Khushi Baby

50 out of 1000 children in the rural District of Udaipur, India do not see their first birthday. A combination of maternal- and child- health and nutrition contribute to a child's early development - the foundation for longterm health and prosperity. Current health systems fail to capture accountable, longitudinal health records at the last mile. Khushi Baby has developed a decentralized health record that stays with the mother/child from family planning, through antenatal care, through postnatal care and child care. This biometrically-secure health record lives inside a culturally-symbolic amulet in a $0.70, battery-less, NFC chip With the Khushi Baby App, frontline healthworkers scan the chip to review/update medical history, see AI-guided high-risk alerts, and provide early interventions  to impact clinical and developmental outcomes. Automated, personalized voice call reminders are sent to families to ensure timely follow-up and uptake of these services. .

50 out of 1000 children in the rural District of Udaipur, India do not see their first birthday. 11.8% of children under 5 are severely malnourished by WHO standards and less than 5% of children between 6 and 23 months of age receive an adequate diet per WHO standards. Addressing the root of poor maternal- and child- health and nutrition is a starting point for any aspirational gains in early child development. Existing solutions to track key indicators during this phase have failed in last-mile, dfor two key reasons: lack of connectivity and lack of supervision. Other mHealth approaches to track data from one antenatal care visit to the next, depend on health workers being able to synchronize data collected on their mobile devices with a cloud solution. The ground reality is that patients move, health workers move, devices change hands, and internet connectivity is not reliable. As a result data is duplicated and longitudinal medical history is broken. Second, lack of supervision leads to deliberate data misentry for the purpose of achieving "targets". MHealth solutions cannot claim such tampered data can be use for AI-based, predictive modeling of future developmental milestones

We serve mothers and children living in 1000 rural villages of Udaipur, India. This population lives on less than 2 USD per day. Mothers on average have a second-grade education. Literacy rates are less than 20% and most beneficiaries are daily wage earners with agrarian lifestyles. Over the past 5 years, our full-time team has spent time directly with the population we are serving and the government heatlh workers that provide access to maternal and child health care and nutrition services. This user-centered ethos was borne out of a Yale-course from which Khushi Baby was derived: Culturally-appropriate design for the Developing World. Our solution has two key features co-developed with the people we serve: a) the health record comes in the form of a culturally-symbolic black thread, already worn by members of the community to ward off the evil spirit; b) our automated voice call reminders are sent in the specific local dialect to family's mobile phones.

Mothers and infants receive the Khushi Baby Pendant - a $0.70, battery-less, culturally-symbolic amulet which contains a Near Field Communication Chip in a durable, waterproof casing. This pendant stores over 150 columns of longitudinal health data from pregnancy through the child's 5th year along with the caregiver's biometric in 1KB of space. Health workers scan this Pendant with an Android mobile application on their smartphones - allowing them to review the health history, see what treatments/interventions are due, see any alerts, and mark what services were provided for the visit. This data is udpated onto the patient's Pendant and travels with the patient to any subsequent visit at any health center. Health workers eventually upsync data collected to a cloud database which provides analytics to health officials via a mobile dashboard. The dashboard allows health officials to schedule personalized voice call reminders to specific patient groups. They can further send key reports to WhatsApp groups of frontline health worker teams to spark specific follow-up. Provided that we are tracking the spectrum of health from family planning onwards, we see a tremendous potential in using early signals in the data to both predict and prevent future early childhood milestone delays. With 25,000 mothers and infants already being followed-up in our system, and another 50,000 projected beneficiaries to be tracked in the coming year, we are developing AI-guided models to best predict future clinical, nutritional, and developmental outcomes. And we are using similar models to find the best mix of interventions to minimize adverse outcomes.

Decentralization - patients move, health workers move, and mobile devices are not always synchronized with a central server. The Khushi Baby health record uses a Near Field Communication chip (50 INR) which stores the patient's entire health history. The health worker can scan the chip to view the history. The mobile application indicates any due action items. The health worker updates the status in the mobile app and then scans the chip to save the updated medical history. This entire interaction requires no prior synchronization and no internet at the point of care to identify the patient or retrieve their record.

Data Accountability - in order for a patient encounter to be recorded, the health worker must first sign-in to the app which records GPS and time-stamp. Then the health worker needs to scan the patient’s health card with their mobile device. The mobile app then prompts the health worker to scan the patient’s thumbprint to confirm to authenticate the record. The encounter details are saved on the patients card which serves as a digital, secure, unfalsifiable proof that the patient was seen. A set of unique indicators such as the time per patient encounter and variance of particular indicators isused to determine the probability of fraudulent data entry. In addition to detecting data fraud, the mobile app is designed to gamify high quality data entry.

We use Near Field Communication technology to store longitudinal medical records within a compressed amount of memory. Health workers use an Android mobile application to scan this NFC chip (the patient's health record) to review data stored on the chip. This medical record is unlocked after comparing the live thumbprint of the patient against the biometric template stored within the NFC chip. The medical record contains 150+ columns of data is viewed and updated with the Android application, before being saved onto the NFC chip at the end of the encounter. Later the updated records are also synchronized with a cloud database. An automated voice and WhatsApp messaging service is used to schedule messages to specific patient groups. These messages are informed through AI-driven models which are trained to predict and prevent adverse maternal and child health outcomes.

Data for Action - systems that simply collect data for reporting are not currently structured to incentivize improvement in outcomes. The Khushi Baby platform has been designed to empower and educate frontline health workers through thoughtful user experience design. Frontline health workers are alerted to key actions they can take through standardized protocols displayed in the mobile app. High risk reports are automated and shared on WhatsApp groups of health worker teams to drive improved coordination and follow-up. Patients receive individualized voice-based reminders in their preferred dialect. Health officials can call team leaders directly from the mobile dashboard to follow-up on specific performance metrics.

Our system tracks the health of 25,000 mothers and infants across 400 rural villages in the Udaipur. In the next 12 months we will reach 320 health workers tracking 80,000 mothers and infants across 1000 villages. In the next three years, our goal is to become a standard for the state of Rajasthan and for this solution to be scaled to 17,000 health workers tracking the health of 1.5 million pregnancies every year. At minimum we aspire to be fully deployed in 3 full districts (out of the 33 in the districts in the state) and reach 700,000 mothers and children with the support of catalytic matching funding from the Gates Foundation.

Within the next year our platform will be scaled across the entire District of Udaipur in Rajasthan, India, where 50 out of 1000 children currently do not see their first birthday. In the next five years we aspire to be a universal standard for decentralized health records across multiple states in India. Our ambition to scale requires that we partner with the government and establish a standard for medical records that is based on the user experience between the patient and provider at the last-mile level. Only by using technology, can we transform this interface no matter where the patient goes to receive services. And only with an interface that is used by patients and providers can we begin to collect accountable data that can be used for long term outcome improvements for areas like early child development

1. Financial Capital to activate catalytic matching funding

2. Political Capital as an outsider to scale the model at the state level against other competing mobile health apps (that lack the fundamental health record paradigm we are advancing)

3. Recruiting Talent to the last mile

1. Approach CSRs, funding agencies, and local and state governments to buy in to the idea of a District pilot by visiting our field sites and seeing our innovation in action. We will also aggressively strive to hit our outcome targets to show that our innovation is not only a pilot, but a growing model at a sizable scale.

2. Universalize the system within one district to serve as the blueprint and model for scale-up across other districts in the state and across India

3. Obtain sufficient financial capital to pay a premium for the best talent to stay competitive with the best companies in the region. We need to acknowledge that mission-driven talent may not be prepared to move to rural Udaipur for example and will need to strategically balance opening satellite offices in major cities when capital is sufficient to continue our growth and innovation with the best talent.

Full-time: 25

Part-time: 5

We have an interdisciplinary team of designers, developers, public health practitioners, and data scientists that is focused on the end-to-end design, development, deployment, and monitoring of a comprehensive maternal and child health tracking solution. Half our team - comprised of our field monitors - is locally based in the villages where our beneficiaries and health workers live. We are not the best to deliver this solution because of our pedigree from world class institutions like UNICEF, JPAL South Asia, Harvard Medical School, and the Indian National Health Mission. We are the best placed to deliver this solution because we are firmly rooted in the communities we serve at the last mile and have the multitude of perspectives to not only present a paradigm shifting technology, but actually execute that idea in the most rugged environments.

Rajasthan State Ministry of Health and Udaipur District Government: government partner in implementation of the technology

UNICEF: financial support, access to broader networks and implementation sites

GAVI: financial and technical support, access to new government partners (within India and beyond) for collaboration and deployment of the technology at scale

Johnson and Johnson: financial and technical support, access to new implementation opportunities and collaborations

Harvard Medical School: research mentorship and financial support

Yale Global Health Leadership Institute: organizational mentorship and other technical support

NCORE Incubator: organizational, financial, and technical mentorship and support

We provide a platform as a service for comprehensive maternal and child health tracking to Ministries of Health. The government is the prime payer. The mothers and infants receiving enhanced services from schemes that implement a program that advances our technology platform are the beneficiaries. Our bundle includes the software and hardware technology solution along with the implementation vehicle to execute the program deployment at the grassroots level. We consider the government as a prime payer due to its ability to standardize this form of longitudinal health record data collection for over 70% of the population in its state. At the same time, we do approach hospital networks focused on maintaining a similar track of their patients within their systems, and even smaller private clinics  who are looking to stay connected with their patient population. Beyond our technology platform, our domain expertise in analytics in maternal child health also affords us opportunities to consult with large health providers who are servicing these populations. We also offer opportunities to providers in the education, insurance, and microfinance sectors who are looking to newly connect with new rural populations through vehicles such as our opt-in voice call reminder platform.

1. Pool together CSR and donor funds to activate 1.5M USD in matching funding from the Gates Foundation

2. Demonstrate at a District level, a model of the universalized platform being used to track every mother and infant's health

3. Use the evidence we have already collected from our 2 year Randomized Controlled Trial and the additional data we generate, to onboard new Districts in new States, leveraging a combination of local CSR and government funding to support an expanding footprint

4. Generate alternative streams of revenue by customizing similar solutions for maternity hospitals and other maternal child health providers (low hanging fruit) who collect similar data streams. Also engage with service providers looking to connect with our population base and charge commission for opt-in enrollment in education, insurance, and microfinance services.

5. Convince a major health provider to consult our services in design of a larger scale program

6. Convince a major health provider to subscribe to our model for rapid deployment and scale-up on a recurrent basis

7. Become a standard for decentralized, longitudinal health records for maternal and child health/nutrition, and begin to expand to new verticals such as new disease categories and education.

We are applying to Solve to get connected to a network of like-minded changemakers, seasoned technology scalers, creative impact investors, and new team members who can take Khushi Baby and make it a solution for not just 25,000 mothers and children, but for 10 if not 100x that scale. We are a young technology non-profit and we do not have all the answers. We have limited experience at building things at scale. But we have a bold vision for the future and strongly believe that our nugget of an innovation has a bold potential. We want to redefine the interface between provider and patient around the world and we belive a portable, accountable, digital health record touching a mobile phone with a user-centered interface can break down that barrier and unlock informed, longitudinal care.

We would like to partner with technology partners that have scaled solutions to millions of users in developing settings and who have talent who can help us architect our solution to be scale-ready. The time is critical for such a partnership to be forged as we are in the midst of scaling up our platform to the District level. We look to lean on the experience of others who have taken such technologies to the next level while allowing our core expertise - the local connection and grassroots innovation to be our forte. We also want to connect with a strategic investor who can allow us to hire the sales team to focus on scale-up while preserving our ability to make the best platform for maternal and child health in India.

To date the Khushi Baby system has tracked the health of 25,000 mothers and infants across 400 villages in Udaipur, India. Within the next year, the Khushi Baby system will be scaled to reach another 55,000 beneficiaries across 1000 villages in the district. The proposed project aims to utilize existing and forthcoming Khushi Baby data for machine-learned models to develop a precision approach to public health outreach for improved maternal and child health behaviors and outcomes - especially for early child development. The end goal is to establish a sensitive, specific, and real-time, high-risk score for each pregnant woman - by incorporating over 300 columns of high-quality data related to her current health status and behaviors, past health status and behaviors, and social determinants. One score, generated through machine-trained models, would be able to predict the probability of future clinical outcomes such as miscarriage, stillbirth, maternal death, neonatal death, infant malnutrition, and missed developmental milestones. These scores would be used to deduce a set of interventions (reminder messages, awareness messages, household visits) which best reduce the risk of adverse clinical outcomes. The novelty in this proposal comes from our ability to rigorously enforce data standards and  quantify data quality - through over 30 automated data quality metrics derived from our domain expertise in the field.

To date the Khushi Baby system has tracked the health of 25,000 mothers and infants across 400 villages in Udaipur, India. Within the next year, the Khushi Baby system will be scaled to reach another 55,000 beneficiaries across 1000 villages in the district. The proposed project aims to utilize existing and forthcoming Khushi Baby data for machine-learned models to develop a precision approach to public health outreach for improved maternal and child health behaviors and outcomes - especially for early child development. The end goal is to establish a sensitive, specific, and real-time, high-risk score for each pregnant woman - by incorporating over 300 columns of high-quality data related to her current health status and behaviors, past health status and behaviors, and social determinants. One score, generated through machine-trained models, would be able to predict the probability of future clinical outcomes such as miscarriage, stillbirth, maternal death, neonatal death, infant malnutrition, and missed developmental milestones. These scores would be used to deduce a set of interventions (reminder messages, awareness messages, household visits) which best reduce the risk of adverse clinical outcomes. The novelty in this proposal comes from our ability to rigorously enforce data standards and  quantify data quality - through over 30 automated data quality metrics derived from our domain expertise in the field.

To date the Khushi Baby system has tracked the health of 25,000 mothers and infants across 400 villages in Udaipur, India. Within the next year, the Khushi Baby system will be scaled to reach another 55,000 beneficiaries across 1000 villages in the district. The proposed project aims to utilize existing and forthcoming Khushi Baby data for machine-learned models to develop a precision approach to public health outreach for improved maternal and child health behaviors and outcomes - especially for early child development. The end goal is to establish a sensitive, specific, and real-time, high-risk score for each pregnant woman - by incorporating over 300 columns of high-quality data related to her current health status and behaviors, past health status and behaviors, and social determinants. One score, generated through machine-trained models, would be able to predict the probability of future clinical outcomes such as miscarriage, stillbirth, maternal death, neonatal death, infant malnutrition, and missed developmental milestones. These scores would be used to deduce a set of interventions (reminder messages, awareness messages, household visits) which best reduce the risk of adverse clinical outcomes. The novelty in this proposal comes from our ability to rigorously enforce data standards and  quantify data quality - through over 30 automated data quality metrics derived from our domain expertise in the field.