AI-Supported Sexual and Reproductive Health Information and Care-Seeking among Women in India

In this study, women are provided access to a digital health-support tool focused on sexual and reproductive health, delivered through mobile phones. The core intervention is Myna Bolo, an AI-enabled conversational tool that allows women to ask questions and receive responses in a private, anonymous, and on-demand format. The tool is designed to lower barriers to asking sensitive questions by allowing women to seek guidance at any time, without needing to immediately approach a family member, provider, or in-person service.

Depending on study assignment, women may receive different forms of support built around this core tool, including informational messages, access to an interactive text-based chatbot, and/or access to a more proactive voice-based AI system. The intervention is designed to study how privacy, anonymity, responsiveness, conversational interaction, and modality shape women’s information-seeking, support-seeking, and care-seeking behavior in the context of sexual and reproductive health.

2026-03-03

2026-09-30

The primary outcomes are:
(1) engagement and health-related information-seeking;
(2) health knowledge and symptom interpretation; and
(3) support-seeking and healthcare utilization.

Primary outcomes will be measured using a combination of survey-based measures and administrative data from the intervention platforms.

The first primary outcome family, engagement and health-related information-seeking, will include measures such as whether the participant used the tool where applicable, intensity or frequency of engagement, and whether the participant sought information or guidance for a health concern during the study period.

The second primary outcome family, health knowledge and symptom interpretation, will be constructed from survey questions assessing knowledge of women’s health topics, correction of common misperceptions, recognition or interpretation of symptoms, and understanding of whether and when additional support or care may be needed. Where appropriate, related variables may be standardized and combined into composite indices.

The third primary outcome family, support-seeking and healthcare utilization, will include measures of whether the participant took action in response to a health concern, including use of formal healthcare, telehealth, informal support, online support, or other problem-solving behavior. This family may include indicators such as any attempt to solve a health concern, any formal care use, and specific care channels where applicable. Where multiple measures reflect a common underlying construct, variables may be standardized and combined into summary indices.

Higher values for constructed indices will be coded to reflect more engagement, greater knowledge or more accurate symptom interpretation, and more support-seeking or healthcare utilization. This follows the design documents, which organize outcomes around utilization, health-seeking, knowledge/misperceptions, and related behavioral responses.

Secondary outcomes include attitudes, stigma, and norms related to women’s health; comfort discussing health concerns; perceived privacy and trust in the support channel; agency-related measures; health expenditure and time costs; telehealth take-up; informal care use; condition-specific health issue measures; hemoglobin and anemia indicators; and IFA / iron-pill compliance.

Secondary outcomes will be measured using survey responses, administrative platform data, and, where relevant, linked telehealth or referral records. Some outcomes may be analyzed as individual variables, while others may be standardized and aggregated into summary indices.

Attitudes, stigma, and norms outcomes will be based on pre-specified survey items related to menstrual myths, reproductive-health beliefs, and related social perceptions. Agency-related outcomes will capture whether and how women take action around health concerns and other relevant survey-based dimensions of decision-making or locus of control. Informal care and expenditure outcomes will capture whether women rely on home remedies, chemists, or social networks, and the time or money spent addressing health concerns.

Hemoglobin testing will be treated as an objective secondary outcome measured across arms where feasible, and IFA compliance will be measured with a standardized compliance threshold where appropriate.

Participants in this study are randomly assigned to receive different forms of access to an AI-based conversational support tool in the context of women’s sexual and reproductive health in urban India. The tool allows users to ask questions and receive responses through a private, anonymous, and always-available messaging-based interface. Some study arms also receive additional forms of support alongside access to the tool. The intervention is designed to study how AI-based tools affect information-seeking, symptom interpretation, and care-seeking behavior in a setting where privacy, stigma, and convenience may be especially important.

Not available

Randomization is done by computer at the individual participant level, with eligible women assigned in equal proportions across four study arms before the baseline or activation of any intervention arm. Randomization is done on the listing sample collected before the baseline study.

Individual woman / individual participant.

No

NA

3,000-4,000 women.

Around 4,000 women total, allocated equally across four arms: 1,000 control; 1,000 notifications only; 1,000 notifications + text chatbot; 1,000 notifications + text chatbot + proactive voice AI.