Development and Feasibility of a Group, Household Level Intervention To Improve Preconception Nutrition in Nepal

Background: In a setting such as Nepal with malnutrition and persistent poor maternal and infant health outcomes, developing interventions to improve the nutrition of preconception and pregnant women is essential. Aside from factors related to access and knowledge about diverse and adequate diets, inequitable gender norms and women’s low status and autonomy also feed into a lack of emphasis on women’s nutrition. Methods: In this paper we describe the three phases of the design of an intervention for newly married women, their husbands, and mothers-in-law in rural Nepal. We rst discuss ndings from a mixed methods formative phase, and how that fed into the design of the intervention itself, in close partnership with community members. We the describe the intervention, and ndings from a mixed methods feasibility and acceptability pilot among 90 participants. Results: Our formative work highlighted not just lack of awareness about nutrition, but also how factors such as women eating last, and limited mobility, played into poor diets. Low lack of communication and household and community inequitable gender norms were also important factors. Thus we designed Sumadhur, an intervention that brought groups of households (newly married wife, husband, and mother-in-law) together weekly for 4 months to strengthen relationships and gain knowledge through interactive content. We found Sumadhur to be highly feasible and acceptable by all respondents, with most (83%) attending 80% of sessions or more, the majority (73%) of participants reported “no diculties” in attending sessions, and 99% reporting that they would like it to continue. Pre/post surveys showed a decrease in the proportion of women of women eating last and increase in knowledge about nutrition in preconception and pregnant. Qualitative interviews suggested that respondents felt it made large impacts on their lives, in terms of strengthening relationships and trust, understanding each other, and changing behaviors. Conclusions: We show how a designing an intervention in close partnership with the target recipients and local stakeholders can lead to an intervention that is able to target complicated and culturally held practices and beliefs, positively benet health and wellbeing, and that is very well received.

Results: Our formative work highlighted not just lack of awareness about nutrition, but also how factors such as women eating last, and limited mobility, played into poor diets. Low lack of communication and household and community inequitable gender norms were also important factors. Thus we designed Sumadhur, an intervention that brought groups of households (newly married wife, husband, and motherin-law) together weekly for 4 months to strengthen relationships and gain knowledge through interactive content. We found Sumadhur to be highly feasible and acceptable by all respondents, with most (83%) attending 80% of sessions or more, the majority (73%) of participants reported "no di culties" in attending sessions, and 99% reporting that they would like it to continue. Pre/post surveys showed a decrease in the proportion of women of women eating last and increase in knowledge about nutrition in preconception and pregnant. Qualitative interviews suggested that respondents felt it made large impacts on their lives, in terms of strengthening relationships and trust, understanding each other, and changing behaviors.
Conclusions: We show how a designing an intervention in close partnership with the target recipients and local stakeholders can lead to an intervention that is able to target complicated and culturally held practices and beliefs, positively bene t health and wellbeing, and that is very well received.
Trial registration: ClinicalTrials.gov NCT04383847, registered 12/05/2020 https://clinicaltrials.gov/ct2/show/results/NCT04383847?view=results Background Gender inequality and low levels of empowerment among women contribute to poor health outcomes for women and children in South Asia. 1 Indicators of women's status and autonomy, including the gender development index, are lower in South Asia than other LMICs and strongly correlated to child stunting and wasting, among other adverse outcomes. 1 Low maternal education, limited ownership of land by women, poor female labor force participation, and low decision-making power among women are further associated with adverse child health outcomes and poor care-seeking in Nepal. 2,3 A preference for sons means that daughters born into families with more siblings are at increased risk of acute and chronic malnutrition and poor growth. 4 Married women who have not yet given birth or have not yet had a son have the lowest status in their marital households and are often con ned to their homes, restricting their ability to seek health care. They are further harmed by customs such as eating last, not speaking in front of household elders, and being at the will of their mothers-in-law. 5,6 Restrictive gender norms and practices are associated with poor maternal and child health outcomes through delayed/low careseeking, poor nutrition, depression, etc. Exposure to physical violence and abuse (such as food restrictions and denial of health care) from husbands and in-laws perpetuates inequality and worsens maternal, infant and child mortality and morbidity. [7][8][9] Young, newly married women in the preconception stage are at the lowest status in their marital households. In much of Nepal, marriages are arranged or at least heavily involve the family. Newly married women often eat last, cook and serve all other household members, and have limited mobility or decision-making power. [10][11][12] About half of women in Nepal experience a pregnancy within the rst year of marriage, making nutrition in early marriage critical 13 . Newly married women in food insecure households often have less access to diverse foods in their husband's compared to their parents' home. However, better relationships between daughters-in law and mothers-in-laws have been found to mitigate the association between household food insecurity and lower consumption of high-quality foods upon moving into the marital home. 5 Most interventions aiming to improve pregnancy and birth outcomes globally, including in Nepal, focus on women once they are pregnant, or pre-marriage (in adolescents), thereby neglecting newly married women in the important preconception period. Group-based intervention models have been shown to be successful at increasing reproductive and other health knowledge, as well as empowerment, for women in South Asia. [14][15][16][17][18] Interventions promoting gender equality and challenging restrictive gender norms can positively in uence women's health. 19 Group interventions are especially important, and impactful, for addressing household and community gender norms, and have been found to increase health knowledge in South Asia. [14][15][16]20 There have been a large number of group pregnancy interventions, and many have successfully improved IFA uptake, consumption of nutritious foods, and reduced anemia. 18 Engaging family members is especially important in South Asia, where co-residence early in marriage is still common and decisions are often made at a household level. 21 A 2020 systematic review of nearly 60 programs targeting gendered social norms and health found that more than half speci cally engaged males to decrease violence and rede ne household roles and responsibilities, though the majority (67%) named females as the intended primary bene ciaries. Approximately three-quarters (74%) of evaluations identi ed signi cant changes related to both health and gender. 19 Gaps remain, as only few evaluations measured nutritional outcomes. 19 Household interventions in Nepal involving women, their husbands, and in-laws have demonstrated improved livelihoods for young married women, increased earnings and savings, and reduced food insecurity. 17,22 However, we found no studies that targeted whole households; instead, publications to date recruited women, mothers-in-law, and husbands from separate homes, or only engaged some family members.
To address the intersecting community, household, and individual factors contributing to poor nutrition and early pregnancy among newly married women in Nepal, we conducted formative research, followed by a community-based participatory design phase, to develop Sumadhur (meaning "Best Relationship"). Sumadhur is a group intervention that engaged members of households with a newly married woman, her husband and her mother-in-law weekly over 4-months to provide interactive, informational content about nutrition and women's health, address inequitable gender norms and practices, strengthen household relationships and communication, and improve the status of newly married women. We then piloted this intervention among 90 participants (30 households, split into 6 groups) in rural Nepal. In this paper we rst describe the ndings from our formative work (Part 1), followed by describing our development process for the intervention (Part 2) and nally present the ndings from our pilot (Part 3).

Methods
Part 1: Preliminary research: Our team conducted mixed methods formative research in the communities in which we would develop our intervention. We rst conducted in-depth qualitative interviews with 20 household triads (newly married women, their husbands, and mothers-in-law) in rural Nepal. We then conducted an 18-month longitudinal study with 200 newly married women, followed every 6 months (4 rounds of data collection) between 2018 and 2020. More details about study design, recruitment, and data collection for this formative work, along with other ndings not presented here, have been published elsewhere. 5,23−25

Analysis
Descriptive statistics from the baseline survey data are presented (frequencies, percent). Details of the analysis of the qualitative data has been published elsewhere. 23,24 For the purpose of this paper, we reviewed the codes and themes, both looking at the data as individuals and then as triads.
Part 2: Development of Sumadhur: Through a community engaged design approach, we collaboratively developed and piloted Sumadhur, an intervention to in uence these pathways among newly married households in Nawalparasi District of Nepal. First, throughout the course of the mixed methods formative study described above (more than a 2-year period), we held bi-annual meetings with a District Advisory Committee (DAC), initially comprising ve members and later extended to 11 members to accommodate more community stakeholders. The DAC was comprised of representatives from district public health o ce (head), maternal and child health focal person of district public health o ce, civil society organization working on nutrition, and women and child health, local level elected ward members, etc.Their main roles were to suggest study areas for the study and intervention, participate in dissemination meetings, suggest potential intervention based on ndings, support the core team members to implement the study and monitor the intervention activities.
Based on the analysis of data from the formative phase and input from our district advisory committee, we began drafting a list of educational sessions. At this point, we also reached out to other researchers and implementers working in Nepal and India who had group interventions that addressed some of the topics that we were covering. Many of these groups shared their intervention protocols with us and gave us permission to adapt some of their content/materials to our population and target. Speci cally, we learned from the RANI project (India) 26 , the MIRA project (mother and infant birth activities, Nepal) 27  Group sessions were organized in a convenient location to all the participants such as ward o ces, schools, local government agriculture, and livestock building o ce hall (within their village, not more than half an hour walking distance, within approximately 3km radius). VDRC hired three facilitators with Bachelor's degree level of education to be trained for the intervention. Two facilitators, one each for Sunwal Municipality and Palhinandan Rural Municipality were selected and trained. The third facilitator was responsible to coordinate and help facilitators in both the municipal areas. The facilitators were provided a ve-day intensive training before the launch of the intervention. The training was facilitated by the core team members from CREHPA and VDRC. During the training, the facilitators were oriented on facilitation skills, content, and mock sessions. A eld o cer from CREHPA attended most of the sessions.
Core team members from CREHPA and VDRC also observed selected sessions and discussed lessons learned and provided feedback to the facilitators to improve in the subsequent sessions.

Analysis
For the quantitative data, we use descriptive statistics (frequencies and percent). Primary feasibility and acceptability measures include satisfaction (5-point Likert scale), if a participant would recommend to a friend, number of sessions attended, and main barriers. For the qualitative data, we used thematic content analysis. 33 Using inductive and deductive approaches, themes that pertained to the concepts of interest were identi ed and additional concepts or themes that emerged spontaneously during the interviews were also included. Once the key concepts were identi ed, a codebook was developed by a three-person team (NDS, AC, AM) through an iterative process. A quarter of interviews were double coded, and where discrepancies arose, these were discussed until agreement was reached.
All phases of this study received ethical approval from the University for California, San Francisco and the Nepal Health Research Council. The trial was registered at ClinicalTrials.gov NCT04383847, registered on 05/12/2020.

Part 1: Formative research:
Women were marginalized in terms of eating practices and diet, with almost half (47.5%) reporting that they ate last always or most of the time and 88.5% not meeting minimum dietary diversity standards (Table 1).
In the qualitative data, women described not eating enough and not eating foods they preferred, with one wife explaining "No one cares if I eat, what I eat. No one asks how I am feeling…. I feel no one cares about me." (Formative Phase Wife #1, Age 18). All household respondents described ordered eating patterns, with newly married women eating last. This was often explained by "culture" or "tradition" or as a practical solution to the need for someone to be serving the food, combined with a willingness on the part of the wife to eat less, as described by one husband: I: So, all four of you eat together?
R: No, my wife doesn't eat with us. She eats after we all nish eating.
I: Why is that so?
R: This is because we may need to add something when we eat. If she is eating as well, it becomes di cult for her to give us additional food. Therefore, we eat at rst and after we nish eating, she eats it.
I: So, does the food becomes su cient for her? R: Yes, it does. We usually tell her to cook food if it is insu cient. But she eats less and tells the food is su cient for her. (Formative Phase Husband #8, age 20) Some household triads reported positive and loving relationships, but many wives felt that they were not cared for "He doesn't help me in my work since he has married me and brought me here…My husband never asks me anything regarding my health and wellbeing. (Formative Phase Wife #11, Age 20)" Quantitative data supported low levels of communication between spouses, with less than half of wives reporting having discussed the number of children they wanted with their husbands (42.5%) and a third (32%) feeling comfortable talking to their husbands about sex (Table 1).
Mobility was also strictly limited, with only 50% having ever left the house since marriage. In depth qualitative interviews highlighted how isolated and lonely wives were, as described below: R: I like to go but who allows me to go? I don't have anyone who accompanies me. Restricted mobility was also seen as a potentially large barrier to women being able to participate in an intervention, especially alone: Do you think your wife will be willing to participate in the program? R: Yes, she will be interested to participate if the program is within the house. If it is outside, she will not be able to participate.
I: Why can't she participate?
R: It is not in our culture to allow newly married women to go outside the house.
I: And will your mother participate in such programs?
R: Yes, she will be willing to participate.
I: Will she be allowed to go outside the house and participate?
R: Yes, she will be. She doesn't have to ask anyone. She can simply inform us and go in the program.
I: So when can your wife go outside the house and participate in the program? R: It depends; the culture here is that we have to stop her from going outside as long as we can. The people here aren't educated and they think that women should be kept at home and men should go outside and be the bread earner. (Formative Phase Husband #9, age 21) In terms of the design of the intervention, respondents reported that they would prefer an intervention that engaged household members in addition to the women, and that it not be moderated by a community health care worker. Furthermore, involving husbands and mothers-in-law was also seen as an acceptable way for wives to be allowed to leave the home.

Part 2: Development
Our ndings highlighted nutrition as a high need area, which also aligned with our goals of improving preconception health and wellbeing of women (and ultimately maternal and infant health). These ndings, in combination, suggested that strengthening household relationships by bringing the three key players together (newly married women, husbands and mothers-in-law) was important, and also could be a vehicle to address household and individual practices. Bringing groups of households together could further get at community norms around expectations of women's role and eating dynamics. We hypothesized that strengthening relationships and addressing inequitable norms could improve women's household status, increase mobility, and increase access to food.
Based on the preliminary ndings, and through the community-engaged process described above, we developed Sumadhur, a 4-month long, weekly group intervention for triads (wives, husbands, and mothers-in-law) that covers nutrition, anemia, intrahousehold food allocation, prenatal health and pregnancy care, gender inequitable norms and practices, fertility planning and contraception, and couples and household relationship dynamics (Table 2). Each session combined educational information with interactive topic-related games and activities that helped build relationships and break social and gender norms. As can be seen in Table 2, most sessions included all three household members, but a sub-set only included the wife and husband, for more sensitive topics. Detailed discussion and feedback with our partners and DAC informed the decision about which sessions should all have household members, the number of households that should be in each group, additional input about speci c session content. Except in two sessions, all sessions were moderated by the trained moderators from VDRC; those two other sessions had health workers come in to provide more detailed information on family planning and biology/menstruation (Sessions 14 and 10).

Part 3: Intervention pilot results
i. Feasibility and acceptability: A total of 44 households were approached initially. Ten households did not agree to be a part of the study at the rst approach citing various reasons (inability to manage time, no good relation among wives and mother-in-law, the wife having gone to her maternal place for a long time). Four households backed out at the last moment citing fear of COVID-19. Therefore, an additional 4 households were approached. A total of 30 households consented to take part in the intervention. One household was again replaced after presurvey as the husband obtained visa for foreign employment. Those households were divided into 6 groups (each group comprising of 5 households). The sessions were conducted weekly in ve groups, however, one group requested that they be conducted biweekly. Most of the sessions were conducted in the daytime as per the participant's choice and convenience. The sessions lasted an average of 81minutes (73-91minutes).
A total of 31 households participated, however, only 28 husbands answered the surveys. One of the husbands had gone abroad immediately after the pre-survey and did not attend any sessions. Another husband attended two sessions and then went abroad. Table 3 shows the socio-demographics of the participants.
Sumadhur was acceptable and feasible, with 97-100% of participants reporting that they would recommend it to a close friend and that they talked to someone about something they learned (Table 3). Most (83%) attended 80% of sessions or more, and the majority (73%) of participants reported "no di culties" in attending sessions; participants identi ed health issues (n=12), personal work obligations (n=4), and personal household obligations (n=4) as the primary attendance barriers. Satisfaction rates were high, with 100% or respondents being very or somewhat satis ed. Additionally, there was a desire for the intervention to continue, with 99% reporting that they would like something like this to continue in the future. Additionally, despite initial hesitancy about combining husbands, mothers-in-law, and wives, 95% of wives reported that it felt "good" to attend session with their in-law.
Respondents also enjoyed the program, including interacting with others in their community, as one wife explained how it built relationships with others in her community: I felt very good to participate with other community members. It felt like being in a family when people from different community, caste and family structure participated in the training. I didn't feel that we represented different family or community. It felt as though we all were from the same family and were attending the training for mutual bene t. Maybe I felt that because everyone was cooperative, understanding and helpful….They all had their own de nition and understanding on the topics covered. … in earlier days, my neighbors didn't call me by my name. But after attending the training, they call me and ask me about the things I've learned from the training. I used to tell them the things I've learned. I also told them to participate in such trainings in future. After the training, I felt that the community here is very supportive. Other women of my age come to me and ask me about the training. I tell them about the things I've learned and also suggest them to maintain peace and mutual cooperation in the family. (Pilot Wife 13, age 21) A husband also discussed how he appreciated the group dynamics and engagement of other community members: R:I felt comfortable to be in a group. There were other members with whom we could interact and know them better. Had it been only my family in the training, it would have been less interactive. In the institute, we try to have more students to make the class better. Similar approach was used in the training and I liked that part. It wasn't uncomfortable with other community members. Also, it is very important to give training to community members as well. Here, the community follows traditional practices. I think such training programs will help to change their thought process.
I: What do they follow?
R: In our community, daughters-in-law cannot come outside freely, woman cannot go outside their house to work, mothers-in-law and daughters-in-law don't interact much etc. Such things need to be changed. It was somewhat similar in my household as well. But after the training it has changed a little. As I said earlier, we eat together and interact much more than before. My wife and my mother interact more and this makes me feel good. (Pilot Husband #25, age 19) Despite initial concerns about constraints on women attending due to restricted mobility, Sumadhur provided an opportunity for some to leave their homes for the rst time since marriage: R: I was shy to attend the program at the beginning, which later on decreased. I couldn't understand few words spoken by the facilitator, but I hesitated to ask for clari cation as there were other male participants in the program. With time, I felt comfortable and my hesitation also decreased. In addition to immediate dietary changes, participants liked the information related to nutrition. When participants were asked which topics they found most useful, anemia and iron folic acid were reported most often by 61% (n=55) of participants. Topics including household relationships, gender inequality, household eating patterns, and stress and anxiety were also among the top ve most useful.
We found that while engaging with sensitive and stigmatized topics together was new for many households and groups, by the sixth session comfort and engagement was reported by the facilitators to have increased across the participants.
Even though people are educated, they feel shy in front of their family members. After participating in the training, the participant gained con dence to speak and express what they feel to their family members. The training helped in their personal development. One participant did express her feeling to me after the training. She said that she used to feel shy in front of her family members. After attending the session, her mother-in-law includes her in family conversation and encourages her to speak. In my opinion, through the training we brought closeness among daughters-in-law and mothers-in-law. If they can continue this bonding, this will help the family to become strong. (Facilitator) ii. Preliminary impact of the intervention As a result of the intervention, nutritional knowledge and practices improved. There was increased awareness of the need for preconception, pregnant, lactating or postpartum women and adolescents to eat more ( Figure 1). Eating patterns also shifted, with a decline in the number of daughters-in-law reporting that they ate last (from 43% to 3%) and increase in the proportion reporting that the household ate together usually or all of the time (from 37 to 52%) (Table 4).
Overwhelmingly, respondents had absorbed information about what types of foods women should eat before and during pregnancy. Intertwined with increased knowledge about nutrition itself, respondents described changing dynamics around support for women, order of household eating and awareness of the importance of strengthening relationships, as described in the two quotes below: Nowadays, my mother-in-law helps me a lot with my household work. As I am pregnant now, she asks me to eat egg daily saying it's good for my health…Even if I don't say anything, she tells me to have snacks saying that I might be hungry. She also says that I should eat more in this condition. She says eating frequent and nutritious food will be bene cial to both me and my baby inside me. (Pilot Wife #16, age 23) My wife is pregnant, she also learned things related to pregnancy and childbirth.... Now, both my mother and my wife are trying to establish a good relationship. She is eating nutritious food and also asks for my help. In earlier days, she didn't ask for my help in any work. I also have been helping her in household work. ...we all eat together now...My wife eats more nutritious food like green vegetables, egg, meat, cereals and lentil. We tell her to rest more and prevent her from doing any heavy work. If there's any important work, we take suggestion from other family members before doing it (Pilot Husband #30, age 27) Myths that mothers who eat too much or take iron tablets will gain weight, and other misconceptions about nutrition and health, were also dispelled, as this respondent explained: I used to hear that to prevent unborn baby from gaining weight, pregnant mothers were prevented from eating nutritious food and iron tablets. But I came to know that all these weren't true. We have to eat nutritious food, take adequate rest and consume iron tablets on time. I've been implementing these things in my life as well. (Pilot Wife #13, age 21) Respondents also described getting to know each other better, for example one husband told about learning something surprising about his wife: "I didn't know that my wife was somewhat scared of childbirth. She said that she is scared of complications during her delivery. We told her not to stay worried as staying worried might cause health problems during pregnancy." (Pilot Husband #30, age 27) Two mothers-in-law described how it helped build a relationship with their new daughters-in-law, and how the family-group approach made them less shy about participating: Going together with my son and daughter-in-law, I came to know many things. It was easier for me to go together with my son and daughter-in-law. I would have been shy to talk too if I had gone alone, but going together with them became easier for me…..I have felt changes even if it's a little. There are changes in everything regarding behavior of the family members, working environment, eating habit, con icts. I had always wondered how would my daughter-in-law be. As she was recently married, I had not known her well. I got an opportunity to know her while going to the program. We used to go together, talk on the way and discuss about the things discussed after we would come back. We perform household work together, talk to each other and share our things. I had a negative attitude towards daughter-in-law earlier which has changed now. My daughter-in-law treats me well and so do I. (Pilot MIL #9, age 54) I have found it so good that I cannot express in words. (She said with much of excitement, and with happiness). I really liked the part where we learnt that we should not discriminate between a family who has sons with those who have daughters only. But, in villages if the family has female child only, people talk a lot behind their backs. It is not just sons who can take care of the parents even daughters can look after them….. When I went along with my son and daughter-in-law, if there were things that I didn't understand, my daughter in law would make me sit and used to explain by saying that we were taught these things today. And when we went together, the villagers would also say that it is so nice seeing her going with her son and daughter in law to the program. I used to feel very good. (Pilot MIL #25, age 43) A newly married women shared about how her relationship with both her husband and mother-in-law changed: R: Yes, there has been changes. The behavior towards me when I was newly married has changed after taking part in the intervention. I had to perform all the household chores by myself before but the scenario has changed now. Everyone helps household chores. There is equal distribution of food among all the family members. We don't ght with each other. There has been changes in husband's and mother in law's behavior. Previously, my husband used to force me for sexual relation but he doesn't do it now. He asks about my health, respects me and loves me. I am happy now.
I: Are you surprised listening to views of your family members?
R: I am surprised seeing the love and care towards me nowadays. I used to feel being neglected previously but now husband, mother-in-law, sister-in-law, brother-in-law care and love me. They help me in household chores, everyone eats together. I have understood that If a person gets good information then it is bene cial to whole family. (Pilot Wife #3, age 20)

Recommendations
All participants felt that others in their communities and other districts should have access to this program. Participants also raised that other family members should be invited to join, such as other daughters-in-law in co-resident households with multiple sons and families living together, or, most commonly, the father-in-law. One woman explained: People in the villages are still a bit ignorant, and my father-in-law comes home drunk sometimes. Well, my husband, my mother-in-law are attending the program, but my father-in-law does not understand, he needs to be taught, even he needs to be invited to the program. Everyone in the home should participate in such programs, and they should be informed about everything, such as "what would make it good and better at home". Well, it so happens that at my home, we often have arguments and I want to make them understand (Pilot Wife #26, age 20) My father-in-law was very curious to know about this training. He used to ask us every time. If he attends the training, he can learn many things from it. This is because, the older generation mindset takes time to change. Family and society cannot change and become progressive until and unless such older generation's mindset changes. It is important to bring changes in belief system of the head of the family.
In order to bring such change, training like "Sumadhur" can play an important role. If head of the family bring change in their belief system, other members also follow it. Therefore, I think it is very essential to include father-in-law in the training. (Pilot Wife #13, age 21)

Discussion
Targeting triads from multiple households in an intervention designed to improve preconception and pregnancy nutrition as well as strengthen relationships and address inequitable gender norms was found to be feasible, acceptable and showed improvement in behaviors. Participants were overwhelmingly positive, nding that this intervention made marked changes on their lives. Furthermore, even culturally entrenched behaviors such as the order of eating in the household 12 , were seen to be malleable through this 4-month intervention, suggesting that perhaps some practices that have hitherto been assumed to be hard to change might actually be able to be shifted when people have additional information and understand each other's perspectives more.
Despite restricted mobility of newly married women in this setting, and the potential for this to be a barrier to an intervention that occurred outside of the home and with other community members, we had high rates of participation from newly married women. Not only this, but the opportunity to participate in this intervention served as a way for newly married women to be able to leave the house prompting downstream effects on their autonomy. A related unexpected nding was the impact of the intervention on mothers-in-law. While we hoped that the intervention would help build relationships between mothersin-law and their daughters-in-law, as it did, we also found that mothers-in-law felt empowered by the opportunity to participate in the intervention, leave the house, and interact with others in this new way. Few interventions engage, much less target, mothers-in-law (or older women), and this population deserves more focus and attention. 34 We identi ed one past intervention that engaged married women, their husbands and mothers-in-law to reduce violence, address gender norms, and provide economic opportunities for young women, which was found to be effective at impacting some of these outcomes. 35 Overall, the intervention was successfully implemented and welcomed by district level stakeholders.
Therefore, it could be scaled up with few minor modi cations in the content, approach and preparation of information, education, communication/behavior change communication materials. The modality of conducting sessions weekly seemed appropriate as it does not overburden the participants with information. Since the sessions were conducted in local language, development of materials in local language would have an additional advantage. It is possible that an intervention such as this may be more interesting and applicable to rural and relatively less educated participants, and thus, more research is needed about its applicability in a more urban setting.
Part of the success of this intervention was the community engaged process, from formative research, to design of the intervention, and roll-out. The local government and district level stakeholders showed their commitments to support this kind of intervention, therefore the engagement of such stakeholders from the early stages of the design would be bene cial and key to successful implementation of future intervention such as this. Additionally, the design and content of the intervention was derived directly from the ndings of the in-depth, multi-year formative work, which helped make the intervention salient to participants. Our team was able to shift focus based on the voices of community members, for example, our initial design idea had been an intervention in households given to household members individually by community health workers and our formative phase helped us see the need for different types of facilitators to provide information in a group structure.

Conclusion
Through a community-engaged process informed by mixed methods research, we were able to design a feasible and acceptable intervention that not only addressed norms and practices, but actually through its very structure (bringing young newly married out of the house, bringing households together) challenged these norms and led to empowerment, strengthened relationships, and changes in health knowledge and behavior.

Declarations
All manuscripts must contain the following sections under the heading 'Declarations': Ethics approval and consent to participate: All phases of this study received ethical approval from the Authors' contributions: NDS led the manuscript writing, oversaw and participated in data analysis, was the PI on the grant and oversaw all aspect of study design and execution. AM: led the quantitative data analysis; AC: led the qualitative data analysis; MD: contributed to study design and oversaw data collection and implementation, contributed to manuscript, LG: participated in the design of the intervention content and support manuscript preparation; MJ: provided mentorship in intervention design and feedback on analysis and writing, SW: provided mentorship on nutrition, overall study, and feedback on analysis and writing; MP: local primary lead of the study, supported study design, oversaw implementation of the intervention and all data collection, supported manuscript preparation.      Figure 1 Proportion of respondents agreeing that certain people need to eat more than others, pre and post intervention