Artificial Intelligence | Lesson 6.2

Ada Health

The Problem

Globally, more than 4 billion people lack access to basic healthcare services. Financial and geographical barriers are the primary reasons for this mishap. Also, there is a global shortage of healthcare workers, particularly in rural areas the condition is worse.

In today’s society, individuals are increasingly turning to Google to find answers to their health questions before consulting a doctor. Hence, there is a need for a patient-facing health guide. Even though life expectancy is rising, chronic diseases are increasing.

The Challenge

The world is facing a severe global health challenge – over 4 billion people lack access to basic health care services, leaving millions without the care they need and deserve. This alarming situation is mainly attributed to major financial and geographical barriers. In developing countries, three out of four people live in rural areas (The World Bank, 2017). Still, healthcare resources are mostly concentrated in urban areas, resulting in limited access to appropriate care. The demand for affordable, effective, and quality health services is high. It cannot be met by the traditional, costly model of delivering healthcare solely through doctors and nurses, which are most often not accessible in rural settings. The massive shortage of more than 7 million health workers urges us to think about new solutions that can be deployed quicker. Rising smartphone penetration in low-income and middle-income countries (LMIC) provides the unique opportunity for mobile AI health applications to facilitate new access to healthcare for vulnerable communities and fundamentally transform healthcare delivery.

The Solution

Ada is designed to relieve the pressure healthcare services are facing. We can bridge the gap between consumer expectations and the actual capability of legacy healthcare systems.

Ada’s AI-powered platform provides personalized symptom assessments in real-time. Once a user shares their symptoms, Ada asks a series of questions, searching through infinite symptom combinations and thousands of possible conditions to find the most probable cause.

To increase global access to high-quality care, Ada Health applies its AI technology in two ways. Through its consumer-driven approach, Ada empowers individuals with instant access to credible, personalized medical expertise through its AI-powered app. Through its frontline health worker approach, Ada equips semi-skilled health workers (such as community health workers, pharmacists, nurses and midwives) with clinical decision support tools tailored to support primary care delivery. The Ada Foundation was founded by Ada Health in 2018 and is committed to putting Ada’s ground-breaking technology into the hands of thousands of community health workers all over the world to enable better-informed decisions, support diagnosis, and improve appropriate care navigation of patients. The Ada foundation’s mission is to overcome these challenges and bring life-saving, high-quality care and decision support to millions. We can achieve this by empowering community health workers and local drug shop dispensaries with Artificial Intelligence (AI) via mobile technology.

Launched in 2016, Ada helps individuals to understand and manage their health through a personalized health assessment, while providing earlier health information and decision support to doctors. With an approachable, user-friendly interface, anyone can enter their describe their situation. With a doctor’s warm, friendly tone, Ada guides users through a series of questions related to their symptoms and provides them with a list of most likely conditions. Ada has more than 3.5 million users globally, who have completed over 5 million assessments. Ada offers a secure platform for collecting patient data and providing guidance on possible conditions and next steps.

Their medical reasoning engine was designed over the past seven years by a team of over 100 doctors, data scientists and engineers to solve the puzzle of diagnosis once and for all. With over 12,000 diseases and an infinite number of potential symptom constellations, it is simply impossible for a human to be able to identify all of these possible combinations. Through the use of AI, Ada solves this issue within minutes and supports decision-making in currently over five languages.

Figure 6.1 | Ada Health user interface.
Attribution: User Interface for Mobile App. Ada Website. All rights reserved.

Ada’s Value Proposition

Globally, healthcare is complicated, inaccessible and focuses on reactive versus proactive care, and Ada is shifting the healthcare model by providing instant primary healthcare regardless of where you are.

By complementing the strengths of human doctors with machine intelligence, Ada has built a comprehensive medical knowledge base covering hundreds and thousands of conditions, symptoms, and cases. Ada is able to consider the patient’s complete medical history, along with data from a broader range of sensors, devices and other sources, when making a health assessment.

In terms of usability, Ada is known for its approachable and conversational interface. It has been proven to be engaging to the users. Ada’s assessment is designed to work much like a good doctor would, determining the most relevant, helpful questions to ask, with intelligent algorithms enabling Ada to ‘get to know’ each individual better over time, providing a more comprehensive health picture and increasingly tailored guidance.

Business Model

Hundreds of millions of people currently lack access to healthcare and, with life expectancy rising and chronic disease rates increasing, health services are coming under huge pressure. Ada’s AI-powered health platform can help patients and doctors by enabling anyone with a smartphone to access high-quality, personalized health information while giving clinicians more time to focus on patient care. There are many opportunities in the healthcare industry for Ada to grow, and Ada is partnering with health providers, organizations and other businesses, where Ada can deliver valuable cost savings and improve access, care and treatment.

Market Opportunity

• According to the World Health Organization (WHO), the shortage of health workers globally is 7.2 million. This number is expected to grow to 12.7 million by 2035.
• As a result, the WHO has endorsed task shifting as a potential solution, whereby certain tasks are delegated to unspecialized health workers when appropriate to make better use of the available human resources.

Organization Goals

Ada Health will:

• Develop prototypes of the Ada tool adapted for use among different types of semi-skilled health workers.
• Introduce an effective referral system across stakeholders within primary healthcare.
• Form partnerships with relevant stakeholders, such as local governments and the private sector, to implement Ada tools on the ground in Sub-Saharan Africa and South Asia.

Differentiating Factor

Ada is an AI-powered health management platform that supports both patients and clinicians. Their system is independent of the location and it is instantaneous. To use Ada’s service, you only need access to the internet. Ada has been focused on providing very detailed, high-quality medical insights and decision support. To develop such an accurate system, they have spent more than seven years on R&D and curating a world-class medical knowledge bank.

Analysis

Scouting AI Opportunity

As we discussed in module 4, when it comes to AI opportunities, the organization has to analyze what they are capable of and what their competitors can and cannot do. They had a team of machine learning researchers and physicians in this case. They also were able to negotiate with the hospitals they were working for to collect data. Their goal was to provide healthcare to places with a deficiency in healthcare providers. There is no possibility to travel to those places and physically provide the service they need. Even if they could, the service would be limited to a specific location at any time. Their competitors were not using AI because they did not have the team and data, so they have to travel to provide healthcare.

For their product to work, they needed to analyze the risk. They made a business assumption: (a) People would trust the result of an AI-powered app to learn about their health and (b) take the diagnosis seriously. Their market research showed that individuals are increasingly turning to Google for their health questions before consulting a doctor. There is clearly a need for an accurate AI-based health guide. Life expectancy is rising, yet the chronic disease is increasing. Also, they have had many third parties analyzing their app so that individuals could trust the output of their model. Hence, their assumptions seem reasonable.

As for technical assumptions, they assumed that they could collect all the needed data and develop a model accurate enough to diagnose and give people health advice. Many research papers confirm AI’s capabilities in healthcare (Dabowsa, Amaitik, Maatuk, & Aljawarneh, 2017; Kumar, Koul, Singla, & Ijaz, 2022).

The next step would be breaking down the process into smaller bits. Figure 6.2 demonstrates the breakdown.

Figure 6.2 | Ada Health process breakdown.

Their process starts with the user typing their complaint in the app. The text will be analyzed, and the type of action will be determined. Based on the action, the corresponding AI classifier will be used to determine the possible actions and the corresponding responses to the user’s query. The app will show results if the AI certainty exceeds a certain threshold. Otherwise, it will ask the user for more information.

We have all the information needed to create the Framing Canvas for this business. You can find the business canvas for this business in Figure 6.3.

Figure 6.3 | Framing Canvas for Ada Health