Dubai-based healthcare technology startup TachyHealth has developed an epidemiological compartmental model to predict the likely peaks of the Covid-19 curve in the UAE, with data collected from the literature reviews, comparative analysis, World Health Organisation (WHO) and the World Bank.
TachyHealth makes value-based healthcare a reality by working at the intersection between the players and providers, the medical insurance and hospitals.
Dr. Osama AbouElKhir, CEO of TachyHealth, told TechRadar Pro Middle East that compartmental models are used for the mathematical modelling of communicable diseases in which the population is divided into groups (compartments) with the assumption that individuals in the same compartment have the same attributes.
“The SIR (susceptible-infected-recovered) epidemic model is one of the compartmental models in which real-time data query was performed and visualised, then the queried data is used to build the SIR predictive model based on daily observations of confirmed, death, and recovered cases from Covid-19,” he said.
Being a medical doctor by profession, a data scientist by passion, and working in the public health domains in the UAE and Saudi Arabia for health authorities, AbouElKhir said that the UAE was one of the first countries in the GCC to have a Covid-19 confirmed case.
“Everything relates to case zero or the first case. We started the data modelling from the start date and taking into consideration the reproduction number or how many persons are going to be infected from one infected person and the recovery period, days taken by a patient to recover,” he said.
TachyHealth assumed the epidemiological and statistical data such as the recovery period to be from 12 to 14 days. Also, the basic reproduction number is assumed to be between 2 and 2.6 persons as referenced by the WHO in its March 6th report, as well as the population of each country as referenced by the World Bank. The basic reproduction number is one of the commonly used metrics to check whether the disease will become an outbreak.
Three case scenarios in UAE
- Scenario 1: Peak date is expected on April 29th with the assumptions that 5% of the population is susceptible, reproduction number= 2.6 and the recovery rate = 14 days.
- Scenario 2: Peak date is expected on May 16 with the assumptions that 2.5% of the population is susceptible, reproduction number= 2.3 and the recovery rate= 13 days.
- Scenario 3: Peak date is expected on May 28th with the assumptions that 1% of the population is susceptible, reproduction number= 2.6 and the recovery rate= 12 days.
“Our confidence of the UAE model exceeds 80% due to the high transparency in the reported data in addition to the fact that UAE has one of the highest Covid-19 tests to the million population in the world. We are working to publish it in international scientific journals. We have almost reached the peak in UAE and that is why the government is starting to ease the containment policies and restrictions while maintaining some mitigation efforts” AbouElKhir said.
“After the pandemic, digital-first mindset is going to be the approach for patients as well as hospitals and healthcare players in general. WHO also needs to invest in big data and artificial intelligence as they can strengthen the communicable diseases surveillance, foreseeing outbreaks and pandemics even before they happen” he added.
AbouElKhir said that its artificial intelligence and data science solutions bring intelligence and automation to the end-to-end value chain from the point of data entry by the physicians to the medical coding and billing by medical coders to the claim management at both sides of the spectrum through a suite of solutions.
The solutions help health insurance companies to better manage the big data related to medical claims, maximise the efficiency of the process, and improve the medical auditing practices. They also help hospitals in minimising the denials and rejections, improving the revenue cycle management, supporting doctors at the point of entry with evidence guidelines, and automate the medical coding by converting any part of the medical record (physician's notes, lab tests, radiology results, discharge summary, etc.) to the right coding output.
The company has developed the applications through a team that encompasses machine learning PhDs and researchers, software engineers in addition to the clinical and healthcare consulting executives and working with the insurance companies and hospitals to automate their workflow in a bid to cut cost and time.
AbouElKhir said that there is increased pressure on healthcare stakeholders and executives to minimise the operational costs without impacting the staff and patients.
“This is where AI comes into play and helps them to do better with lower cost. Due to the digital transformation of the healthcare sector, we are seeing increased queries from insurance companies and hospitals in UAE, Saudi Arabia and Egypt. Our aim is to make the healthcare more efficient with a better value,” he said.
Before the Covid-19, he said that investors were hungry to invest but due to the low oil prices, they [investors] are very cautious in investing.
“We are planning to close a seed funding of $500,000 within a month. We have already invested close to $200,000 since inception. We're also working on a deep learning model for Covid-19 detection in medical images and with other companies and startups to bring remote patient monitoring and telehealth soon to support the healthcare providers in handling the surge in a better way,” he said.