VoxMedica closer look

VoxMedica

Improving the efficiency and empathy of the care team by allowing them to focus on the patient, not the clinical note.

What % of physicians are reporting about burnout in 2023?

Wrong answer

53% of Medscape’s 2023 study doctors said they were stressed

Try again

PROBLEm

31%

of clinicians cited “paperwork” contributes most to their feelings of exhaustion

WE ARE IN THE ERA OF PROVIDERS' BURNOUT

40%

I'm from the family with two generations of doctors. And I saw it myself how much time of my mom's and granny's life the documetation took. And one more thing, even being raised among doctors, as a mom of a child with congenital pathology, I had to record the physician's consultations by myself, and then google for a long-long hours after it to get the basic understanding of what doctors said during the consultation.

of providers time is spent on documentation

1H ~4H

15%

of clinicians cited “paperwork” contributes most to their feelings of exhaustion

Play

of patients remember what is said during a consultation.

Real story from real doctor

SOLUtion

‘Every minute a doctor saves on paperwork is another minute dedicated to saving lives.’

target market and potential

“Communication is documentation” will become a new standard in traditional care settings.

The tools will shift the role of clinicians from content creators to content editors. Will this make healthcare more human?

DREAM

What Calm($2B) and headspace ($1B) did for meditation, we are going to do to reduce providers' burnout

Hypothesis: If we use AI to transcribe medical dictations, then the transcriptions will be as accurate as human transcriptions or even more accurate. It will be produced in a shorter time, which improve doctor-patient interactions with a focus on quality of care and patient engagement.

product

An AI-powered medical transcription service accurately converts voice-recorded medical consultations into text for future diagnostic or treatment recommendations. Doctors can focus solely on their patients instead of their notes. The perfect balance of efficiency and empathy!

What did we build in a day?

Dockerized, micro-service based, pluggable,MS Azure ready solution

2.Rails API

3.Upload to Azure

4.GPT-3.5 is triggered

1.Audio recording

Audio is downloaded

8.Success!

7.Share with provider

6.Summarization

5.Voice-to-text

Live Demo

COMPETItors

We are not alone, but we are unique

Note-taking technology is perhaps the most developed space.

Dreams, Ideas, Plans

Customized summary for the provider and for the patient

Privacy and security

Medical Codes Extraction

Health plan draft

Patterns identification

Missing data analysis

FUTURE

Build reliable, cost-effective in-house solution

Support clinical workflows that vary

Trainings for adoption rate

Internal processes integration

Empathy through personalization

Referrals

Next Best Actions

TRACtIoN

Simple math

EFFiciency

FACTS

growth

2h

Active listening and empathy are two key elements of building strong patient-provider relationships

$ >60,000

Daily provider's saved time

Possible increase in revenue seat/year

25-95%

Increased profit for 5% increase in customer retention rate

75%

Nearly 80% patients would recommend an empathic physician to other individuals

Reduction in documentation time

TEAM

"Let the magic happen!"

Yuliia Karpenko

Oleksii Novitskyi

Operational and inspirational magician

Ruby on Rails magician

Taras Lavruk

Ihor Kuibida

Ruby on Rails magician

Quality Assurance magician

SoftServe Healthcare & Life Sciences

yuliia.karpenko@teladochealth.com

Let's amplify care, clarity, and compassion in healthcare with Teladoc!

Thank you!

Patterns identification

• What happened to other patients like mine?
• Analyse symptoms and signs, compare them to previous patterns or cases
• A disease is recognised when the actual pattern fits
• Recommend provider on the next best actions
• Provide patient with addiitonal educational resources

Missing data analyses

• Only about 20% of patients are linked to a standard care guideline
• possible use of the data imputation methods
• Gather data to create/build an evidentiary basis
• Identify gaps and suggest the recommendations to providers on additional questions/labs/etc.

Referrals

• Prepare internal and external referrals for services needed/recommended
• Prioritize next best referral actions to take through omnichannel marketing tests and closed loop experiments to optimize yield from what you could do (Next Best Actions) and to best fit targeted segments of patients.

Support clinical workflows that vary

• Create ease of use and customizability to be adopted across the different specialities

Training for adoption rate

• Gather feedback from physicians
• Send follow-up instructions
• Set up trainings to increase the use of AI technologies

Customized summary for the provider and for the patient

• Generate templates for different medical services.
• Summary for provider differs from summary for patient: first is more technical, second is more "humanised" (with explanations, NBA suggestions, simpler terms, etc.)

Medical Codes Extraction

• Categorize (I and II) CPT Codes on-fly
• Dynamic calculations for member MH pdf generation

Next Best Actions

• Cross-product recommendations
• Checklist with
• Labs
• Meds
• Images
• Edu resources
• Reminders/schedulers

Health Plan

• Consolidated place with all available information on patient health
• NBA
• Graphs on the tracked data:weight, mood, meds, etc.
• Add behavioural science and/or gamification: checklists, achievements, etc.

Empathy through personalization

• Add block for non-medical notes about patient for providers
• Set reminders for providers before consultation to review this section
• Increase empathy and connection in doctor-patient communication

Privacy and security

• An additional layer of security with a pseudonymization algorithm that systematically removes all portions of the transcript that contain personally identifiable information. It will masks names, addresses, dates, phone and fax numbers, SSN, medical record numbers, health plan beneficiary, account, certificate or license numbers, vehicle, device or serial identifiers, URLs and IP addresses. The masked version is the one we end up giving to the LLM.
• Move ChatGPT from OpenAI to Azure to get HIPPA BAA

Internal processes integration

• Extend knowledge base to support CC agents with the most up-to-date information
• Gather and analyse analytics on the user requests, feature issues, etc.
• Real-time tips to medical staff in certain cases, including, for instance, advising a nurse how to process a patient complaining of neck pain or telling a 911 caller when to give CPR instructions