MedEssist’s Joella Almeida on going slow with AI in healthtech

joella almeida saas north
AI’s accuracy problem could spell trouble for patients and practitioners.

To MedEssist CEO Joella Almeida, Canada’s healthcare system is “fundamentally broken.”

“Canadians have a massive problem accessing healthcare right now. Fundamental wait lines at [emergency] rooms and walk-in clinics are just to the brim,” Almeida said.

She detailed the grim reality in a sit-down interview with BetaKit at SAAS NORTH 2023: forty-person lines at clinics in the early hours and years-long waits for a family doctor, scenarios that are likely familiar to most Canadians.

“If you just throw this into a healthcare setting, you are going to be impacting patients in a novel way that you probably did not expect.”

So how can tech help? Well, MedEssist is looking to tackle part of Canada’s healthcare crisis by providing pharmacies with education, technology and external partnerships to launch a more comprehensive healthcare site. 

“The whole point of this is creating new models of care that have never existed before, so there are systems where pharmacists, nurses, physicians can all collaborate together to actually deliver healthcare,” she added.

MedEssist employs various technologies in aiding pharmacies and has recently started incorporating artificial intelligence with its latest clinical support tool. This AI taps into a growing database of over 18,000 curated clinical and pharmacy-specific resources, assisting healthcare providers in answering clinical queries during practice.

The adoption of AI in the healthcare field is accelerating, offering new possibilities in diagnostics, treatment planning, and operational efficiencies. But Almeida described MedEssist’s approach as “really slow, very methodical, very dedicated.”

RELATED: #CDNtech talks AI disruption and the state of SaaS

Why? It’s not just a privacy issue, she explained. Almeida’s concern primarily revolves around accuracy and the potential risks that inaccuracies in large language models (LLMs) could present to patients.

“MedEssist is deployed in 500 pharmacies across Canada and the US, and we are serving about roughly one million patients,” she said. “The LLM that I’m using is 97 percent accurate. But guess what? Three percent inaccuracy is 10,000 patients. Is that a risk I’m willing to take?”

For Almeida, integrating AI has been a process of thorough vendor review, testing across various healthcare settings, and gradually adding complexity to ensure the AI is helpful to healthcare professionals while reducing the risk of harm to patients.

“If you just throw this into a healthcare setting, you are going to be impacting patients in a novel way that you probably did not expect,” she added.

Listen to Joella Almeida’s full conversation below, and catch more conversations from SAAS NORTH 2023 on The BetaKit Podcast.

Isabelle Kirkwood

Isabelle Kirkwood

Isabelle is a Vancouver-based writer with 5+ years of experience in communications and journalism and a lifelong passion for telling stories. For over two years, she has reported on all sides of the Canadian startup ecosystem, from landmark venture deals to public policy, telling the stories of the founders putting Canadian tech on the map.

0 replies on “MedEssist’s Joella Almeida on going slow with AI in healthtech”