanitation; 3) information through SHG outreach activities on cord care, kangaroo mother care, breastfeeding, family planning, and sanitation; 4) invitation by a SHG member to attend health meetings; 5) home visit by a SHG member who shared information on healthy maternal and newborn care practices; 6) health messages from a SHG member during pregnancy; 7) receipt of health leaflets; 8) received a congratulatory letter; 9) viewed a health video within the last three months; 10) attended a local baby shower during last pregnancy; and 11) attended community evening meetings within last three months. Group members who were exposed to any of these activities were considered exposed to SHG activities.
A group's characteristics could influence women's expressions of empowerment, exposure to health-related program activities, as well as the practice of healthy behaviors. Hence, we measured SHGs' characteristics captured through the SHG leaders' interviews and included them in the regression analysis while assessing the relationship between women's empowerment and reproductive and maternal health practices. Group characteristics were captured by six questions focusing on SHG duration (of three or more years), regular meetings (six or more in the last three months), regular weekly savings activities, regular register updates, internal lending, and regular loan repayments; the last five characteristics were key determinants of a functional SHG by Rajiv Gandhi Mahila Vikas Pariyojana, focused on group economic and administrative functioning. A Likert scale interpolated SHG strength from these six dichotomous variables, ranging from 0 to 6 and further categorized as 0 to 2 for 'needs improvement', 3 as 'moderate', and 4 to 6 as 'good'.

Results
The mean age of respondents was 28 years (Table 3). About 40% of respondents had no formal education, more than half were from scheduled castes or tribes, and were from low wealth index households. Only 11% of respondents were employed. About three quarters of women had been exposed to at least one health related SHG activity during their last pregnancy or following the birth of their youngest child. Approximately one-third of interviewed women belonged to SHGs that were not functioning optimally, while 41% belonged to stronger groups that met regularly for group activities. Most SHG members (84%) stated that their group had been formed more than three years ago, while around half reported that their group had conducted less than six meetings in the prior three months, less than half the expected number. Nearly all members (97%) saved money weekly as part of the group's activities. A quarter of group members (26%) had taken out a loan through the group, and another 37% belonged to a group that had recently repaid or was currently repaying a bank loan. The Chi-square test of association suggests that formal education, household asset ownership, and exposure to SHG activities are positively associated with antenatal care, delivery preparation, postnatal care, and use of a family planning method (Table 3).
Overall, over a third (39.6%) of the population interviewed had at least four antenatal care visits during their last pregnancy, around three-quarters (72.6%) had performed at least one delivery preparedness, a third (31.3%) had at least one postnatal care visit within seven days of delivery and 41.3% of The correlation matrix of sub-domains of collective and individual empowerment shows that over 80% of the correlations noted were significantly associated with one another (Table 4). Collective efficacy was positively correlated with other subdomains of collective empowerment such as social cohesion, collective agency and action, as well as the sub-domains of individual empowerment such as self-confidence, decision-making, self-esteem, and financial inclusion. Similarly, self-esteem was positively associated with all reported sub-domains of collective and individual empowerment. Financial inclusion was, however, negatively associated with sub-domains such as mobility, decision-making, collective agency, and collective action, signifying that women with assets and access to funds are less likely to engage in collective action and have limited ability to exercise freedom of movement or make decisions on their own. Freedom from violence is positively associated with social cohesion, collective efficacy, self-confidence, self-esteem, and financial inclusion, indicative that women who do not experience spousal violence are more likely to have trust in their group to support them in times of need, or to work together for positive change. Those women were also more likely to perceive their own worth positively, and own assets and resources.
Results from Model A of the logistic regression show that more empowered women were associated with four or more antenatal care visits during their last pregnancy (β = 0.02, SE = 0.004), preparation for their last delivery (β = 0.03, SE = 0.005), and a postnatal care visit within a week of delivery (β = 0.01, SE = 0.004) ( Table 5). Findings from Model B show that antenatal care and postnatal care visits were primarily influenced by individual empowerment, while delivery preparedness was driven by collective support from SHGs, as well as individual empowerment. In Model C, where relationships between the outcome variables and each subdomain were tested individually, collective efficacy, women's self-confidence, freedom of mobility, self-esteem, and financial inclusion had positive and significant associations with antenatal care. Women's preparedness for delivery during their last pregnancy was significantly associated with nearly all the sub-domains of collective and individual empowerment, except for freedom of mobility and freedom from spousal violence. Postnatal care within a week of delivery was positively associated with social cohesion, self-confidence, and financial inclusion. Generally, combined domains of collective and individual empowerment did not appear to affect family planning method use, but when we observed the specific sub-domains of empowerment, family planning method use was positively associated with collective agency, self-confidence, and experience of less spousal violence within the past 12 months.

Discussion
This paper measures empowerment extensively to describe both collective and individual empowerment among women in SHGs. Collective empowerment encompasses cohesion, whereby a member believes she will be supported by her SHG in time of need; efficacy, when a member believes her group can work together for positive change, agency indicates a group has assisted its members in seeking services, and action denotes that a respondent has herself participated in activities for social change. An extensive array of questions also relate to women's individual empowerment within six domains: self-confidence in engaging with people of authority and new situations, freedom of mobility in leaving home for various activities, autonomy in decision-making within the household and in matters of the family's and a woman's own wellbeing, self-esteem as a measure of a woman's  c Sample size is smaller as respondents who were pregnant were not asked about current FP use perception of her own worth, access to financial resources and making those decisions independently, and freedom from experiences of violence, specifically spousal violence. This study's findings reveal strong correlation between collective and individual empowerment among women who belong to SHGs, suggesting that collective empowerment and individual empowerment occur simultaneously for women in self-help groups. This supports previous research that suggests SHGs increase women's abilities to access financial resources and opportunities, which further manifest as greater autonomy and propensity for autonomous decisions 58 .
Women who were individually more empowered were more likely to seek four or more antenatal care visits, signifying that a pregnant woman who is more empowered will be more likely to express greater self-confidence and leave her home to interact with a frontline health worker, in addition to having greater access to financial resources for utilizing those services. Other studies employing a few individual empowerment measures separately found similar results 9,11,23,24,[34][35][36]59,60 .
Delivery preparedness was associated with both collective empowerment and individual empowerment. Women are more likely to prepare for their infant's delivery if they express higher levels of empowerment related to self-confidence, decision-making, self-esteem, and financial autonomy. Collective support from fellow SHG members is also reflected in a positive association of social cohesion, and collective efficacy, agency and action. SHG members are more likely to assist fellow members in delivery preparations, and those women exercise their abilities to plan, save money, and consult with others in determining next steps.
Postnatal care was associated with higher levels of overall combined empowerment, primarily driven by greater association with individual empowerment. A woman received postnatal care if she visited a frontline worker or health facility, or if a frontline worker visited her at home. Literature suggests that domains of individual empowerment such as increased self-confidence and financial autonomy are associated with a postnatal care visit 9,11,12,59 , which is substantiated by our findings. Our findings show association of increased social cohesion with postnatal care; other studies have not acknowledged the role of collective empowerment.
In this study, family planning is not found to be associated with any aggregate empowerment scores -overall or combined, collective or individual -but increased collective agency was positively associated with family planning use, reflecting women's confidence in the group's support. Individual self-confidence was positively associated with family planning use, reflecting a woman's ability to exercise choice and will within her home. Family planning often serves as a proxy for inter-personal relationships and communication between couples; hence, women reporting lower experiences of spousal violence tend to use family planning methods more. This is corroborated by the literature 14,61-63 . Changing behavior around family planning use takes time, and further effort is needed -through SHGs -to meet women's need for comprehensive services. Programs working to address reproductive and maternal health behaviors through SHGs have an opportunity to focus on family planning messages, encourage discussions of family planning practice, and strengthen links with the health system to improve women's access to family planning services. Our findings are consistent with other studies that note positive associations of certain elements of individual empowerment, such as self-confidence and freedom from violence and family planning use 13,15,16,[18][19][20][21]23,24,32,60 . Only one other study that looks at the association of SHG membership with family planning method use shows positive results 64 . These studies neither account for collective empowerment nor capture a wide array of individual empowerment elements, and as a result our study contributes to the global body of literature by studying the association of collective empowerment and individual empowerment with family planning method use.
Limitations -A key limitation of this study is that the associations of empowerment with health practices in relation to SHGs are from one point in time. This prevents us from looking at changes in empowerment over time, how these changes affect uptake of reproductive and maternal health behaviors among SHG members and determining the interplay of the sub-domains of collective empowerment and individual empowerment with one another. Further longitudinal studies with SHGs are needed to observe change in empowerment patterns affecting reproductive and maternal health practices over time. In addition, exploring questions of collective empowerment with other women in a group would reduce bias from endogeneity. Adequate time and effort are needed to collect comprehensive information from respondents, through local, well-trained investigators. Our study acknowledges this need by using well-trained and experienced investigators to interview respondents in private settings.

Conclusions
Through the deployment of an extensive tool capturing women's empowerment both collectively and individually, we found that collective and individual empowerment are correlated when associated with women's SHGs. Furthermore, both collective and individual empowerment are independently and jointly associated with better reproductive and maternal health practices such as antenatal care, delivery preparedness, postnatal care, and current family planning use. These results suggest that SHGs are a powerful mechanism for increasing women's realization of their rights and opportunities, enabling them to access services and improve the quality of their lives. While this study demonstrates an association of empowerment with better reproductive and maternal health practices, these results should be interpreted with caution. While these findings are not generalizable to non-SHG populations, they may be applicable to similar sociocultural contexts and settings within India and globally. Future interventions should utilize SHGs to share messages about reproductive and maternal health practices and programs specifically encouraging women to exercise their individual and collective expression within their homes and communities to accelerate these healthy behaviors. Further research on the multi-dimensional domains of collective and individual empowerment presented in this paper is necessary for achieving the maximal effectiveness of SHGs. As women have opportunities to join SHGs and lead healthier lives, through information and support gained in SHGs, societies will be able to achieve gender equality at an accelerated pace and reap the benefits of a more equitable society.

Underlying data
The underlying data have been adequately deidentified; however, due to the nature of the evaluation, they are not openly available. The data can be accessed by contacting the owner on the link in Harvard Dataverse (see below) or by emailing Ms. Christina Tse, Population Council at publications@ popcouncil.org. A valid request, which will be considered by Population Council, is required to access the data.

Extended data
Harvard Dataverse: Extended data for 'Identifying the association of women's empowerment with reproductive and maternal health practices using a cross-sectional study in the context of self-help groups in rural India'. https://doi.org/ 10.7910/DVN/MHZ5MD 39 This project contains the following extended data: • Questionnaire: UPCMP HH midline tool.pdf • Questionnaire: UPCMP SHG leader midline tool.pdf • Questionnaire: UPCMP Woman midline tool.pdf Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).

Consent
Written informed consent for publication of the participants' details was obtained from the participants.