This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on https://www.researchprotocols.org, as well as this copyright and license information must be included.
More than half of the women of reproductive age (aged 15-49 years) are anemic in India. The uptake of and adherence to iron folic acid (IFA) supplements remain low despite sustained efforts to increase their use. With India’s burgeoning digital environment, mobile phones offer a potential medium for increasing their uptake, especially when combined with interactive voice messages that deliver entertaining stories infused with norms-based educational messages.
This study aims to investigate whether a norms-based entertainment education mobile health intervention can increase self-efficacy for IFA adherence among women of reproductive age in Odisha, India.
Mobile reduction in anemia through normative innovations (mRANI) is a randomized 2-arm study that includes assessments before and after the intervention. All study participants will be recruited from the intervention arm of the parent reduction in anemia through normative innovations trial only. Although the usual practice is to randomize participants either to a treatment arm or a usual care control arm, we will assign the mRANI control group to another entertainment education–based treatment group that is designed to improve bystander intervention to reduce violence against women. Data collection for the mRANI study is embedded in the parent trial and will include baseline and end line assessments. The primary outcomes are self-efficacy for IFA adherence and violence against women–related bystander intervention. The inclusion criteria for the mRANI study are participation in the parent trial and phone ownership. Women (approximately n=400) who meet the mRANI inclusion criteria will be randomly assigned to the IFA arm or the bystander arm. Ordinary least squares regression with robust SEs will be conducted to assess between-group comparisons at the end line. A mediation analysis will be conducted to examine whether social norms and interactivity mediate the relationship between intervention exposure and primary outcomes in both arms. Real-time monitoring data will offer insights into intervention receptivity and audience engagement.
Data collection for the mRANI study is integrated within the parent trial. Household surveys were conducted between February and March of 2021. Responses on the mRANI study’s primary and secondary outcomes were collected from 381 participants. The data analysis is expected to be completed by October 2021.
This study will provide evidence on whether a mobile health norms–based entertainment education intervention can increase self-efficacy for IFA adherence and violence against women–related bystander intervention.
PRR1-10.2196/26252
More than half of the women of reproductive age (aged 15-49 years) in Odisha, India, are anemic [
The RANI project targets nonpregnant women who are currently underserved by the government’s anemia reduction programs. Various government initiatives in India (eg, National Nutritional Anemia Control Program, National Iron Plus Initiative, and Anemia Mukt Bharat) provide iron syrup to infants and iron tablets to both adolescent girls in schools and pregnant women linked with the health system. Non–school-going women of reproductive age and those not pregnant are neglected in most national and state programs, although they are listed as government priority populations. Research on this group is also scant: for example, the Indian National Family Health Survey collects data only for pregnant women [
Considering these insights, the RANI project has adopted a social norms–based approach for improving IFA uptake and adherence. The strategy includes the following: (1) participatory learning delivered through in-person activities and games; (2) a series of tablet- and computer-based health communication videos that target various audiences (pregnant women, nonpregnant women, husbands, and mothers-in-law), followed by discussion sessions; and (3) monthly community-based testing of hemoglobin levels of 15 women in each village, followed by a discussion about trends in anemia and village-level comparisons (based on the hemoglobin readings) with neighboring communities. A full protocol of the intervention and evaluation design is published elsewhere [
Intervention components, exposure, and sample size.
Group | Exposure | Components | Sample size, n |
RANIa control | Usual care | None (usual care) | 2000 |
RANI treatment, not in mRANIb | All RANI treatment components | Interactive learning sessions, RANI communication videos, and hemoglobin testing | 1700 |
RANI treatment+mRANI treatment (mIFAc) | All RANI treatment components+entertainment education and mRANI IFAd adherence arm | All components in RANI treatment that are not in mRANI+entertainment education story and mHealthe anemia intervention | 206 |
RANI treatment+mRANI control (mBIf) | All RANI treatment components+entertainment education and mRANI bystander intervention arm | All components in RANI treatment, not in mRANI+entertainment education story and mHealth bystander intervention | 205 |
aRANI: reduction in anemia through normative innovations.
bmRANI: mobile reduction in anemia through normative innovations.
cmIFA: mobile iron folic acid.
dIFA: iron folic acid.
emHealth: mobile health.
fmBI: mobile bystander intervention.
In the formative evaluation phase, the RANI project sought to identify key issues faced by women in Odisha that would shed light on their anemia-specific behaviors. A primary theme that emerged was the role of gender norms, which illustrated the multiple ways through which women experienced social, structural, and individual-level disadvantages and barriers [
Violence against women is a global problem, with more than 1 in 3 women worldwide reporting physical or sexual violence by an intimate partner or nonpartner [
Recent years have seen tremendous growth in both the scope and number of mobile health (mHealth) interventions worldwide, especially in low-resource settings [
India is one of the largest and fastest growing markets for information and communication technologies, with 1.2 billion phone and internet subscriptions, respectively [
Following these recommendations, the mobile RANI (mRANI) study will center its efforts on improving self-efficacy for IFA adherence as one of its primary goals. We have powered our mRANI study for this outcome as well. As described below, the control arm for the mRANI study is another intervention that focuses on reducing violence against women by promoting bystander intervention, which also uses self-efficacy as its outcome.
Our overall approach is anchored in the entertainment education (EE) literature [
As the research and practice of EE has evolved, so has the theoretical scholarship focused on understanding how it works [
Another component of the intervention we explore in this study is the role of interactivity in EE. We propose two competing mechanisms of change that we will test in this project. On the one hand, when the program promotes interactivity, user engagement with the content is expected to increase [
The objective of this study is to investigate the ability of a norms-based EE mHealth intervention in increasing self-efficacy for IFA adherence among women of reproductive age in Odisha, India. Our hypothesis is that women in the intervention arm will display greater positive improvements in self-efficacy, pre- to postintervention, than women in the control arm.
In addition to testing this hypothesis, we will also ask questions about the underlying process of change. In particular, our secondary aims are two-fold: to determine (1) the extent to which social norms and (2) interactivity mediate the relationship between intervention exposure and study outcomes.
The mRANI study is a randomized 2-arm study that includes assessments before and after the intervention. All study participants are recruited only from the intervention arm of the parent RANI trial (the trial itself has a usual care control group and a social norms–based treatment group, see the study by Yilma et al [
In this study, we introduce an innovative approach. Although the usual practice is to randomize participants either to a treatment arm or to a usual care control arm (this is also the design of the parent RANI trial), we will assign the mRANI control group to another EE-based treatment group designed to improve bystander intervention to reduce violence against women. We will call this the reciprocal control double treatment (RCDT) design. In this RCDT design, the treatment group for one intervention (mobile iron and folic adherence [mIFA] to improve IFA consumption self-efficacy) will serve as the control group for the other intervention (mobile bystander intervention [mBI] to improve bystander intervention self-efficacy), and vice versa.
The decision to center the second arm on bystander intervention (as opposed to another health issue) was based on two considerations. First, anemia and violence against women are empirically unrelated [
Taking this RCDT design into account, the hypothesis we proposed earlier will be tested by comparing changes from baseline to end line in self-efficacy to take IFA tablets in the treatment arm with the corresponding change in bystander self-efficacy to intervene in the control arm (and vice versa).
The RANI project is currently underway in 2 blocks (administrative units below the district) in the Angul district of Odisha, India [
The mRANI intervention was developed based on the literature as well as quantitative analyses conducted using the RANI study’s formative and baseline surveys. Using these data, we developed a mobile ownership and user profile and identified an optimal intervention delivery channel. We first conducted a feasibility analysis to determine whether a phone-based intervention would be viable in our study sites. Our data from baseline assessments (n=1874) of the parent RANI trial revealed that 43.97% (824/1874) of women owned mobile phones, and 42.63% (799/1874) of those who did not outright own a mobile phone could borrow one from a close family member with permission. Only 13.39% (250/1874) of our study participants were classified as those without access to a mobile phone. Accounting for these structural limitations, to be eligible for the mRANI study, participants have to be mobile phone owners with an active phone line.
We also conducted formative research in the form of qualitative story testing to select an overall storyline for both arms. Once the overall storyline was finalized, we developed subplots for each arm, integrated them into the broader storyline, and translated them into scripts. Finally, we conceptualized and developed interactive components. Audio recording and production of both EE) programs took place as a final step.
EE programs will be delivered with 13 episodes over 5 weeks, each lasting approximately 3-5 minutes. A total of 3 new episodes will be released every week. The exact day and time of delivery will be personalized according to participant preferences. Episodes will be delivered using an interactive voice response (IVR) system that can provide individualized messages to participants and interact with them through voice recognition or a touch-tone keypad. Prior studies have demonstrated the effectiveness of the IVR system in mHealth interventions, especially in rural settings with populations with limited educational and digital literacy [
The overall story was kept similar across both mRANI arms—that is, the mIFA arm and the mBI arm. However, we inserted messages about either taking IFA tablets (in the mIFA arm) or intervening to prevent violence as a bystander (mBI arm) at strategically chosen points in the storyline.
Interactive components for the mRANI study have been conceptualized at 3 levels: program-driven, audience-driven, and responsive interactions (
Entertainment education (EE) program release announcements
EE program jingles
EE program episode teasers
EE program episode recaps
Like or replay EE episodes on an interactive voice response system
Call a mobile reduction in anemia through normative innovations hotline
Quizzes and questions on the following:
Topics related to prosocial themes
Character identification
Engagement with the narrative
Agreement with major plot points
Satisfaction with the narrative
Perception of the most important problems in the village
We calculated the required sample size based on the findings of Constantino et al [
All women of reproductive age who meet the inclusion criteria for the larger RANI trial will be eligible to participate in the mRANI trial. For the parent RANI trial, inclusion criteria included being a woman aged between 15 and 49 years, speaking Odiya, and residing in our study sites. In addition, for the mRANI trial subsample, we require participants to be mobile phone owners with an active phone line. Before the start of the study, data collectors tested the phone numbers provided by participants to investigate which phone lines were still active.
Upon confirmation of the operating phone numbers (n=411), we listed eligible women in random order using a random number generator. We selected every other woman, starting with the first woman, for the IFA arm (206/411, 50.1%); similarly, we selected every other woman, starting with the second woman, for the violence against women arm (205/411, 49.9%).
The primary evaluation outcome for the mIFA arm is self-efficacy for IFA adherence. IFA use and anemia among women of reproductive age will also be measured as secondary outcomes. The primary evaluation outcome for the mBI arm is bystander self-efficacy in preventing violence against women.
The mRANI study data collection will be embedded within the overall RANI trial, similar to its intervention delivery strategy. Women between the ages of 15 and 49 years residing in treatment (n=2061) and control (n=2049) clusters have been recruited and randomly selected to participate in the impact evaluation of the RANI project. Baseline survey of the RANI participants took place in September 2019, and an end line survey was scheduled to take place in February 2021. Baseline data collection for the larger RANI project served as formative data toward developing the mRANI EE programs and segmenting potential participants. The study team will conduct the impact evaluation for the mRANI study using data from end line assessments, controlling for baseline estimates.
The average time to complete a survey during the baseline and midline data collection was 55 minutes for all RANI measures and outcomes. The mRANI-specific modules take an average of 15 minutes to complete. Translation and back translation of the entire survey instrument, including mRANI-specific modules, will be conducted to check the accuracy of the questionnaire’s translation. Pretesting of mRANI-specific modules will be conducted in nonsampled villages to ascertain reliability and appropriateness in this intervention context. A description of the data collection methods for the parent trial can be found in the parent trial’s protocol [
We will collect demographic information from the RANI sample, including age, education, number of children, and caste or tribe membership.
Self-efficacy in adhering to an IFA regimen was conceptualized as the extent to which women feel confident that they can continuously take IFA when facing a number of relevant barriers. At baseline, we asked the participants the extent to which they agreed they (1) could take IFA every week when not pregnant, (2) believed they could easily take IFA, (3) could take IFA even if their husband or father does not want them to, and (4) could take IFA even if their mother-in-law or mother does not want them to. Responses were coded on a 5-point Likert scale and averaged into a scale for self-efficacy for adherence at baseline.
We will measure self-efficacy for adherence at the end line using a modified scale that includes additional barriers for adherence. Participants will be asked the extent to which they are confident they (1) will continue to take IFA every day (for pregnant women) and every week (for nonpregnant women), (2) are anemic, (3) are not anemic, and (4) receive advice from others to discontinue taking IFA. Barriers were identified during the formative research phase of the RANI project [
The RANI project operationalizes three forms of social norms: descriptive, injunctive, and collective. This study conceptualizes IFA-related descriptive norms as the extent to which women believe others regularly consume IFA. At baseline, women were asked, “what proportion of a) non-pregnant, b) pregnant and c) adolescent girls take IFA regularly?” Responses were recorded as
IFA-related injunctive norms were conceptualized as the extent to which women believe others expect them to consume IFA regularly. At baseline, women were asked, “how many women in your community think you should take IFA tablets regularly if you are pregnant?” as well as if they are not pregnant. Responses were recorded on the same scale as descriptive norms, ranging from
Previous studies have conceptualized collective norms as the prevalence of a focal behavior within a community [
This study conceptualizes bystander self-efficacy as the extent to which women feel confident in their ability to intervene when they witness another woman experiencing violence. At the end line, we will ask women the degree to which they agree with 8 statements stating, “If I witness violence against a woman, I am confident I could intervene by distracting (e.g. beating pots and pans); delaying (e.g. checking in after witnessing violence with the victim); delegating (e.g. seeking help from SHG group members or influential community leaders); documenting the violence (e.g. calling a hotline or recording on a cell phone) as a bystander.” Responses will be scored on a 5-point Likert scale, ranging from
Violence against women–related descriptive norms were conceptualized as the extent to which women believe that others intervene when witnessing violence in their community. We will measure descriptive norms by asking participants to report how many women in their community and friends would intervene if they saw a woman experiencing violence. Responses will be recorded as
Violence against women–related injunctive norms were conceptualized as the extent to which women believe that others expect them to intervene if they witness a woman experiencing violence. This study will operationalize these injunctive norms by asking participants the extent to which they agree with the following statement: “If you saw a woman experiencing violence, [referent] will expect you to intervene,” in which the referent will be (1) most women in the community, (2) most friends, and (3) most women in self-help groups. We will record responses on a 5-point Likert scale ranging from
The operationalization of violence against women–related collective norms will mirror the mIFA arm, except that we will use intentions to engage in bystander intervention strategies in lieu of behaviors [
Prior EE interventions leveraging IVR systems have reported success in their ability to facilitate and track audience engagement in real time and at scale [
Number of calls sent
Number of calls received
Number of entertainment education (EE) episodes heard in full by the mobile reduction in anemia through normative innovations (mRANI) participants
Number of interactions with interactive components
Number of likes on episodes
Number of replays of episodes
Number of calls and queries on the mRANI hotline
Number of referrals to community and clinical linkages
Number of seconds spent listening to episodes
Number of seconds spent relistening to episodes
Number of seconds spent listening to jingles
Number of seconds spent listening to episode teasers
Number of seconds spent listening to episode recaps
Percentage of participants who accurately recall prosocial themes addressed in the episodes
Percentage of participants who accurately recall mRANI characters
Percentage of participants who accurately recall major mRANI plot points
Percentage of participants engaged with the narrative
Percentage of participants satisfied with the narrative
Percentage of participants who identify with Malati (protagonist) or Dolly (protagonist’s daughter)
Percentage of participants who identify anemia or violence against women (depending on their assignment) as the most important problem at the end of the mRANI EE program
The full RANI trial, including the mHealth intervention, has been reviewed and approved by appropriate institutions in the United States and India. These institutions include the George Washington University institutional review board in the United States; the Sigma Science and Research in New Delhi, India; and the Indian Council for Medical Research’s Health Ministry’s Screening Committee. The trial has also been registered with the Clinical Trial Registry of India.
Participants will be contacted with preintervention delivery for validation of phone numbers, followed by orientation and enrollment. The orientation and enrollment activities will take place a week before the mRANI intervention delivery commences. A structured script will be prepared to orient potential participants to mRANI and inform them of the objectives and intervention delivery channel and duration. For participants who express interest in enrolling in the mRANI study, we will obtain verbal informed consent. Next, we will note their preferences for the day and time to receive the mRANI EE programs. Participants will also receive contact information for additional questions and feedback.
Given the short duration and digital delivery format of this intervention, we anticipate low levels of exposure to adverse events for mRANI participants. In addition, considering that the mRANI educational messages are strategically masked within an entertaining narrative tested in the field, we expect minimal sensitivity and negative reactions from our audience or their family members. However, the study team acknowledges that violence against women is a sensitive subject matter in Angul, Odisha. The mRANI study will adhere to the risk mitigation plan in place for the larger RANI trial [
The primary analyses planned for the mRANI impact evaluation will test the hypotheses that participants assigned to the treatment arm will have higher levels of self-efficacy in IFA adherence (primary end point) and self-reported IFA use (secondary end point) at follow-up and that participants assigned to the control arm will have higher levels of bystander self-efficacy for prevention of violence against women. We will test these three hypotheses in three ways. First, because participants are randomized to treatment and control groups, a simple comparison of those groups should produce an unbiased estimate of the treatment effect. Therefore, we will estimate, for each end point, the linear regression model
Although these analyses should produce unbiased estimates of treatment effects, they may miss key opportunities to increase statistical power and precision through two analytic elaborations. The first is to introduce cluster fixed effects to the model, which effectively results in a within-cluster comparison of the treatment and control group participants. As treatment assignment is within clusters, cluster and treatment assignments are orthogonal, so the addition of cluster fixed effects should increase statistical power and precision by reducing the residual error variance. In this case, the analytic model for each end point will be
Another opportunity to improve statistical power while retaining unbiased estimation is to control for pretreatment variables that are likely to be highly correlated with the primary and secondary end points. These include measures of the same variables taken before randomization: baseline and midline self-efficacy for serum hemoglobin levels, IFA adherence, and IFA use. Note that bystander self-efficacy for violence against women was not measured at baseline or midline and thus cannot be included as a covariate. In addition, sociodemographic variables assessed at baseline and midline, such as age, education, and parity, can also be considered for inclusion. We will choose covariates for inclusion based on their marginal associations with each end point. The resulting model for each end point will be
This approach will produce three estimates of the treatment effect for each end point: b1, g1, and d1. All are unbiased, and we expect them to be similar across the three specifications. The main difference we expect would be in the SEs, which we expect will be highest in the first and lowest in the third specification. For transparency, we will report the estimates from all three specifications for each end point.
To assess the extent to which social norms and interactivity mediate the relationship between intervention exposure and primary outcomes in both mIFA and mBI arms, we will use a structural equation model to estimate the effects of the intervention on the mediator variables, the effects of the mediators on the 2 self-efficacy variables (IFA use and bystander intervention), and the direct effect of the treatment arm on both self-efficacy variables. From this model, we will estimate the percentage of the intervention effect on each aspect of self-efficacy mediated by each of the mediators [
Data collection for the mRANI intervention is integrated within the parent RANI trial. Household surveys collected responses for the primary and secondary outcomes of the mRANI study between February and March 2021. End line data were collected from all 381 (192/381, 50.4% women in the IFA treatment arm and 189/381, 49.6% women in the violence against women’s control arm) study participants. Data analysis is expected to be completed by October 2021.
The primary aims of the mRANI intervention are three-fold. First, we wish to determine whether an intervention delivered through a rather thin medium, an IVR channel, can propel behavior change. By relying on a social norms–based theory [
The second aim of the study is to determine whether embedding an mHealth trial within the intervention arm of a larger field trial can further add to the effects of the parent trial. We will use these data to inform the next iteration of the RANI intervention. Finally, we also wish to determine whether the reciprocal control double intervention design can yield meaningful results.
This study does have a few noteworthy limitations. First, the generalizability of our study is confined to women who are cell phone owners. Although ownership is increasing rapidly in India [
In spring 2021, at the end of the RANI evaluation, we will hold a virtual convening in Bhubaneswar, India (the capital of Odisha), to present the mRANI findings along with the main trial findings. We will invite key stakeholders from Angul, such as district officials, program planners, researchers, and policy makers working on anemia reduction efforts in Odisha. We will also disseminate findings locally back to the communities where the intervention will take place through smaller presentations and dissemination materials.
To our knowledge, an mHealth intervention has not been used to reduce anemia among nonpregnant women, the subpopulation that makes up the largest number of women with anemia [
Finally, although the inception of this mRANI intervention predated the COVID-19 pandemic, the need to reach women via mobile phones rather than in-person settings is critical now more than ever. Given the uncertainty of when and how in-person interventions will unfold, contactless interventions are critical to be able to continue providing information to communities staying home. Furthermore, violence against women was rampant before the COVID-19 pandemic; however, research shows that the incidence is rising worldwide and specifically in India [
Extended details on story testing.
Overall mobile reduction in anemia through normative innovations entertainment education narrative.
Episode summaries for both mobile reduction in anemia through normative innovations arms.
entertainment education
iron folic acid
interactive voice response
mobile bystander intervention
mobile health
mobile iron and folic adherence
mobile reduction in anemia through normative innovations
reduction in anemia through normative innovations
reciprocal control double treatment
The authors would like to acknowledge Dr Peter Orton for his instrumental support in translating our story ideas into vivid entertaining scripts. The authors are also grateful for the thoughtful feedback and input provided by Dr Michael Long. This work was supported by a grant from the Bill and Melinda Gates Foundation (grant OPP1182519) to the George Washington University (principal investigator: RNR). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
IP and RNR were responsible for the conception of the study, its objectives, the design, the rationale, the development of research questions, and the drafting and revising of the protocol. IP also provided digital health and monitoring and evaluation expertise. SB contributed to intervention delivery and development, data collection, recruitment, implementation, and risk mitigation procedures. HY contributed to the sampling plan, measures, and outcomes procedures. JBB contributed to statistical analysis procedures. ES contributed to the
None declared.