We’re building Cardios, an intelligent system to solve heart failure monitoring. We want to help 26 million heart failure patients around the world to live healthier by encompassing the cardiovascular consultation and precision medicine into a self-monitoring platform. Ultimately, we want to reduce re-hospitalizations and improve patient clinical outcomes.
As medical students, Gabriel and I have spent time on hospital wards and we've seen the devastating impact that heart failure can have upon patients. We see the scale and the severity of the problem almost every day when we're and the wards, hence we are determined to find a solution to help patients.
We realised that re-hospitalizations are very common; 1 in 4 HF patients are re-admitted to hospital within 3 months. The economic impact of HF is huge, encompassing annually almost 2% of total NHS spending in the UK, $39 billion in the USA and $108 billion worldwide. Research suggests that 70% of these costs are from re-hospitalizations alone. Up to 50% of these re-hospitalizations are considered preventable. The scale and the severity of the problem has inspired Gabriel, Jaime and I to work on Cardios so that we might be able to identify deteriorating patients earlier and prevent them from being re-admitted to hospital.
What it does
We are creating a data analytics platform to analyse key information collected from automated phone calls, smartphones and wearables. We plan to use readily available smartphone data (e.g. walking distance/steps to assess exercise tolerance, sleeping time and quality to assess paroxysmal nocturnal dyspnoea) and actively acquired data (e.g. answers to research-verified Framingham Criteria questions and measured pulse rate/heart rate variability using smartphone cameras) to build an accurate model of a patient's heart failure status. The acquired data from a variety of sources would be stored on the cloud to allow for effective analytics.
Research Verified Metrics: Cardios will generate actionable insights by using research-verified questions, based on the internationally accepted Framingham criteria, to evaluate the patient’s cardiovascular state in the same way a cardiologist would do in a consultation. Patient’s would be able to respond by selecting a number from 1-10, thus providing quantitative data for a conventionally qualitative index. These key cardiovascular questions would be independently pushed to patients either through an automated phone call or a smartphone application and responses would be stored in the cloud to provide a long-term overview of the patient’s condition. Automated phone calls running on Google Voice API would ask the Cardios questions and patients would respond by saying the number aloud. This would widen our reach to patients who may not have smartphone access. Through using intelligent data analytics Cardios would identify deteriorating patients earlier and inform patients so that physicians can review their management plan before the need for re-hospitalization.
Patient Education: Another key aspect of improving heart failure management is to offer patient education regarding the disease and to optimise lifestyle changes. Using data from existing smart phone sensors (e.g. steps walked, sleep cycle etc…) we will also offer personalised lifestyle advice and responsive suggestions. We aim to maximise adherence through a rewards-based system.
Growth: The Cardios system’s machine learning system will learn to get better at identifying trends and patterns with the more patients it experiences. Cardios will be personalised to each user by tailoring patient risk/deterioration parameters based on interaction with the Cardios system and any additional information the user chooses to provide, such as genetic data if the patient has used private genome sequencing services, such as 23 and me. In the future, we hope to offer genome sequencing as an optional service through Cardios, so each user experience is as personalised and as effective as possible. Through precision medicine, we hope to revolutionise the field of cardiology and achieve huge improvements to patient outcomes. After validating our system for heart failure, we intend to be leaders in applying digital monitoring and precision medicine to other major chronic health conditions.
How we built it
We are in the early stages of building the prototype and then will move towards a phase 1 clinical trial
Challenges we ran into
Obtaining large heart failure datasets to validate and test our model.
Accomplishments that we're proud of
1.) Winners - The Telegraph STEM Innovation Award 2018 (out of 8000+ applicants) 2.) Top 10 Solution - London Mayor's Entrepreneur 2018 (out of 450+ applicants)
What we learned
We've learned the importance of creating an easy-to-use interface for Cardios, especially given that most heart failure patients are over 55 years old. Moreover, since not all patients may have access to smartphones, we have decided to use automated phone-calls and natural-language-processing to widen the accessibility of Cardios. At the very minimum, patients would just need a phone number to access the system.
What's next for Cardios
Our primary aim is to finalise our prototype and run a phase 1 clinical trial to quantify the parameters of the Framingham criteria. The clinical trial would involve enlisting 50-100 heart failure patients for 3-6 months to test our prototype and determine intervention thresholds. We have discussed our concept with a London-based cardiologist to validate our approach. During the trial, we will collect data to evaluate patient adherence, patient education and overall patient outcomes.