Dr. Kev Lim, CEO and co-founder of Qmed Asia, reveals in an interview with HealthTechAsia how the Malaysian digital health startup seeks to ensure accuracy and relevance in gen AI, while also laying the foundations through digitalisation.
Healthtech company QueueMed Healthtech Sdn Bhd (now Qmed Asia) emerged in Malaysia in 2019 out of a frustration familiar to many of us: queues. Dr. Kev Lim, a busy doctor with paediatric training, faced long and unpredictable waits at the clinic, from 9 in the morning until early afternoon, while bringing his daughter for follow-up appointments. “It was a terrible experience,” he recalled. “As a doctor I was always thinking of the best medications. As a patient, I had a different perspective altogether.”
Recognising the potential of technology in healthcare, he partnered with friends Dr. Tai Tzyy Jiun and Nic Tai after his daughter’s medical condition had stabilised. Together, they launched a user-friendly solution aimed at enhancing the patient waiting experience and expectations: a patient portal. The application enabled patients to monitor live appointment queues from their phones and receive reminders when their turn was approaching.
“It was a simple solution,” observed Dr. Kev. “Yet, there was no adoption in the market at the time. We pitched our solution to the Ministry of Health in Malaysia, but they weren’t keen as they didn’t like change,” he shrugged.
Then the pandemic struck. The outbreak proved a blessing in disguise.
“We got lucky,” he acknowledged. With the implementation of Standard Operating Procedures such as social distancing, Malaysia’s Ministry of Health contacted QueueMed and invited them to provide their solution to help alleviate patient crowding.
What started with four government clinics quickly expanded to 400 as doctors began recommending the digital health solution to their own colleagues.
It wasn’t long before doctors from private hospitals and clinics approached QueueMed with their own pain points, once they realised they shared common ground with the founding team.
“We started to offer more comprehensive solutions leading to our three main modules: Journey, Connect, and AI for Healthcare,” said Dr. Kev.
Digitalising healthcare
Journey is Qmed’s core offering and where the business began, aiming to streamline patient flow and enhance operational efficiency in healthcare.
Building on their queue management solution, which has seen a reduction in patient waiting times from approximately two hours to less than 30 minutes, Qmed also provides hospitals and clinics with self-registration kiosks to alleviate the registration bottleneck (Dr. Kev likened it to the self-service kiosks in McDonald’s stores). They also offer a teleconsultation module and remote monitoring solutions integrated into the patient journey.
Connect meanwhile encompasses Qmed’s medical device integration solutions, offered in partnership with CelcomDigi, Malaysia’s biggest telecommunications conglomerate, combining connectivity with clinical management solutions.
Dr. Kev explained that Malaysia lags behind in healthcare digitalisation, compared to the likes of South Korea, Taiwan, and Hong Kong. “Many devices are old school, for example radiology departments are still offering X-ray scans in CD format. Now let’s say the patient wants a second opinion. They can take the CD to another hospital, but they may not be able to read it.”
“It’s the same with data from patient vital monitors,” he continued. “The data doesn’t flow into the EMR seamlessly. Nurses often write down vitals on paper during patient monitoring and then manually enter them into the computer. There is a lot of delay and human error.”
Qmed’s answer to such delays is a gateway that communicates with radiology devices. The X-ray data flows into the gateway, and is pushed into a PAC server and sent to the cloud, the Qmed founder revealed.
With this, the radiologist can perform teleradiology from home or another hospital by accessing imaging studies stored in the cloud PAC and collaborating with healthcare teams remotely, enabling a faster turnaround time. The patient also has X-ray data stored in an app that can be shared with another hospital for a second opinion.
The same approach applies to patient vital monitors. Vital sign data flows to the portal and sent to the EMR with Early Warning Signs (EWS) to trigger alerts accordingly so that the caregiver is immediately informed.
Alongside expanding its services, the company ventured into markets beyond Malaysia, rebranding from QueueMed to its current name, Qmed Asia.
“We tailor strategies for various countries, identifying similar pain points in places like Singapore, Indonesia, and the Philippines,” Dr. Kev explained. “In Singapore, for instance, we’re launching our self-registration kiosks and queue management system.”
AI-driven guidance for patients
Healthcare’s digitalisation is paving the way for another opportunity: AI.
“Quite simply, without patient data, we cannot do AI. AI is trained with data,” Dr. Kev emphasised. “AI can only be implemented once a digitalisation layer is in place.” He highlighted as an example the need to digitalise X-ray images so that AI algorithms can analyse them.
In the past two years, Qmed has ventured into AI in Malaysia in partnership with Intel, focusing on assisting radiologists in interpreting chest X-rays and aiding pathologists in detecting cancers.
The company is also piloting generative AI with a tool, Qmed NORA, enabling patients to check their symptoms at home. “In Malaysia, if patients are not feeling well with flu or fever they typically go to the emergency department, contributing to congestion,” said Dr. Kev.
Qmed NORA will triage patients, determining whether they should visit an emergency department, consult a GP, or manage symptoms at home. The tool also enables patients to upload photos of conditions such as rashes for analysis, and can use location data to aid in detecting diseases like dengue and COVID-19.
In Southeast Asia, large language models like ChatGPT are not widely adopted in healthcare, according to Dr. Kev. This is largely due to the region’s linguistic diversity, where English is not universally spoken. The models occasionally generate hallucinated information, he added, while there is the issue of patient data potentially going directly to companies such as OpenAI.
Another limitation highlighted by the Qmed CEO is that off-the-shelf models don’t provide localised responses when discussing symptoms. “Take dengue, for example,” he said. “ChatGPT relies on WHO international guidelines rather than country-specific clinical practices for dengue, such as those in Malaysia.”
Furthermore, large models lack citation of references, which could pose a challenge for doctors seeking to verify the information.
Qmed trained the large language model for Qmed NORA themselves, using Malaysia’s Clinical Practice Guidelines. “Now it understands Malaysian healthcare and can provide specific citations based on the guidelines,” Dr. Kev explained.
Empowering doctors
The Qmed founder sees AI ultimately serving as a powerful healthcare assistant. “AI won’t replace doctors; it will act as a co-pilot,” he said. “Next year, doctors’ laptops will become more powerful still, incorporating Small Language Models for medical scribing.”
With this technology, a doctor and patient’s conversation can be automatically transcribed by generative AI on the laptop and shared with the doctor in the standard consultation format, requiring only a quick check by the physician.
Patients, meanwhile, will benefit from personalised healthcare. “Every patient will receive highly customised treatments because we will gather more data points from hospitals, such as heart rates, sleep patterns, X-rays, and their medical history,” he offered, pointing to the rise of smart devices and wearables. “We will know more about our patients than we know about ourselves.”
Amidst a patient data explosion, Dr. Kev emphasised the need for a more open approach to regulation. “Until now, teleconsultations have lacked a regulatory framework. It took 10 years to reach this stage. However, in the era of generative AI, significant changes are occurring.”
Sandboxes are the solution, he suggested, offering the example of the Government of Singapore in Southeast Asia. “There is a sandbox available for any company with an innovative digital health solution to test on real patient data and demonstrate its safety before market entry. This is the way forward, instead of having a hospital conduct a proof of concept only for the Ministry of Health to later disapprove.”
From seeking permission to cut queue times in clinics, to validating AI systems in a sandbox environment, it seems that digital health practitioners like Qmed Asia have come a long way.
Leave a Reply