"Are Drag Queens the Future of Healthcare? Healthvana's Bold Vision Explored"

How is Healthvana Engaging People About Their Health?

In an unexpected and innovative approach: Drag queens.

“Drag queens embody acceptance and authenticity,” said Gabriella Palmeri, VP of Partnerships at Healthvana, during her presentation at VB Transform. “You're receiving empathetic and candid responses, wrapped in a fun persona that helps to de-stigmatize health issues.”

With this unique strategy, Healthvana is currently serving tens of thousands of patients and aims to expand its reach to one million by the end of 2024.

“We help digitize the healthcare experience,” said Ramin Bastani, CEO of Healthvana. “You may not have heard of us, but our patients truly love our service.”

Reaching Patients Where They Are

Bastani opened the session with a poll:

“How many people here use a patient portal?” (Hands were raised.)

“How many of you actually like your patient portal? Any fans in the room?” (Few hands went up.)

“Okay, so they’re mostly disliked,” he concluded.

Healthvana’s platform aims to revolutionize this perception, claiming to be one of the largest patient-facing AI systems in the U.S. This AI bot is linked to patients’ health histories, providing instant responses based on past interactions.

“As a patient, it knows your health history and can converse in any language, with responses delivered instantly,” noted Bastani.

Upon logging in, patients have the option to engage in a traditional human conversation, chat with a standard AI, or interact with the drag queen AI.

Palmeri shared that patients use the platform to engage in serious discussions, receiving immediate answers to urgent questions about medications. The AI can also entertain lighter queries—such as, "Should I be worried about it raining men?"—responding with, "If it's raining men, grab your most fabulous umbrella and enjoy the blessing!"

Remarkably, Palmeri pointed out that 80% of patient interactions occur with the drag queen persona. Additionally, 25% of conversations happen outside clinic hours, allowing Healthvana to reach patients who are typically hard to engage.

A Slow, Deliberate Build

Bastani emphasized that healthcare providers often struggle to manage the volume of patient messages and require AI support. “We didn’t create this technology out of thin air; it was developed to address a real need,” he explained.

Before the platform was launched, over 100 clinicians rigorously tested it through a careful rollout. Palmeri explained that this was necessary for regulatory compliance and to ensure patient safety and comfort, as interactions with AI differ significantly from those with human providers.

The initial user base comprised just five patients, but it has now expanded to 50,000 across 15 states who are part of Healthvana’s early access program.

Partnering with OpenAI, Healthvana emphasized an iterative approach, continuing to refine the platform based on thorough testing and data from over one million patient-provider messages.

“We believe we’re ahead of many competitors in evaluating real patient interactions,” Bastani added, clarifying that Healthvana maintains a zero-retention policy with OpenAI, ensuring that no patient-provider conversations are stored.

Three Categories of Rigorous Evaluation

Healthvana’s success is largely attributed to its meticulous evaluation process. Palmeri outlined that every AI-generated message is reviewed by two content moderators within 24 hours, assessing accuracy, completeness, and contextual relevance.

“Was the response correct? Complete? Contextually appropriate?” she noted.

Messages that receive a score below five—where five indicates optimal performance—enter a feedback loop for improvement.

Sam Warmuth, Healthvana’s CPO, detailed three critical evaluation categories: model evaluation to identify the best AI framework, single-turn evaluation to determine if the AI resolves user queries, and multi-turn evaluation, which examines the ongoing interaction between the AI and user.

“Multi-channel evaluation involves a continuous exchange where each output informs the next input,” Warmuth explained.

The platform incorporates multiple models interacting with each other, managing data input and communications to enhance user experiences.

Healthvana also employs chain point detection to identify inconsistencies in workflows, ensuring that patient interactions remain smooth and rewarding.

In summary, Healthvana's innovative approach, blending humor and empathy through a drag queen AI, is revolutionizing how patients interact with their healthcare—making the experience not only informative but also enjoyable.