As a visionary and self-described science geek with a hunger to change the healthcare industry at breakneck speed, it’s perhaps unsurprising that Osama AbouElkhir is at the helm of a Dubai startup named TachyHealth - any student of ancient Greek or medicine will recognise tachy as meaning "rapid".
AbouElKhir knows the internal workings of the industry intimately well – its strengths and its flaws. It comes from a background spanning multiple disciplines: clinical, administrative, management consulting and entrepreneurial.
Starting as a medical doctor in his early career he crossed over into hospital operations before making the leap into management consulting for the healthcare industry across the Middle East with Ernst & Young and IQVIA. Some of the more illustrious clients he worked for included Cleveland Clinic Abu Dhabi, Dubai Health Authority, the Ministry of Health in Malaysia and the Saudi Health Council.
Becoming an entrepreneur was a natural next step. Speaking to Omnia Health Insights, AbouElKhir described the business he co-founded as a deep technology R&D company focused on using machine learning and deep learning to solve challenging problems that mostly cannot be solved with traditional tech.
He described success in simple terms: “We try to save lives with technology. We help the healthcare focuses on what really matters - the patients - and in doing so, taking away all the waste, redundancies and inefficiencies.”
Founded in Delaware in October 2018, TachyHealth is a newcomer to Dubai, having been established in the UAE for slightly over six months.
Despite the short tenure, the company has achieved rapid recognition, owing perhaps to a top team composed of machine learning PhDs, medical executives, software engineers, insurance leaders and Fortune 500 professionals.
It was selected for the first Artificial Intelligence Lab in the region, AIVenturelab, powered by Microsoft and Area 2071. It was also picked for the first of its kind mentorship program by HSBC Bank and TiE Dubai that encouraged startups to transform business in a sustainable way (in their case healthcare).
“If it can be automated, it should be automated”
TachyHealth delivers value-based healthcare solutions to help reduce inefficiencies in the system for both healthcare providers and payers (hospitals and insurance companies).
One such inefficiency can be seen in relation to medical claims - like invoices, AbouElKhir said, only more complex as they contained medical information as well as financial.
Claims are generated by the hospitals with each insured patient encounter and sent to the insurance company, who have the function to audit and rationalise the payment and claims before paying out accordingly.
In the event of denial, however, a reason must be provided. It’s a cycle happening the world over, and most insurance companies are using manual methods or rule-based systems through software from 20-30 years ago – a clear opportunity to optimise.
To expedite the process, TachyHealth built a hybrid model that combines the best of machine learning algorithms with the best of the medical role-based system to automate the claim processing and adjudication. Today TachyHealth’s solution, named AiClaim, is able to automate 80 percent of the process.
In addition to the congestive engine, a real-time insights dashboard meanwhile offers clients easy insight into the analysed claims data, to inform actions for better clinical and financial outcomes.
On the hospital side meanwhile, TachyHealth is piloting two solutions.
The first, AiGuide, is an AI software for physicians that eliminates inefficiencies at the electronic medical records (EMR) point of entry.
Studies show that clinicians spend much of their time on documentation: if a physician sees a patient for 20 minutes, AbouElKhir revealed, they will chart the records for an additional 10-15 minutes after. In the event that the insurance company needs further explanation, the physician will need to revisit the records and resubmit accordingly in what is overall a cumbersome process.
AiGuide eases the burden through its easy to use experience that leverages natural learning processing and intelligent search optimisation built on evidence based guidelines from NICE and the BMJ. The physician can be guided through order sets and medications so that time spent charting is reduced in half – five minutes instead of 10.
The second hospital solution, AiCode, takes this streamlining a step further – through an automated medical coding system for medical records.
Medical coders sift through hundreds of pages of medical documentation each day to understand diagnoses and procedures before ensuring that the correct medical codes are assigned. Again, many hospitals are using manual ways or legacy software of doing this, AbouElKhir reminded.
TachyHealth instead plugs medical charts into the AiCode system for immediate conversion into medical code. This frees up coders’ time – they simply verify rather than assign codes - so that more records are charted per day and the insurance company is billed correctly.
AbouElKhir underlined the importance of security and privacy: through using Microsoft Azure datacenters in the UAE, any data containing protected health information stays within the country.
Pandemic pressure
Physician feedback on TachyHealth's AiGuide has been positive - doctors have reported satisfaction with seeing a user interface designed for and customised to a real practice. Sitting as a layer upon the EMR, the interface is a seen as a welcome addition to a system whose complexity has grown over the years (some systems date back to the 1970s).
Medical coders are also encouraged by AiCode solution, reporting a 70 percent elimination of redundant work through using it.
Business prospects look promising. According to AbouElKhir, TachyHealth has commenced work with an insurance company in Kuwait and is in talks with a major group in the UAE over the deployment of hospital solutions.
While the adoption curve was previously “slow”, the current pandemic has changed all that – private healthcare providers are in survival mode (with elective surgeries for example postponed) and looking to identify ways to improve and transform without the extra cost.
Realigning, not replacing, workers
With resistance to adoption no longer a problem, a challenge now is how to realign the workforce. AbouElKhir was keen to stress that automation wasn’t about replacing workers but improving processes so that people did a better job in less time.
Another challenge is a disconnect between research, software and healthcare industry communities in the way they collaborate and communicate with one another.
The research community is 10 years ahead, AbouElKhir explained, while the software community is focused mostly on traditional software engineering, and industry of course adopts. The disconnect is especially manifested, he argued, in the way machine learning advancements are translated back to industry.
To advance the conversation, TachyHealth has opened a “Tachy Transform Clinic” inviting clinics and hospitals to ask questions about how to use cutting edge technologies in practice.
The virtual clinic is offered through LinkedIn Live, the platform’s live video streaming service, using hashtag #TachyTransform, with the first launched on 11 April.
Unleashing innovation
Looking further ahead – 12 months from now, to when COVID-19 will (hopefully) be an event in the rear-view mirror – AbouElKhir sees a digital-first world better prepared for any emerging disease outbreak.
Anything considered redundant in healthcare today will be “cut off” or automated, particularly administrative costs not relating to the patient. R&D will be “nourished” better, solving problems such as contact tracing and population health big data analytics; public health institutions will receive higher funding than today; and the national healthcare expenditures will be increased, particularly in the Middle East.
AbouElKhir takes encouragement from developments in the US, where he believes a major change is happening on the regulatory front. He explained that traditionally years of work are required with the FDA for software, device or medication regulatory checks; this is now being accelerated.
For instance, the agency has announced a fast track approval for medical devices, software, and therapies that use AI. Also, the Department of Health & Human Services (HHS) finalised rules to provide patients more control of their health data which will allow for more data sharing and research based on this.
A leaner regulatory era combined with open healthcare platforms and patients owning their data would unleash powerful innovation – innovation that would furthermore benefit the patient, AbouElKhir believes.
It’s a compelling outcome.