RATI - Reduction of Anemia through Technological Innovation

Anemia is a significant public health concern in India, affecting a significant proportion of the population, particularly women and children. According to the National Family Health Survey (NFHS) 5, an estimated 57% of women and 67% of children under five years of age in India suffer from anemia. A staggering number, especially in a country where about 10% of its population is under the age of 5 (Census 2011). This high prevalence of anemia has far-reaching impacts on the health, productivity, and well-being of individuals and communities.

To address this issue, India launched the Anemia Control Program, also known as Anemia Mukt Bharat, in 2018. The program aims to tackle fundamental challenges related to program financing, implementation, and last-mile delivery. The strategy targets six groups of beneficiaries with focused interventions, including prophylactic supplementation with IFA tablets, which are essential for treating anemia and preventing its recurrence.

However, the success of the program depends on the availability and accessibility of IFA tablets, which remains a significant challenge. Critical in preventing anemia and its recurrence, IFA consumption remains an essential challenge among all age groups. The consumption of IFA for the recommended 180 days is an abysmal 26%, and for 100 days, it is 44%. On the surface, the problem is directed toward poor adherence to the tablet among the beneficiaries. However, as we delve deeper into the issue, we can identify several key challenges and focus areas of AMB guidelines that contribute to the low distribution of IFA tablets among the beneficiaries.

Studies have identified both demand and supply-side challenges that affect the poor coverage of tablets. Demand-side challenges include poor knowledge about anemia, a lack of importance for tablet consumption, and forgetfulness, among others. Supply-side deficiencies are mainly linked to poor supply chain management at multiple levels of the health system. Both of these factors end up leaving a sizable chunk of the population without coverage.

To overcome these challenges, it is essential to adopt a multi-pronged approach that addresses both the demand and supply-side challenges. On the demand side, there is a need to increase awareness about the importance of IFA consumption and the consequences of anemia.

On the supply side, there is a need to strengthen the health system's supply chain management to ensure a steady supply of IFA tablets to the beneficiaries. This can be achieved through the adoption of innovative technology solutions that can track, monitor, and assist in the last-mile delivery of IFA tablets. 

RATI Bot was developed with the idea that, with the increasing penetration of smartphones in rural areas of India, each household has access to at least one smartphone. It provides us with a unique opportunity to engage directly with beneficiaries and front-line workers by providing them with easy-to-understand and quality IEC materials discussing anemia and its management, with a focus on calcium and folic acid. With the idea of strengthening the supply side, the bot, apart from knowledge dissemination, was also aimed at providing easy tools for calculating the required quantity of IFA, calcium, and folic acid for the FLWs.

WhatsApp based, AI enabled, interactive BOT in vernacular languages, with the following key features:

1.Personalized learning on key concepts in anemia prevention and management.

2.Personalized schedulers and reminders

3.Information on government and civil society efforts and schemes

4.Quizzes and gamification

RATI, through the WhatsApp API, has a two-way interaction of request and response with the server. This server, based on need, can be integrated with other data sources like HMIS, RCH Portal, E-Aushadhi, and ABDM-E Hospital. Utilizing the data, RATI engages with different target beneficiaries, providing them with curated content and also providing tools for informed decision-making.

RATI has been piloted in one district each of Madhya Pradesh and Gujarat in partnership with Nutrition International.

The target population is basically split into 2 categories :

• Beneficiaries
• Health workers

Beneficiaries: With the increasing penetration of smartphones in rural areas in India and each household having access to at least one smartphone, it provides us with a unique opportunity to engage directly with beneficiaries.

Beneficiary categories include

• Pregnant women
• Lactating women
• Women in the reproductive age group(18-49 years of age)

Anemia is a huge consideration among this group of beneficiaries. They are also a focus group for government policies and programs on anemia. Social norms and environmental context often leave these groups underserved and with poor access to information and knowledge.

Health worker:

Front line health workers are the backbone of the healthcare delivery system in the country. Healthcare workers focused by RATI include

• Front line workers like Accredited Social Health Activist (ASHA), Auxiliary Nurse Midwife(ANM), and Anganwadi Workers(AWW)
• Supply chain managers like pharmacists
• Key decision makers like block and district managers and medical officers

Though there is an influx of data generated from the ground level, these decision-makers and healthcare managers are often shorthanded on the availability of readable and usable data that can be used to make informed decisions. Further, these groups of key persons are often lost in multiple reporting applications and formats that are standalone and not interoperable.

Further, RATI provides flexibility in the addition of more categories of users based on need.

The Tattva Foundation has been working in the space of digital health since 2014. Tattva Foundation has been closely working with the health systems in different states of the country (India), like Uttar Pradesh, Madhya Pradesh, Gujarat, and Nagaland, among others, and is well-positioned to understand the needs and challenges faced on the ground.

RATI has been developed after a detailed assessment of the challenges in the IFA supply chain in two states of India, namely Gujarat and Madhya Pradesh. Over the course of a few weeks, we carried out a rapid assessment through interactions with the beneficiaries, front-line workers, and healthcare managers to understand the issues, bottlenecks, and challenges faced in the delivery and consumption of iron-folic acid tablets. Informed by the ideas and needs from the communities, RATI was piloted in two districts in both states. Based on the feedback, iterations have been made to the existing model.

Rati currently serves 5,500 users, including beneficiaries and healthcare workers.

We would like to benefit from the MIT Solve community’s learning and guidance to:

• Improve and accelerate a perception change towards the culture of data use among grassroot level communities and organizations such as FLWs.
• Attain a wider acceptance for the solution and trigger conversations regarding it among global stakeholders.
• Overcome the barriers of information asymmetry existing in such systems by  making information accessible to all stakeholders in the program alike for more harmonious supply chain improvement efforts.
• Financial and operational partnerships for scale-up and the next phase of development of the project

With such strategic partnerships, we intend to conduct the following activities:

• Conduct design science research to measure the  efficacy of the model in improving consumption of iron folic acid tablets by targeting both demand and supply side factors.
• Conduct a cost-benefit analysis for the scalability of the solution in low- and middle-income countries. Run a full-fledged model for a period of one year to demonstrate the effectiveness of the solution and take appropriate course correction measures.

The following innovative aspects of our solution make it a novel proposition:

• Theory-informed, replicable, and data-agnostic nature of the design proposition, thus encouraging the integration of other data sources and uptake by other national and global programs.
• The proposed system seeks to provide easy access to information and actionable nudges based on their current behavior for the beneficiaries at their fingertips and provides tools that promote adherence to the tablets. Also, the system aims to strengthen the supply chain of IFA tablets by ensuring the linking of different key stakeholders and data systems already existing in the system.
• The granularity of the solution enables customized and tailored messaging suited to the needs of the user.
• The summarization capability allows for an eagle-eye view of the retrieved information at state and national levels for improved planning, efficient resource allocation, and focused feedback.
• Due to the novelty of the proposed solution and its capability to leverage any existing proxy data sources, our proposition does not increase the burden of data collection on FLWs or other stakeholders in the program. 
• Moreover, supplying the stakeholders with readily available and timely information can lead to improved feedback, supervision, and management, ultimately creating a multiplier effect on increased population health.

We have planned our impact goals for the products in two phases:

Short term

1. Improve the adoption rates and the intent to use the system for a long time among the end users of the platform. For the success of any digital system, it is very important to create a positive perception among the end users towards the usefulness of the system, which should prompt them to have the intrinsic intention to use the system for a long time. Our design approach fundamentally takes into account the expectations and capabilities of the stakeholders during the design process itself. With respect to this, in order to increase the ease of use of the system, the proposed information will be delivered through multiple user-friendly llow-resourcechannels, thereby increasing its access. Moreover, we will be conducting ccapacity-buildingactivities and workshops to overcome the resistance and inculcate a positive perception of the system.

2. Improve the levels of monitoring and positive feedback with respect to IFA distribution and consumption.

The immediate objective of the system is to utilize the existing government-approved messaging to provide the right information at the right time. Apart from that, integration with existing data management systems will enable accurate tracing of the supplies from the state to service delivery points to ensure timely identification of bottlenecks in the supply chain. In order to leverage the platform efficiently, we will hold initial review meetings and demonstrate frameworks and pathways for retrieving and utilizing the right information from the platform. We will also conduct activities with FLWs to improve their capacity to self- monitor and optimize their performance.

Long term

• Improve the demand generation of iron folic acid tablets among the targeted beneficiaries.
• Improve the supply of iron and folic acids to the beneficiaries by reducing the delays, defects and stockouts due to poor stock monitoring.
• Increase the uptake and adoption of the solution at national and global levels through

The system will undergo constant evaluation and course correction to periodically improve it and align its core features with the changing needs and sentiments of users.

- Journal Publications

- Partnerships with global players in anemia control programs

The various indicators we are using to measure our progress toward the impact goals are:

1. Adoption rate and intent to use (to signify information system success)

2. Availability: Regular stock level monitoring to assist in the prevention of stockouts and timely distribution of IFA tablets to recipients

3. The distribution of IFA tablets must be tracked from the central level to the last-mile distribution locations.

4. Adherence is measured by the percentage of beneficiaries who take the recommended dosage of IFA tablets for the approved time.

In order to measure these indicators, we will be conducting multiple field tests, observational studies, and group discussions among the stakeholders.

Our system utilizes the BG Fogg behavioral model as the basis of creating the change.

The BJ Fogg model is built on three major components: Motivation, Ability, and Trigger. When these three factors converge at the same moment, behavior changes.

According to this paradigm:

Behavior (B) = Motivation(M)*Ability (A)*Prompts(P)

The model states that people need the motivation to change their behavior. But they also require the ability to do the desired behavior. In absence of either one of them, it is not possible to elicit a behavior change.

RATI comes in at this precise point where it acts as a medium to improve the motivation of the beneficiaries by

• Providing them tailored and actionable prompts based on their current behavioural patterns in simple, easy-to-understand manner
• Providing them with timely and regular nudges
• Providing with the choice of interaction and type.

At the same time, it improves the ability of health workers by

• Providing them tools for better calculation of supplies of IFA tablets (supply chain management)
• Linking various levels of health system for easy monitoring and review
• Providing information curated to needs of health workers.

Improving both beneficiary motivation and health worker ability puts the users at the threshold which triggers the behavior change.

The core technologies that power our WhatsApp chatbot are WhatsApp Business APIs, chatbot host servers, and dialog management. Apart from this, we are also incorporating artificial intelligence and machine learning into it.

The architecture of our WhatsApp chatbot typically follows a client-server model, with the client being the WhatsApp platform and the server hosting the chatbot's backend infrastructure. Users interact with the chatbot through the WhatsApp application on their mobile devices. The WhatsApp platform handles the messaging infrastructure, delivering user messages to the chatbot server and relaying chatbot responses back to the user. The chatbot server hosts the backend infrastructure responsible for processing user messages, generating responses, and managing conversations. It incorporates various components, including NLP, dialog management, and integration with external systems or databases. Some of the chatbot components are explained below.

Business Logic and Integration: The chatbot's backend incorporates business logic and integration with external systems or databases. This allows the chatbot to fetch relevant information, perform actions, or access data from external sources to provide accurate and up-to-date responses.

API Integration: The chatbot server integrates with the WhatsApp Business API to send and receive messages with the WhatsApp platform. It utilizes the API's capabilities for sending responses, handling user interactions, and managing chatbot sessions within the WhatsApp environment.

Dialog Management Component: The dialog management component governs the flow and context of the conversation with the user. It keeps track of the conversation history, manages session state, and ensures coherent and meaningful interactions. Dialog management decides which responses to provide based on the user's input and maintains context throughout the conversation.

• Context Tracking: The dialog manager keeps track of the ongoing conversation, including user queries, previous responses, and any relevant contextual information. It maintains a conversational state to understand the current context and guide the conversation flow.
• Intent Recognition: The dialog manager identifies the user's intent based on their input. This involves using NLP techniques, such as intent classification, to understand the purpose and meaning behind the user's message.
• Response Generation: Once the user's intent is determined, the dialog manager generates an appropriate response. This can involve retrieving information from a knowledge base, triggering specific actions.
• Multi-turn Conversation: The dialog manager enables the chatbot to engage in multi-turn conversations. It handles user queries, prompts for clarification if needed, and maintains the conversational flow while addressing the user's needs.
• Contextual Recommendations: In some cases, the dialog manager can provide personalized recommendations or suggestions based on the user's preferences, previous interactions, or available data.

The dialog management component is designed to provide engaging and meaningful interactions with users, ensuring smooth conversations and effective communication between the chatbot and front-line health workers or program teams.

Machine Learning and Training: Machine learning techniques are used to train and improve the chatbot's capabilities over time. This involves collecting user interactions, training ML models on this data, and continually refining the chatbot's language understanding, response generation, and conversational abilities.

Target Beneficiaries: With the flexible architecture and continuous update options, the target beneficiaries can be anyone we design for, and with the current progress, we are streamlining our bot specific to health programs and health care service delivery.

The current pilot version of the bot was developed under the umbrella of the Anemia Mukt Bharat Programme, RATI - (Reduction of Anemia through Technological Innovation), for which the target population is basically split into two categories:

• Beneficiaries
• Health workers

1. Beneficiaries: With the increasing penetration of smartphones in rural areas of India and  each household having access to at least one smartphone, it provides us with a unique opportunity to engage directly with different population categories, including

• Pregnant women
• Lactating women
• Women in the reproductive age group

2. Health worker: Front line health workers are the backbone of the healthcare delivery system in the country. Healthcare workers targeted by RATI include

• Accredited Social Health Activist (ASHA)
• Auxiliary Nurse Midwife (ANM),
• Anganwadi Workers (AWW)
• Supply chain managers like pharmacists
• Key decision makers like block and district managers and medical officers

Value Proposition:

Timely identification of target beneficiaries and registration to begin timely delivery of services and information that focuses not only on anemia but moves women along the life course journey from reproductive age group to pregnancy care to infant care through scheduled push messages and reminders for medicine and health checkups, as well as identification of any risks involved,

Aside from data capture tools, health personnel can access on-demand information, assessment tools, and follow-up and referral care tasks. They can efficiently manage their catchment area and improve the program's service delivery by using data-informed automatic scheduled messaging.

Though there is an influx of data generated at the ground level, decision-makers and healthcare managers are often shorthanded on the availability of readable and usable data that can be used to make informed decisions. through a central dashboard that displays the data to track the program compliance of the beneficiaries and front-line healthcare providers and displays the analytics necessary to build a decision support system.

Revenue Streams:

For replication and enhancement of the current bot version and implementing it in different programs and geographies, we are currently exploring the following revenue streams,

• Service Contracts: We are seeking  funding opportunities from government departments or public agencies through bidding or directly entering into service contracts with them.
• Subscription Model: We are also aiming to provide premium features or content through a subscription model.
• Grants from Donor Agencies: funding support from the agencies that offer grants through bidding, and collaboration
• Data licensing and transaction fees: We aim to provide paid insights to governments or researchers.

Impact Measurement:

We track a variety of metrics to gauge our progress towards our impact goals, including:

• Intent to use and adoption rate (to indicate an information system's success)
• Using stock monitoring to close the annual and monthly stock estimation gap
• Beneficiaries' adherence to the recommended dosage of medications
• Reduction in anemia prevalence

We began developing the WhatsApp chatbot to test its feasibility as a low-cost mass-adopted intervention to address the issues identified by rapid assessment research, to understand and identify supply chain issues, and to strengthen the primary health system and its logistics management to improve IFA and Ca consumption and adherence to the Anaemia Mukt Bharat program, which is running in response to high anemia prevalence in multiple Indian states, with a 30,000 dollar grant from Nutrition International.

However, now, based on the success and adoption rate of the bot as a monitoring and information access tool among the programs' target beneficiaries and frontline health workers, we are replicating the model and promoting its use for more health programs with a much broader target user base and geography. For replication and enhancement of the current bot version and implementing it in different programs and geographies, we are currently exploring the following options,

• Service Contracts: We are seeking  funding opportunities from government departments or public agencies that prioritize healthcare workforce development or strengthen public health programs through bidding or directly entering into service contracts with them requiring specific functionalities or services from WhatsApp chatbots, including providing customer support, delivering personalized content, or facilitating transactions.
• Subscription Model: We are also aiming to provide premium features or content through a subscription model. Users who want access to exclusive services, advanced features, or personalized experiences can pay a recurring fee to subscribe. With this model, we are providing data analytics and integration with different systems and monitoring tools.
• Grants from Donor Agencies: We are seeking funding support from the agencies to offer grants specifically designed to support initiatives that improve healthcare delivery, enhance the capacity of health workers, or strengthen health systems. We identify agencies and foundations that align with the goals and objectives of our project, and submit grant proposals to secure funding. We also explore opportunities to collaborate with other organizations and apply for collaborative grants or consortium funding.
• Data licensing and transaction fees: By collecting and analyzing user data, we are able to provide decision support systems and tailor the user experience of a service delivered to them. We also aim to provide insights to governments or researchers. In the future, we are also looking to facilitate medicine delivery, referral, and follow up care, which we will be using for revenue generation through charging a percentage or fixed fee for each successful transaction.

We will be looking at various kinds of revenue streams so that, in addition to delivering to specific clients, we can improve and advance our chatbot and its quality service delivery to consumers.

As we already informed you in the previous response, we began developing the WhatsApp chatbot to address the issues identified by rapid assessment research, to understand and identify supply chain issues, and to strengthen the primary health system and its logistics management to improve IFA and Ca consumption and adherence to the Anaemia Mukt Bharat program, which is running in response to high anemia prevalence in multiple Indian states, with a 30,000 dollar grant from Nutrition International, which includes the system study  and rapid assessment and the BOT development cost.

As per our business model, we are now trying to replicate the same with broader geographies, and we are pitching to two other states of India, one of which is Meghalaya, for their health program monitoring and reporting.

We are also enhancing the RATI and the data coming through it to advance service delivery, and we are seeking funding support from the state and the third-party organization Nutrition International. We are continuously enhancing our product by incorporating AI and ML into it. We are pitching our product to some of the funding platforms, and we are actively applying for the challenges. and seeking funds to continue our product development.