MedBrain

According to The Lancet’s Comission on Diagnostics, 47% of humanity has no access to a medical diagnosis. Access to diagnostic resources is fundamental to quality health care. Primary care settings are considered the diagnostic last mile, and lack of resources in these settings particularly affect poor, rural and marginalised communities, globally.

Two major reasons affecting the global diagnostic gap are a severe shortage of access to medical specialists, and a severe shortage of access to gold standard diagnostic tests.

Maternal & new-born survival is linked to quality of care they receive in Health facilities (Lancet Glob Health, 2019). About 303,000 women globally die from complications, from which 99% (302,000) of women’s death were in developing regions taking 66% in sub-Saharan Africa, due to their poor quality of health care service (HRSA, 2019. p.3; WHO, 2019). In Ethiopia, EMDHS 2019 showed that though, maternal mortality decreased from 676 deaths per 100,000 live births in 2011 to 401 in 2019 and under-5 mortality and infant mortality decreased from 123 and 77 per 1000 live births in 2005 to 59 and 47 per 1000 live births, respectively in 2019 showing no significant reductions in neonatal mortality (33 deaths per 1,000 live births in 2019). 99% of maternal death are avoidable.

The Primary Health Care on the Road and Universal Health Coverage 2019 Global Monitoring Report mentions that the lack of timely diagnosis and price is one of the main obstacles to providing quality health care and increasing access to effective health services, especially in countries low and middle income. But not only do all of us face at least one misdiagnosis throughout our lives, but in low-income countries and low-income households the percentage of misdiagnoses increases.

In fact, numerous studies suggest that there is a clear gap in clinical diagnosis around the world and that this gap varies depending on socioeconomic status and other factors, such as age, gender, and race or ethnicity. Here socioeconomic barriers are especially important since, in the case of diagnoses, barriers could be associated with cost or lack of insurance coverage, or both. Therefore, problems with clinical diagnoses are aggravated in low- and middle-income countries where the population has limited access to basic diagnostic tests due to their cost, in addition to inequalities in access to medical care between different groups. of population throughout the world, especially in the comparison between urban areas and rural areas.

The problem that MedBrain is tackling is the global diagnostic gap, further explained in the MedBrain Dossier.

We present an artificial intelligence-based software to allow  community health workers in rural areas who do not have access to an specialist doctor, to offer specialized medical diagnosis and treatment of equivalent quality to that of an specialist doctor.

MedBrain offers a novel approach to the generation of a diagnostic prediction based on clinical data. MedBrain’s approach is based on the following innovations: A measurement of the diagnostic weight of each symptom, physical sign, and risk factor, for each individual disease. 

The diagnostic weight is also called the likelihood ratio, which is obtained from sensitivity and specificity data. Furthermore, instead of focusing only on individual tags, MedBrain focuses on the combination of tags (clinical patterns). Based on the data inserted about the patient, MedBrain focuses on identifying the probability of specific clinical patterns being present. The most probable clinical patterns are suggested to the healthcare professional, who can determine the presence or absence of the clinical patterns. In contrast with other CDSSs, MedBrain does not use any static decision trees. MedBrain performs a weighted match between the patient’s information and the tags within each disease in the database. Based on the strength of all weighted matches, a Disease Rank is generated. Based on the Disease Rank and on MedBrain’s disease database, MedBrain generates a Tag Rank - which ranks the symptoms, physical signs, and risk factors (present in the MedBrain database) with greater diagnostic potential for that specific patient, at each step of the dynamic interview. At each step of the interview, MedBrain gathers the top 3 tags within the Tag Rank and asks the tags as the new questions. The Tag Rank helps identify the ‘next best step’ to take, clinically, in either the clinical interview or physical examination.

Our target popullation are community health workers in rural areas who do not have access to an specialist doctor.

MedBrain allow them to offer specialized medical diagnosis and treatment of equivalent quality to that of an specialist doctor.

The entire population of rural areas that does not have access to a specialist doctor, thanks to MedBrain, will have access to diagnosis and treatment of specialist doctor quality.
This will improve the health of the entire population.
We improve and contribute to universalizing health coverage.

Sub-Saharan Africa continues to have the lowest health worker density, with only 2.3 medical doctors and 12.6 nursing and midwifery personnel per 10,000 population, resulting in nurses and midwives accounting for more than half of the professional health workforce and contributing to 90% of patient contact.  In Ethiopia in 2022/23, 310,591 health professionals provided health services in public health institutions. Nurses, health extension workers, and health officers were the top three professional categories, accounting for 33.2%, 13.8%, and 10.7% of the total, respectively. General practitioners and specialists made up 5.7%. The health professional density was 1.4 doctors, health officers, nurses, and midwives per 1,000 population, far below the national target of 2.3 per 1,000 and the WHO target of 4.45 per 1,000 population.

Due to Ethiopia's underdeveloped healthcare infrastructure, limited health care workforce, rapidly growing population, and increased fertility, it is imperative to design a system to enhance the quality of service provided by mid-level health care professionals, especially in areas where physicians are unavailable. This is especially important for common emergency conditions, where early diagnosis and management can improve patient outcomes and mitigate the burden on the healthcare system. 

Our Team: https://medbrain.io/

Pol Ricart, MD

Founder & Chief Executive Officer

Pol is a medical doctor and former resident in neurosurgery at the NHS. After acknowledging the global diagnostic gap and the potential of data science in diagnostics, he left his career in neurosurgery to build MedBrain full-time.

As CEO, he leads innovation by identifying market needs and leveraging data science, medical statistics, and gathering clinical feedback from our team of medical advisors and from different stakeholders in the healthcare industry. He runs daily operations, helping lead our medical and engineering departments to ensure OKRs are met.

Pol has been previously selected as a Global Top 50 entrepreneur by The Kairos Society (K50, 2017) and by Singularity University (Global Grand Challenge Awards, 2017), and has been featured as a Top 50 Global Entrepreneurs to Watch in the Next Decade, by Inc. magazine.

Pol studied multilateralism, diplomacy, and humanitarian action at the United Nations (2019).

Iñaki Alegría, MD MSc

Chief Medical Officer

Iñaki is a medical doctor and specialist in paediatrics. He graduated from the University of Barcelona and specialized in pediatrics at the General Hospital of Granollers (Barcelona). He also obtained a master’s degree in international health and cooperation.

In Ethiopia, where he has more than 10 years of professional experience, he has worked as medical director at Gambo General Hospital (www.gambohospital.org ). He has worked on the front line in the response to the COVID19 pandemic, measles epidemics, and participated in the improvement of the national maternal and child health program supporting Ethiopian health institutions. 

His work and career have been recognized with national awards such as the Spanish Association of Pediatrics or the International Health Society and at the international level where the Ubuntu award as a social leader for the defense of health, awarded by the Euro-African forum, stands out. 

Alex Cáceres, BSc MSc

Chief Technology Officer

Alex is an experienced, full stack software engineer. He leads our software development team at MedBrain. Alex constitutes the bridge between the engineering and medical teams, and helps provide technological solutions to MedBrain’s innovation requirements.

He has a postgraduate Degree in Full Stack Web Technologies (UPC Technology Center), and a Bachelor’s Degree in Industrial Electronics and Automatic Control Engineering.

Alex’s technical skills include: Languages/Frameworks: Javascript / Typescript – Node, Express, React. Python – Django. Java – Spring. PHP – Laravel, Infrastructure: MySQL – MongoDB – PostgreSQL. Kubernetes – Docker. CI/CD – Jenkins – Bash Scripting. AWS – G Cloud – Azure DevOps.

Tigist Workneh, MD MPH

Tigist Workneh is a medical doctor and public health specialist currently working as Research Director at MedBrain and at the Medical Research Lounge. She also serves as a research mentor at St. Paul’s Hospital Millennium Medical College and Addis Ababa University College. During the mentorship for clinical and master’s research program, she has successfully advised over 300 students. So far, she has led more than 30 original articles. She is also a reviewer to national and international health journals. 

Agreement signed in Ethiopia and Nigeria for clinical validation

We have been pilot testing MedBrain regularly over the past few months in hospitals and rural health centers in Ethiopia and Nigeria. Before commercial deployment, we are currently focusing on clinical validation, both in terms of its technical feasibility (measurement of diagnostic accuracy) and usability (customer validation, user feedback, A/B tests). To do so, we are currently running two multi-centric clinical trials.

Despite still being in validation phase, MedBrain has received a remarkable amount of interest by hospitals and healthcare organisations worldwide. The following are our agreements to date.

🟢  Research Agreements Closed 🧬

Agreements to run clinical validation, reaching a scientific publication and having also expressed interest in continuing to use MedBrain with a commercial agreement after the validation process:

• Hospital Sant Joan de Déu 🇪🇸
• Diagnext Hospital Network 🇧🇷
• Hospital Beneficente Português 🇧🇷
• Saint Paul’s Hospital Millennium Medical College 🇪🇹
• ALERT Specialised Hospital Addis Abeba 🇪🇹
• Dodola General Hospital 🇪🇹
• Gambo General Hospital 🇪🇹
• Enugu State University Teaching Hospital 🇳🇬
• Annunciation Specialist Hospital 🇳🇬
• Enugu State Hospital Network 🇳🇬
• Parklane Enugu Hospital 🇳🇬

🔵 Letters of Interest 📄

We have documented interest from the following organisations. We would like to establish research and commercial agreements with them in the near future. However, we are currently prioritising our ongoing research activities.

• DKV Digital Health 🇪🇸
• NHS UK West Midlands 🇬🇧
• Centro Hospitalario Sao Tomé 🇸🇹
• Sant Joan de Deu Hospital in Sierra Leone 🇸🇱

We are convinced that Solve can support us through the following areas: 

Financial Support

Strategic Partnerships in Key Regions

•⁠  ⁠Go To Market strategy

•⁠  ⁠Networking/Relationship building with Strategic Partners in Emerging Nations

•⁠  ⁠Reaching Economic Buyers faster

Techical Support:

Expertise in supervised Machine Learning

Legal Support:

Navigating regulatory pathways (ISO, EU Medical Device Class II, HIPAA Compliance, etc.

MedBrain is a tablet/mobile tool to support and help with medical diagnosis. In MedBrain, healthcare professionals can enter the patient's symptoms, signs, risk factors, medical history, and demographics. From this data, MedBrain will generate a dynamic interview based on asking specific questions (or questions for the patient, or physical signs to examine on the patient's body). The answers are always Yes/No, which increases ease of use. This dynamic interview is an iterative process of between 5-20 steps. The answers in the previous steps will determine the questions in the following steps. During this iterative dynamic interview process, when MedBrain reaches a predetermined probabilistic threshold, it will generate a diagnostic prediction.

MedBrain Works Differently

Key differences in our functioning result in a difference in performance.

MedBrain has demonstrated a market-leading clinical diagnostic accuracy:

MedBrain’s diagnostic accuracy (measured with 250 externally-sourced patient cases) significantly surpasses all other documented performances. For a description of our validation process, please visit Scientific Validation.

Differences between MedBrain’s approach and other widely-used CDSSs:

MedBrain offers a novel approach to the generation of a diagnostic prediction based on clinical data, with major differences with respect to the existing CDSSs discussed. MedBrain’s approach is based on the following innovations: A measurement of the diagnostic weight of each symptom, physical sign, and risk factor, for each individual disease. The diagnostic weight is also called the likelihood ratio, which is obtained from sensitivity and specificity data. Furthermore, instead of focusing only on individual tags, MedBrain focuses on the combination of tags (clinical patterns). Based on the data inserted about the patient, MedBrain focuses on identifying the probability of specific clinical patterns being present. 

The most probable clinical patterns are suggested to the healthcare professional, who can determine the presence or absence of such clinical patterns. In contrast with other CDSSs, MedBrain does not use any static decision trees. MedBrain performs a weighted match between the patient’s information and the tags within each disease in the database. Based on the strength of all weighted matches, a Disease Rank is generated. Based on the Disease Rank generated and on MedBrain’s disease database, MedBrain then generates a Tag Rank - which ranks the symptoms, physical signs, and risk factors (present in the MedBrain database) with greater diagnostic potential for that specific patient, at each step of the dynamic clinical interview. At each step of the interview, MedBrain gathers the top 1 tag (with major diagnostic potential) within the Tag Rank and asks the tag as the new question. In other words, the Tag Rank identifies the ‘next best step’ to take, clinically, either in the form of a question to the patient about their symptoms, or an examination of a physical sign on the patient’s body.

As a consequence of the aforementioned technological innovations, MedBrain has managed to go beyond a 29% diagnostic accuracy (market average) up to an 83% diagnostic accuracy. 

Accuracy, sensitivity, and specificity levels close to 80-100% are required in order to consider a diagnostic support tool reliable.

Direct impact on SDG

SGD 3: Good health and Well being

 Ensuring healthy lives and well-being at all ages is essential for sustainable development. MedBrain, through high-precision diagnostic prediction, will allow better decision-making capacity by the healthcare team, guaranteeing good, earlier and more effective treatment, allowing optimal recovery.

 Qualitative relationship:

The advances made in the field of health allow the health of millions of people to improve every day. However, the sector continues to face a number of adversities, which make quality healthcare a constant challenge. Therefore, misdiagnosis and late diagnosis can have serious and widespread consequences. MedBrain will allow a clinical diagnosis to be made with an accuracy of 93%, the highest on the market, which will help achieve greater medical precision and a great improvement in decision-making by healthcare professionals.

 Quantitative relationship:

According to the World Health Organization (WHO), telemedicine can improve access to healthcare in remote and rural areas by up to 60%. Likewise, a report from the Office of Inspector General of the United States Department of Health and Human Services indicated that the implementation of technology in the health sector can reduce medical errors by up to 88%. According to 2017 research from the Mayo Clinic, 88% of patients who ask for a second opinion receive a different diagnosis. While it is estimated that the number of patients who lose their lives each year due to wrong diagnoses is around 40,000 people.

SDG 9: Industry innovation and infraestructure

Innovation and technological progress are key to addressing current challenges worldwide as they can provide more efficient and effective solutions to existing problems. Specifically in the field of health, it has a great impact on people's well-being and the quality of medical care. MedBrain is a computer system that aims to support and improve the clinical diagnosis process in medicine, which presents high innovation thanks to an accuracy of 93% compared to 43.5% of its direct competitors.

Qualitative relationship

According to a study published by the Organization for Economic Cooperation and Development (OECD), improving the efficiency of the health system can increase GDP per capita by 1.5% each year.

Quantitative relationship:

According to a report by Frost & Sullivan, with artificial intelligence, improvements in health outcomes of around 30~40% can be achieved, reductions of up to 50% in the cost of patient care, and generating a strong boost to research of new treatments.

SDG 10 Reduced inequalities

Through the use of MedBrain, medical care is brought closer to all those people with difficulties in accessing primary care, either due to their place of residence or their economic possibilities.

People who live in deprived areas or belong to disadvantaged groups have higher mortality rates due to lack of access to adequate healthcare and health services and may be more likely to develop diseases that could have been prevented through early detection. and treatment. MedBrain will allow access to medical care to all those in remote areas, avoiding the need to travel long distances to visit a doctor.

Our objective are to bring Neonatal Mortality and Child Mortality rates under the United Nations 2030 Target. (SDG 3)

If we focus in Ethiopia: (per 1.000 live births)

UN Neonatal Mortality Taget: <12

UN Under 5 Mortality Target: <25

At this time, in Oromia region, where we are going to start: (per 1.000 live births)

UN Neonatal Mortality Taget: 37

UN Under 5 Mortality Target: 79

Our technological innovation:

MedBrain has created its own medical database using public scientific databases, and clinical evidence has been collected from them for each of the diseases in the database. Additionally, for each disease, a list of Tags (labels) or list of diseases has been created, which can be Symptoms, Signs, or Risk Factors. Each disease contains evidence from multiple different databases, which MedBrain has centralized. Finally, a weight (diagnostic importance) has been assigned to each Tag, this is what is known as Likelihood Ratio (LR). This is the most important feature of the database since for each Tag the sensitivity and specialty of that individual Tag for an individual disease has been compiled from the scientific literature.  Compiling the sensitivity and specificity data scattered throughout the scientific literature, the LR has been calculated internally for each Tag for each of the more than 100 diseases. As a result, MedBrain has an unprecedented database, with evidence on the diagnostic importance of each calculated Symptom, Sign, and Risk Factor, and with all of this centralized within MedBrain.

Secondly, it is worth highlighting the use of the diagnostic algorithms used by MedBrain for the analysis and processing of clinical data with the aim of generating diagnostic recommendations or suggestions. The first step is taken by the user, who inserts the data from the Clinical Interview and the Patient's Physical Examination. Two algorithms then come into action. The first Algorithm 1 (Design Rank) is a mathematical formula that establishes a series of matches between the patient's information and the database, and generates a score for each disease with which there has been a match. Algorithm 2 (Tag Rank), based on the Disease Rank (list of diseases with their probability index of ending up being confirmed as the correct diagnosis), another mathematical formula is used to generate a ranking of the Tags that will provide the most diagnostic prediction. contribute (the 'Next Best Step'). These Tags can be new questions to ask the patient, or new signs to examine in the patient's body. Finally, when the new information suggested in the 'Next Best Step' is re-inserted, the data will be added within the patient's case, and when the patient's case data changes, a new Disease Rank will be generated and so on. until a probabilistic threshold is exceeded and a specific Diagnosis is predicted.

technological innovations registered as intellectual property:

• Measurement of the weight (diagnostic importance) of each label (sign, symptom or risk factor) based on the measurement of the sensitivity and specificity of each label, measured from the prospective aggregation of clinical cases inserted in the platform.

• Use of Boolean operators of combination, hierarchy and sequence of clinical labels to increase the diagnostic accuracy of the tool.

• Calculation of the diagnostic importance of each label at each moment, as a function of the weight of the label in each disease, and a function of the diagnostic probability of that disease at each moment of the clinical case .

MedBrain’s internal (pre-clinical) validation process:

 According to MedBrain’s welcome page, an internal validation with 250 externally-sourced patient cases has been performed. MedBrain staff inserted the patient cases (sourced from the BMJ), one by one, into MedBrain, and reported the results. They offered a comparison between the BMJ’s predetermined diagnosis, and the diagnosis generated by MedBrain.

 The results obtained with the 250 externally-sourced patient cases was a diagnostic accuracy of 83% (208/250) for the Top 1 Diagnosis, 93% (234/250) for the Top 2 Diagnosis, and 98% (245/250) for the Top 3 Diagnosis.

 The dataset of this internal validation has been provided publicly, and can be accessed here. These internal validation results, despite having low quality of evidence (they are internal, retrospective, and non-clinical), pose an interesting insight into the potential diagnostic accuracy of MedBrain. If similar results (in terms of diagnostic accuracy) are reproduced in the clinical research, MedBrain’s impact on global health could potentially be very significant.

The MedBrain team is highly experienced in the key convergent disciplines of clinical medicine, data science, artificial intelligence and clinical research

The core team:

Pol Ricart, MD - Founder & Chief Executive Officer

Iñaki Alegría, MD MSc - Chief Medical Officer

Alex Cáceres, BSc MSc - Chief Technology Officer

Tigist Workneh, MD MPH - Clinical Research Director

24 Months

We are deeply committed to fostering a diverse, equitable, and inclusive environment within our leadership team and across all aspects of our organization. We recognize that diversity encompasses a wide range of social, cultural, and identity-based attributes, and we value the unique perspectives and experiences that each member brings to our team.

Our team's goals for becoming more diverse, equitable, and inclusive are multi-faceted. Firstly, we aim to ensure that our leadership team reflects the rich tapestry of human diversity, including but not limited to race, ethnicity, gender identity, sexual orientation, ability, nationality, and socioeconomic background. We believe that diverse leadership leads to more innovative solutions and better decision-making processes.

Equity is a fundamental principle guiding our actions. We strive to provide equal access to opportunities for all members of our team, recognizing that true equity requires addressing disparities and dismantling systemic barriers that have historically hindered the full participation of marginalized groups. This includes implementing fair hiring practices, offering professional development opportunities, and creating a supportive work environment where every individual can thrive.

Inclusion is at the heart of everything we do. We are dedicated to fostering an inclusive culture where all members of our team feel welcomed, respected, supported, and valued for their unique contributions. This involves actively listening to diverse perspectives, promoting open dialogue, and creating spaces where everyone feels empowered to express themselves authentically.

By prioritizing diversity, equity, and inclusion, we are not only strengthening our organization but also advancing our mission to drive positive social impact through innovation. We remain committed to continuously learning and evolving in our journey towards a more diverse, equitable, and inclusive future.

Our business model is two-fold:

1. B2G agreements with regional governments. Top-down implementation of the MedBrain solution in public hospitals. Economic buyer: Regional government (50% of fee) and public hospital (50% of fee). Shared expenses between the hospital's annual budget, and the health bureau of the regional government. Annual flat fee for product use, per hospital.

2. Monetisation of anonymous clinical data. In the process of providing digital triage, diagnosis and management plan services, MedBrain tracks every variable about the patient's Clinical Presentation (how does the patient present?), Clinical Interventions (what is done to the patient?), and Clinical Outcomes (what ends up occurring to the patient?). Economic buyers: Pharmaceutical companies, to gain healthcare system insights based on RWE (Real World Evidence) and CROs (Clinical Research Organisations), using MedBrain as a source of patients to include in their clinical trials.

The following is a description of the type of variables that are tracked in each clinical case run on MedBrain, which provides insight into the kinds of insights that we are able to provide to Pharmaceutical companies and CROs.

We are currently working with IQVIA Spain on a pilot project to assess our capacity to generate reliable clinical evidence from anonymous clinical data.

____________

In terms of funding, we have successfully raised a €300k pre-seed round (by VC firms Plug and Play Silicon Valley, and Kunsen Health Madrid), and a €100k government grant (Startup Capital, by ACCIÓ).