Evaluating the Effect of Community Engagement on the Birth Preparedness and Complication Readiness in the Community Level Interventions for Pre-eclampsia (CLIP) Trial in Northern Karnataka, India

Background: Increased birth preparedness and complication readiness can promote timely identication of the need for seeking skilled care and arrival at the appropriate facility for pregnancy complications. Engaging communities can support effective usage of local health services, yet implementation of community engagement programs can be challenging. The objective of the paper is to describe the process of community engagement in northern Karnataka, India and its impact on pre-eclampsia knowledge, birth preparedness and complication readiness, pregnancy-related care seeking and maternal morbidity. Methods: Community engagement was conducted in intervention clusters within the “Community-Level Interventions for Pre-eclampsia” (CLIP) trial in Belagavi and Bagalkote districts in rural Karnataka, India. Community engagement attendance was summarized according to participant group (pregnant women and women of reproductive age, mothers and mothers-in-law, community stakeholders, health workers and others). Pre-eclampsia knowledge, birth preparedness, health services engagement and perinatal outcomes was evaluated within trial surveillance. Outcomes were compared between trial arms using a mixed effects logistic regression model on RStudio (RStudio Inc, Boston, United States). Community feedback notes were thematically analysed on NVivo12 (QSR International, Melbourne, Australia). Results: A total of 1,379 community engagement meetings were conducted with 39,362 participants between November 2014 and October 2016. Community engagement activities may have had an effect on modifying community attitudes towards hypertension in pregnancy and its complications. However, rates of pre-eclampsia knowledge, birth preparedness, health services engagement and maternal morbidities among individual pregnant women were not signicantly impacted by community engagement activities in their area. Conclusion: Evaluation of our community engagement program in India demonstrates the feasibility of reaching pregnant women alongside household decision-makers, community stakeholders and health workers. More research is needed to explore the pathways of impact between broad community mobilization to strengthen support for maternal care-seeking and clinical outcomes of individual pregnant women.

Results: A total of 1,379 community engagement meetings were conducted with 39,362 participants between November 2014 and October 2016. Community engagement activities may have had an effect on modifying community attitudes towards hypertension in pregnancy and its complications. However, rates of pre-eclampsia knowledge, birth preparedness, health services engagement and maternal morbidities among individual pregnant women were not signi cantly impacted by community engagement activities in their area.
Conclusion: Evaluation of our community engagement program in India demonstrates the feasibility of reaching pregnant women alongside household decision-makers, community stakeholders and health workers. More research is needed to explore the pathways of impact between broad community mobilization to strengthen support for maternal care-seeking and clinical outcomes of individual pregnant women.

Background
Over 94% of maternal deaths occur in low-and middle-income countries (LMICs), primarily in South Asia and sub-Saharan Africa, and are preventable (1,2). Maternal deaths relate primarily to delays in triage (ability of care provider and women to identify who is severely ill and requires urgent care), transport (ability to get women to appropriate care when needed) and treatment (ability to provide appropriate treatment when care accessed) (3). Increased birth preparedness and complication readiness can help mitigate against delays and ensure timely identi cation of the need for seeking skilled care and arrival at the appropriate facility for pregnancy complications. Birth preparedness includes knowledge of pregnancy danger signs, identi cation of health facility and transport for birth and in case of complications, identi cation of a birth companion and preferred birth attendant, supplies to bring to the facility, funds for any expenses, and identi cation of a support person to look after other children at home as well as compatible blood donors if needed (4).
By identifying women at highest risk of complications at the community-level, transportation and treatment can be targeted to those most in need in a timely manner. Involvement of community members and raising community awareness of health in pregnancy are critical in the implementation of community health programs to reduce delays and improve maternal health services (5). Engaging the community is vital for effective functioning of the health system; however, involving communities in the implementation of the programs can be challenging.
Over the past few decades, community engagement (CE) has emerged as an increasingly effective strategy for harnessing community potential, particularly for health improvement (6-11). CE is de ned as 'a process of working collaboratively with groups of people who are a liated by geographic proximity, special interests, or similar situations, with respect to issues affecting their well-being' (12). It is a dynamic relational process that facilitates communication, interaction, involvement, and exchange between an organization and a community for a range of social and organizational outcomes (13). This strategy has been used for health promotion, research, task shifting and policy making to address a variety of health issues (14)(15)(16)(17). CE enables a more contextualized understanding of the perceptions and contexts of the members of the community, thereby facilitating stronger relationships.
This paper intends to describe the process of CE and its impact on knowledge around pre-eclampsia, birth preparedness and complication readiness, pregnancy-related care seeking and rates of maternal morbidity with task shifting of timely blood pressure recording and timely recognition and referral by Accredited Social Health Activists (ASHAs).

Methods
The Community Level Interventions for Pre-eclampsia (CLIP) India Trial was one of three independentlypowered cluster randomised controlled trials (cRCT); the others took place in Pakistan and Mozambique (NCT01911494). We undertook a two-phased community (including primary health centre (PHC)-level) cRCT in rural Karnataka in Belagavi and Bagalkote districts ( Fig. 1 map) where the incidence of pregnancy hypertension is 10.3% (18). Belagavi and Bagalkote include 1278, and 627 villages, respectively. The adult female literacy rate varies in the two districts, at 70·2% in Belagavi, and 59·3% in Bagalkote, respectively (19). More detail on the CLIP India Trial has been published elsewhere (20). In brief, the unit of randomization was the PHC area and restricted, strati ed randomisation was undertaken to allocate 12 clusters to the intervention and control groups, using population size as a single strati cation factor. The CLIP Trial focused on implementing community-level evidence-based care to reduce all cause maternal and perinatal mortality and major morbidity by supporting early identi cation and prompt referrals for pregnancy complications by the lowest level health care providers i.e., ASHAs. The intervention included engagement of the community and its stakeholders; blood pressure assessment by ASHAs; and usage of PIERS-On-the-Move (POM). POM is a mobile-based application decision aid for community health workers to identify women at risk of adverse outcomes from pregnancy hypertension and guide management, including treatment with oral antihypertensive medication, intramuscular magnesium sulphate, and referral to the hospital depending on the calculated risk (21,22). The project received ethics committee approvals from KLE Academy of Higher Education and Research Deemed-to-be-University (MDC/IECHSR/2011-12/A-4; ICMR 5/7/859/12-RHN) and the University of British Columbia (UBC, H12-03497).
The CE component was carried out in intervention clusters over the entire trial period (November 2014 -October 2016). The CE activities aimed at creating culturally and contextually appropriate discussion to improve maternal health awareness and action around the prevention of maternal morbidity and mortality. CE aimed to improve birth preparedness, nutrition, appropriate antenatal care (ANC) and promote appropriate health care seeking including acceptance of ASHAs for task shifting. Topics included warning signs of pregnancy complications, permission to seek care, identifying a skilled birth attendant, facility for delivery, nutritional aspects, and mode of transport as well as saving up funds for transport and treatment. Discussions also included adverse pregnancy outcomes and success stories in the community, overcoming local barriers to accessing maternal health care and the interventions in the CLIP Trial. Small group meetings involving pregnant women and her family members were also conducted. Many a times these activities were done at late hours of the evening to include decision makers of the family.

Process of facilitating community engagement
Community engagement meetings were primarily held at the anganwadi (AW) centres in the village, which offer basic health services. These were locations where pregnant women and children routinely access ANC, immunizations and other maternal and child health care services. Meetings were also held at subcentres, PHCs, temples, community halls, schools or other appropriate nearby locations for large monthly sessions.
The activities convened multiple levels of the local health care system. This included anganwadi workers (AWW) at the AW centres, ASHAs who worked in the villages and focused on connecting mothers and children to primary care, Auxiliary Nurse Midwives (ANMs) who worked at sub-centres, Lady Health Visitors (LHV) who supervised the ANMs, and medical o cers who coordinated the PHC. In addition, ASHA supervisors and Registry Administrators (RAs) appointed as a part of the trial attended meetings.
Feasibility studies in these communities revealed that pregnant women rarely made health care decisions on their own, and other family members and community leaders were highly in uential in this decision making (23). Consequently, CE activities involved household decision-makers who were often in charge of women's decisions to seek care, along with community leaders, such as elected representatives and teachers, who provided advice and at times provided nancial assistance.
The CE activity schedule was prepared at the beginning of each month by the local study team and shared to the respective community stakeholders. Meetings lasted between 30 and 150 minutes. These meetings were facilitated by the study team members involving the local health o cials led by medical o cers of PHCs. Sessions began with a short formal introduction and description of the purpose, participants were encouraged to actively participate, share personal experiences and discuss case examples. Reviewing local experiences was considered the best way to share information with participants; therefore, recent cases of pregnancy complications, referrals and outcomes were discussed. There were no monetary incentives for participants for attending the meeting. Participants were provided with nutritional snacks and fruits. Later, the success stories of participants were narrated by the participants to the attendees.

Data collection and analysis
A prospective population-based surveillance system was established as part of the CLIP India Trial, modelled on the National Institute of Health Global Network's Maternal and Newborn Health (MNH) Registry (24). Married women of reproductive age likely to conceive within 12 months were identi ed by an annual household survey, they were consented and enrolled into the CLIP Trial once pregnant. Data were collected soon after enrolment, soon after delivery and at 42 days postpartum from the mother in the local language by cluster PHC health care workers identi ed as registry administrators (RAs). Data collection included demographic factors, birth preparedness, care-seeking, obstetric information, preeclampsia knowledge, perinatal outcomes and treatment information at different time points as speci ed above. Data forms were reviewed for completeness and entered weekly into the local database. Deidenti ed and encrypted data were transferred to the central server for analyses. For CE activities, study staff and supervisors documented the number of participants and community feedback after each meeting. Figure 2 illustrates the strategy for assessing CE in the CLIP India Trial, based on the process evaluation framework for the CLIP Pilot Trial in Nigeria (25). CE attendance was summarized according to participant group: pregnant women and other women of reproductive age, the pregnant women's mothers and mothers-in-law, community stakeholders (community leaders, male and female decision-makers), health workers (AWW, ANM, ASHA, ASHA supervisor, LHV, medical o cers, nurses, RA) and others (family members, neighbours, friends, community members).
Pre-eclampsia knowledge evaluated awareness of: 1) abnormal bleeding in pregnancy, 2) high blood pressure during pregnancy, 3) seizures in pregnancy, 4) that pregnancy hypertension can be life threatening; by recalling at least one of the above conditions and at least four symptoms of high blood pressure in pregnancy for the composite pre-eclampsia knowledge score. Birth preparedness and complication readiness (BPCR) evaluated a pregnant woman's plans for: 1) transport in case of emergencies, 2) permission to seek emergency care, 3) money saved for an emergency, 4) identi ed health facility for delivery, and 5) at least two of the components for the BCPR composite score.
Pre-eclampsia knowledge, birth preparedness, health services engagement and perinatal outcomes indicators were summarized between arms with counts and frequencies for categorical variables and medians and interquartile range [IQR] for continuous variables. For health service questions, we restricted the denominators to women who delivered by trial end as these indicators are more likely to occur later in pregnancy. As pre-eclampsia knowledge and birth preparedness were asked throughout pregnancy and all women received community engagement, we included all women in assessing these rates. Outcomes were further compared between trial arms using a mixed effects logistic regression model (to account for clustering) with a logit link. Models were adjusted for age, maternal education, husband education, gestational age at booking, and parity and results are presented as odds ratios and 95% con dence intervals. Quantitative data were analysed using RStudio (RStudio Inc, Boston, United States), which was complemented by qualitative community feedback recorded by study staff. The community feedback notes were imported to NVivo12 software (QSR International, Melbourne, Australia) for thematic analysis.

Results
A total of 1,379 CE meetings were conducted. Among them 586 sessions were conducted in Belagavi district and 773 sessions at Bagalkote district. The median number of sessions per cluster was 204 [IQR 186-281]. Table 1 describes the activities across clusters and the participants reached. As long as pregnant women remained in the same village, they were encouraged to attend all sessions. The other participants, such as stakeholders and decision-makers, were usually different for each session. Pregnant women were most represented at the meetings, followed by mothers-in-law who often accompanied them. Husbands and fathers-in-law were least represented in the meetings as they often had competing priorities; however, overall sessions were well attended by target groups. Staff noted that they appreciated engaging household decision-makers and community stakeholders alongside pregnant women and health workers. Household decision-makers may not have been aware of pre-eclampsia and other maternal health risks and their inclusion in the discussion supported care for pregnant women and increased value in attending ANC.
"The people appreciated these community engagement meetings. They said that they learnt many things about the care of the pregnant women, which they were not aware of earlier. The pregnant women became more aware about their health and their parents also started giving more support and care to her." "Earlier, it was di cult for the woman to take permission from the decision makers in the family to seek antenatal care. But after conducting the community engagement meetings, the views of the decisionmakers were changed so that the seeking of permission for antenatal care became easy." Tables 2 and 3 present summaries of pre-eclampsia knowledge and birth preparedness, the rates for both (composite and components) were not signi cantly different between arms. Only a minority of participants from both intervention and control (2-6%) could name at least four symptoms of high blood pressure in pregnancy or had overall pre-eclampsia knowledge. In contrast, there were higher levels of birth preparedness among women in both intervention and control with a majority of women reporting that they had arranged transport, has permission for emergency care, and identi ed a health facility for delivery. Only saving funds for an obstetric emergency was lower, with about half of women declaring that they had funds saved up. Con dence intervals of odd ratios are wide re ecting high heterogeneity between study clusters [see Additional Files 1 and 2].  CE feedback indicates an increased awareness of antenatal care as an important component of birth preparedness. Staff noted that through CE activities, decision-makers and community leaders gave more value to preparation for delivery and pregnancy complications.
"The women became more aware of the antenatal and postnatal care. The decision-makers and the community leaders started giving more support in antenatal and postnatal care of the woman." "Most of the participants found the discussion about the pregnant woman's care to be unique, as they were not used to prepare in advance, in anticipation of complications." While staff noted that there were changing opinions on the value of ANC and birth preparedness in the community, Table 4 illustrates that health services engagement by pregnant women in the CLIP Trial were similar between arms. Almost all women had at least one ANC visit and three out of four women had at least four visits. Approximately one in ten women visited the PHC for reason other than routine care and about 3% of women were admitted to a health facility for reasons other than delivery, staying a median of three days. One in twenty women experienced a maternal morbidity and there was no signi cant difference between arms (OR 0.95, 95%CI: 0.61-1.50, p = 0.84). **Data taken from CLIP India primary trial paper and includes different adjustment factors (20) Community support for maternal health CE activities supported mobilizing community leaders to better support maternal health locally. There were examples of community leaders advocating for appropriate maternal health care seeking with household decision-makers as well as pushing for change to support local health systems.
"The views of stakeholders also changed so much that the arrangement of transport, nancial support, counselling the decision-makers became effective. For example, in the initial period of the study, there was one incident where the husband, the male decision-maker of the family, of the woman with eclampsia refused to take her to the hospital but after counselling by the community leaders and his family doctor, he agreed to take her to the higher care hospital." "The community leaders recognized the signi cant role played by the ASHAs and the PHC personnel in promoting the health of the pregnant women. The PHC did not have an ambulance of its own and the community leaders saw to it that a new ambulance was made available at the PHC. They also built a room above the PHC building as a meeting place for the AHSAs to continue the work actively."

External factors and barriers
Barriers of poverty, health infrastructure gaps and poor quality of care at facilities were frequently highlighted in CE discussions.
"People accepted all the topics. They only found the issue of arranging money to be a problem because they had meagre earnings." Additionally, the cost of treatment at private hospitals was often brought up in discussions in conjunction with the lack of services at government facilities. Some participants pointed out that there was no obstetric specialist at a nearby secondary level hospital for a period of time. Staff re ected that community members sometimes highlighted the need to strengthen care at facilities in addition to raising community awareness.
"There were some suggestions from the community about the improvement of services at the local government health centres." At referral facilities, participants reported that they often had to wait a long time before they would be provided services. They also reported that rural, uneducated people may encounter stigma and discrimination when seeking care at urban tertiary facilities.
"The participants wanted to know how to get immediate admission and care at higher and bigger tertiary care hospitals." "They also demanded about early admission and prompt service at the tertiary care hospitals as there are some uneducated persons in the community who face some problems in getting the services in bigger hospitals."

Discussion
Our evaluation of the CLIP CE program in India demonstrates the feasibility of reaching pregnant women alongside household decision-makers, community stakeholders and health workers. This initiative involved almost 40,000 participants in 1,379 sessions over a period of two years. Bringing different parts of local communities together to discuss pregnancy care may have helped to shift local opinions on the value of ANC to monitor for pregnancy complications and mitigate risks. However, rates of pre-eclampsia knowledge, birth preparedness, health services engagement and maternal morbidities remained similar across arms in the CLIP Trial. According to our evaluation framework (Fig. 2), our study found no evidence that the outputs were associated with our expected outcomes.
Although research ndings related to the effect of community participation to improve access to maternal health care services have been mixed (26), a previous study in Uttar Pradesh in Northern India found signi cant effects of community health worker home visits and monthly community meetings on improved self-recognition of problems faced during pregnancy, birth preparedness components and knowledge of pregnancy danger signs (27). Differing outcomes between our ndings and the study in Uttar Pradesh highlight two issues: rst, the importance of context, and secondly, challenges in measuring impacts of community mobilization. Uttar Pradesh has much higher maternal mortality rates than Karnataka (28), suggesting that CE may have a stronger effect where there is lower baseline access to maternal health services. Secondly, in contrast to targeted community meetings with traditional maternal-care providers and birth attendants, our CE strategy was focused on broader community mobilization and strengthening support for maternal care-seeking among community stakeholders, which has previously been discussed in the literature as challenging to measure (5,29).
CE involved shifting opinions on maternal health and cultures of care, behavioural change and mobilization and can take more time than the limited timeframe within a clinical trial. Logic models can help to understand the pathways of change, however, some of our expected short-term effects in birth preparedness via community mobilization may take more time than measured during the trial. A systematic review found increased knowledge of birth preparedness and complication readiness as a result of interventions but increased knowledge did not always correspond to increased utilization of maternal health services (30). While the CE program was successful in reaching a wide audience of participants and community feedback suggested increased support for maternal health care-seeking, these may be upstream indicators or insu cient to change clinical outcomes.
A systematic review on the effects of community participation on improving uptake of skilled maternal and newborn care highlighted the importance of qualitative research to understand unforeseen impacts and socio-political factors (26). While quantitative results did not reveal an effect in our study, staff qualitatively noted that effects were noticeable in the opinions of community leaders and household decision-makers in their support for pregnancy care. Effects of CE on community leaders and household decision-makers were not adequately captured in our evaluation framework that focused on measuring effects with only the pregnant women enrolled in the trial. Additionally, the discrepancy between the quantitative and qualitative results may also highlight the in uence of external factors where community members gained appreciation of maternal health issues but remained limited in their capacity to access care. Poverty and quality of care available at local health facilities emerged as signi cant barriers shared in community feedback, which is in line with previous research on the determinants of maternal health care services utilization in the area (23).
Nevertheless, qualitative ndings that suggested improved awareness and attitudes towards pregnancy hypertension and its complications among community leaders and household decision-makers is promising. Previous research in the area found lack of knowledge around pre-eclampsia and eclampsia (31). Hypertension speci cally as a pregnancy condition was unknown and there was no speci c terminology in the local language (Kannada) for pre-eclampsia.
CE has signi cant repercussions for the implementation of health care interventions and policy making. In addition to the individual-level factors, community-level factors play a vital role in understanding whether or not a woman is able to accept maternal health interventions. However, since CE was one of the components of a complex intervention, the effects assessed do not re ect CE alone. Intervention of multiple factors makes the impact assessment much more complicated.

Conclusion
Our study found that CE activities may have had an effect on modifying community attitudes towards hypertension in pregnancy and its complications but did not nd an impact on individual women's preeclampsia knowledge, birth preparedness and care-seeking indicators. Contemplation of speci c individual factors and community capabilities in engaging in the health seeking behaviours would contribute to the vision of supporting people-centred methodologies to deliver health promotion. Through the complex social interactions inherent in the current CE activities, appraisals over time will evaluate the effectiveness and sustainability of these interventions.

Declarations
Ethics approval and consent to participate Ethics approvals were obtained from KLE Academy of Higher Education and Research Deemed-to-be-University (MDC/IECHSR/2011-12/A-4; ICMR 5/7/859/12-RHN) and the University of British Columbia (UBC, H12-03497). All participants provided a written informed consent prior to study participation.

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
PVD, RJD, MBB, SSG, AAM, LAM conceptualised the trial and components of the intervention. MBB, SSG, AAM, USC, GMK, AK, UYR, SGB, GIM, NVH, CCK, AV, MV implemented and co-ordinated monitoring of the trial. JB and MWK performed the quantitative and qualitative analyses. AK and MWK wrote the rst draft of the manuscript. All authors provided feedback and review of the manuscript. Figure 1 Belagavi and Bagalkote Districts, Karnataka State, India Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.