About
Vital.io: Guides patients during an ER or hospital stay. Advanced AI predicts wait times,…
Articles by Aaron
-
How to do a LinkedIn AMA
How to do a LinkedIn AMA
With an increase in posts & content from business leaders, LinkedIn has become a place to learn from the best. Only…
14 Comments
Activity
8K followers
-
Aaron Patzer shared thisAlmost none of my CEO friends have visited me in New Zealand (lived here 10+ years). Fortunately WSJ was kind enough to help my social life with a feature piece on our farm: https://lnkd.in/eYzRNnNK If we're friends, colleagues, or someone I've known for ages, come, I'll host. You'll feel healthier for it.After Selling His Company, He Could Have Lived Anywhere. He Chose New Zealand.After Selling His Company, He Could Have Lived Anywhere. He Chose New Zealand.
-
Aaron Patzer shared thisCourtney's story is tragic, but all too common - cancer or pre-cancer found years in advance, but never disclosed. A few years back, Courtney left hospital after complications from Crohn's. She was happy a bowel bleed was under control and she could go home. But lingering on page 4 of CT scan was an indication of a possible cancerous lump. No one disclosed it. No one followed up. And sadly, it was a cancer. We built Vital Guard because 60-70% of "incidental findings" (ones you weren't looking for) are never disclosed or followed up on - mostly because hospitals simply didn't have the manpower to do anything. Vital automates everything, from AI that detects neoplasms and aneurysms, to searching through doctors notes and follow-up appointments for signs of disclosure and follow-up, to acknowledge and education for the patient, to automating clinician outreach for the highest risk cases.Aaron Patzer shared thisOur CEO, Aaron Patzer, built Vital Guard because of someone he lost. Her name was Courtney. She was 29. A finding on her scan was documented, but no one told her about it. By the time it was discovered, it was too late. This is the story of why Vital Guard exists. Watch it here: https://lnkd.in/gg393Tz2
-
Aaron Patzer shared thisWe've already saved lives with this product. 60-80% of neoplasms & aneurysms found "incidentally" during an ER visit are never disclosed to the patient. For example, during our pilot, we found a Bosniak 4 lesion the patient was never aware of. Our AI found it in radiology, then scanned doctors notes & discharge to see if it had been disclosed. Since it was serious, it was marked for human outreach (vs. simple text message with education & acknowledgement). Patient started chemo literal weeks later. Other companies (Eon is likely the best) identify incidentals. But no one is checking whether they've been disclosed to patients, and automating the follow-up. In our pilot, a single radiology PA can manage an entire region of hospitals. 1 FTE instead of an after-care team you don't have funding for. This is how AI should be used.Aaron Patzer shared this🚀 Introducing Vital Guard™ Vital Guard is an AI-driven solution that protects patients and physicians by ensuring incidental findings don't fall through the cracks. ✔ Automatically prioritizes uncommunicated findings ✔ Supports closed-loop patient outreach ✔ Reduces clinical risk while saving staff time Because discovering a finding isn't enough — communication matters. 👉 Learn more: https://lnkd.in/gCMVzTqy #IncidentalFindings #PatientSafety
-
Aaron Patzer shared thisOn an health system AI Committee and secretly haven't a clue how AI actually works under-the-covers? We developed tools clinicians, operations and technical people can use. Happy to have worked with individuals from Stanford, MedStar, HCA, CHLA, Hackensack, IU, Matter, and more here. Big tech (won't name names) actually joined us for a hot-minute...but wanted to own all the tools, and all the IP. We want this to be unbiased, non-profit, a standard to be used by anyone with no cost/licensing or ulterior motive.Aaron Patzer shared thisToday marks the launch of the AI Care Standard™ for Patient Communication. AI communicates directly with patients through portals, chatbots, care navigation tools, and automated outreach. Yet healthcare has lacked a clear, consistent way to evaluate whether AI systems meet the same standards expected of clinician–patient communication — until now. The AI Care Standard™ establishes these guardrails. Developed by a select group of health system leaders, patient safety experts, clinicians, and AI practitioners, the Standard defines what safe, responsible patient-facing AI must deliver in real-world care — with clear expectations for accuracy, clarity, transparency, and accountable governance. The Standard is anchored by 10 Core Pillars that define what responsible patient-facing AI must deliver in practice, and an accompanying Evaluation Framework tool that allows organizations to assess real systems against those expectations. Together, the Core Pillars and Evaluation Framework move the conversation from abstract AI principles to practical accountability in real-world care delivery. And it provides a shared reference point the industry can rely on as patient-facing AI becomes integral to everyday care. Built to protect patients. Designed to support clinicians. Created to reduce risk and uncertainty as innovation continues to advance. Learn more about the AI Care Standard™ → https://lnkd.in/g4mJ5DEs #AICareStandard #PatientSafety #PatientFacingAI #HealthcareAI Raj Ratwani PhD Bridget Duffy, MD Terry Adirim, MD, MPH, MBA Steven Collens Beth D. Dev Dash MD Christine Dow Karen Drenkard, PhD, NEA-BC, RN Kathy English Alistair Erskine MD MBA jeffrey Gold MD Carole Hemmelgarn Tina HsinTing Liu,CPA,FACHE Kelly Nye Aaron Patzer Kelsey Powell Jeremy Rogers Marla Sammer MD, MHA, FAAP Janae Sharp FHIMSS Michael A. Rogozinski RN Justin Schrager MD Nicholas Sterling, M.D., Ph.D. Julie Wheelan Adrienne Woods MPH
-
Aaron Patzer shared thisToday we launch AICareStandard.com - a set of scoring tools & principles to ensure patient-facing health AI is safe to use. This is an industry collaboration with individuals at MedStar, HCA, Stanford, Highmark, IU, Hackensack Meridian, Children's Los Angeles, Texas Children's, patient safety advocates like Patients for Patient Safety, directors or former directors of AONL, the VA and Cleveland Clinic. This is to be used by health systems and tech vendors alike. Because right now, big tech is in an arms race, almost regardless of patient safety. And at the same time, health systems are notoriously conservative, with chatbots that almost reflexively say "Go to the ER or call 911" for anything. This is a practical middle ground, from the actual experts both building and evaluating healthcare AI in the real world. Use it for internal tools. Insist your vendors use it. Pressure big tech to use it - they initially joined the coalition, then bailed when they couldn't own all of the IP. This is a free, non-profit collective, beholden to no one. Proud of work like this. It will save lives? That's not hyperbole. There are dozens of deaths linked to chatbots: https://lnkd.in/enxpgkbj You wouldn't let an unlicensed clinician practice medicine. And while this isn't a "license" for AI, it's the start of a free-market practice (think Hitrust or SOC2) for everyone in healthcare.
-
Aaron Patzer shared thisComing Feb 11th, in collaboration with individuals from MedStar, Stanford, IU, HCA, Texas Children's, Highmark, AARP, patient advocacy groups and more, we'll be announcing guidelines and tools to ensure AI safely interacts with patients. There are too many chatbot deaths/suicides from big tech, and at the same time a too conservative approach by health systems. This sits in the practical, yet patient safe middle.Aaron Patzer shared thisWe would never allow an untrained clinician to provide medical guidance to a patient. And yet AI is already doing exactly that — quietly, at scale, every day. These messages shape understanding, decisions, and behavior. They carry real clinical consequences. Healthcare has clear expectations for humans who speak to patients. It’s time those same expectations applied when AI does. On Feb. 11, that line gets drawn. #PatientSafety #PatientCommunication
-
Aaron Patzer shared thisWhen a journalist had an emergency visit…and likes your software so much they spontaneously write an article… Here’s to guiding millions of patients + family through the ER, translating doctor speak notes, and explaining lab results & post discharge steps like a human.Aaron Patzer shared thisThank you to Teta Alim and Healthtech Magazines for capturing the innovation and real-world value our Emergency Department application is delivering at MedStar Health. Serving over 7 million patients across the US, we're committed to transforming how healthcare systems support patients during their most vulnerable moments. Proud to partner with forward-thinking organizations like MedStar! Read the full article here: https://lnkd.in/es4tX-85Improving Patient Communication and Transparency in the EDImproving Patient Communication and Transparency in the ED
-
Aaron Patzer shared thisOne of the most fun interviews I've done (at HLTH). We discuss why almost no one in healthcare is making consumer-grade, human-centric software. Shout out to Bridget Duffy, MD ... when asked who I have been taking inspiration from in healthcare as of late, it's her!Aaron Patzer shared thisSerial founder Aaron Patzer shares the very human goals behind his new health tech solution, Vital.io from HLTH USA . Vital’s mission is simple: fix broken communication in healthcare. Patzer realized patients and families feel lost because doctors speak “left cerebral infarction” and patients hear “huh?” With software that translates results and medical language into human terms, Vital delivers guidance, not just data. The best part? They guide millions of patients post-discharge—no more silence after you leave the ER. This is human-centered tech that solves a real, daily pain for both patients and professionals. Link to episode in the comments. #HealthTech #PatientExperience #AI #Founders #PersonalizedCare #DigitalHealth #PlainLanguage #Innovation #HumanCenteredDesign
-
Aaron Patzer shared thisNow that we're serving 7M patients a year for ER & Inpatient, we thought we'd expand to urgent care. AI predicted wait times, human descriptions of lab results (e.g. "Measure strain on your kidneys" instead of BUN = 46 mcg/dL), clear post-discharge follow-up instructions. Test sites over the last 6 months have 60%+ patient utilization, and leakage prevention worth ~$250k/urgent care if you're a health system that owns nearby specialty practices and/or hospitals.Aaron Patzer shared thisProud to share big news: Vital Urgent Care is officially live—bringing real-time transparency to urgent care with no app, no login. Patients get: • an SMS link to a mobile dashboard • Accurate wait times (94%+) • Place-in-line tracking • Real-time visit updates • Plain-language education • AI-powered discharge checklist All integrated with major EHRs. Because the experience matters as much as the medicine. Details in the full press release here: https://lnkd.in/d5y3f9fEVital Launches Real-Time Urgent Care Platform That Elevates Patient Experience and Boosts Satisfaction -- No App or Login RequiredVital Launches Real-Time Urgent Care Platform That Elevates Patient Experience and Boosts Satisfaction -- No App or Login Required
Aaron Patzer commented on a post
6d
Vital.io can do it. 65% patient adoption on real-time guidance through ER, Inpatient, Urgent Care. 35-50% reduction in LWOBS by setting expectations on wait time (using AI), walking people through what lab results mean (average 30+ min in a web-app). AMA/LWOBS is worth $500-1500/visit, so a 3% => 2% reduction is very quantifiable. 30% increase in follow-up retention within system if you explain in patient terms why it's necessary, remind them, find an alternative if the one given in acute care doesn't take their insurance, etc. It's taken 5+ years, but we can contractually guarantee PEX scores, revenue, and outcomes now.
Aaron Patzer commented on a post
1mo
Also, consciously didn't post the full video on LinkedIn. It's frankly too sad, and felt a bit too much like monetizing a friends death. It feels a bit weird to even mention it in the context of a "product launch" but this was one of the big motivating factors for me in prioritizing the engineering and investment in development here.
-
Aaron Patzer liked thisAaron Patzer liked thisLate to the party for late interaction models... but it's never too late! I've *finally* taken the time to understand the ColBERT paper and interested to see what I can do with these ideas to improve my own embedding models.
Experience & Education
-
Vital Software
View Aaron’s full experience
See their title, tenure and more.
or
Publications
-
Prediction of Emergency Department Hospital Admission Based on Natural Language Processing and Neural Networks
Methods Inf Med
Abstract
OBJECTIVE:
To describe and compare logistic regression and neural network modeling strategies to predict hospital admission or transfer following initial presentation to Emergency Department (ED) triage with and without the addition of natural language processing elements.
METHODS:
Using data from the National Hospital Ambulatory Medical Care Survey (NHAMCS), a cross-sectional probability sample of United States EDs from 2012 and 2013 survey years, we developed several…Abstract
OBJECTIVE:
To describe and compare logistic regression and neural network modeling strategies to predict hospital admission or transfer following initial presentation to Emergency Department (ED) triage with and without the addition of natural language processing elements.
METHODS:
Using data from the National Hospital Ambulatory Medical Care Survey (NHAMCS), a cross-sectional probability sample of United States EDs from 2012 and 2013 survey years, we developed several predictive models with the outcome being admission to the hospital or transfer vs. discharge home. We included patient characteristics immediately available after the patient has presented to the ED and undergone a triage process. We used this information to construct logistic regression (LR) and multilayer neural network models (MLNN) which included natural language processing (NLP) and principal component analysis from the patient's reason for visit. Ten-fold cross validation was used to test the predictive capacity of each model and receiver operating curves (AUC) were then calculated for each model.
RESULTS:
Of the 47,200 ED visits from 642 hospitals, 6,335 (13.42%) resulted in hospital admission (or transfer). A total of 48 principal components were extracted by NLP from the reason for visit fields, which explained 75% of the overall variance for hospitalization. In the model including only structured variables, the AUC was 0.824 (95% CI 0.818-0.830) for logistic regression and 0.823 (95% CI 0.817-0.829) for MLNN. Models including only free-text information generated AUC of 0.742 (95% CI 0.731- 0.753) for logistic regression and 0.753 (95% CI 0.742-0.764) for MLNN. When both structured variables and free text variables were included, the AUC reached 0.846 (95% CI 0.839-0.853) for logistic regression and 0.844 (95% CI 0.836-0.852) for MLNN.Other authors -
Patents
-
System and method for building a script for a web page using an existing script from a similar web page
Issued US 9,996,441
-
Multiple alignment genome sequence matching processor
Issued US 6,983,274
-
Feedback cycle detection across non-scan memory elements
US 6,986,114
-
Replicant simulation
US 8,161,448
-
System and method for building and repairing a script for retrieval of information from a web site
US 9,779,007
-
System and method for categorizing credit card transaction data
US 7,840,456
-
System and method for providing price information
US 9,286,639
-
Temporal replicant simulation
US 7,979,820
Recommendations received
6 people have recommended Aaron
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content