AI Clinical Tools for Pulmonary and Critical Care Workflows

Every minute counts in the ICU.

Between managing ventilator settings, adjusting sedation, coordinating with consultants, and documenting everything in Epic, we're constantly context-switching. In the outpatient pulmonology clinic, it's a different kind of chaos—back-to-back patients with complex sleep studies, interstitial lung disease management, and endless prior authorizations.

What if we could reclaim some of that time?

Three AI clinical tools have changed how I work: Doximity, OpenEvidence, and UpToDate Expert AI. Each one solves a different workflow problem, and together they've made me faster and more confident at the bedside and in clinic.

In this article, I'll break down how these tools work, where they fit into pulmonary and critical care workflows, and which situations call for each one.

OpenEvidence is a generative AI medical search platform built specifically for physicians. You ask a clinical question, and it gives you a fast, evidence-based summary pulled from high-quality, peer-reviewed studies.

What sets OpenEvidence apart is its integration with the journals that matter most to us. Multi-year partnerships with NEJM and JAMA Network mean you're not just getting summaries—you're seeing actual tables and figures from the original research.

I use OpenEvidence when I need to answer a specific clinical question quickly. A patient with severe ARDS on high PEEP—should I prone them? What does the evidence say about timing? OpenEvidence pulls the relevant trials, synthesizes the findings, and shows me the data in seconds.

It's also fantastic for those niche questions that don't have a clear guideline. PICO-style queries—like "Does high-dose vitamin C improve outcomes in septic shock?"—get thoughtful, evidence-based responses that cite contradictory studies when they exist.

The bottom line: OpenEvidence is best when you need to quickly synthesize medical literature on a focused clinical question.

UpToDate has been our go-to clinical reference for years. Expert AI takes that trusted content and adds a conversational interface that actually thinks like a physician.

What I love about Expert AI is the transparency. Every response shows you the specific UpToDate topics it used, the assumptions it made about your question, and the step-by-step reasoning it followed. This addresses my biggest frustration with most AI tools: the black box problem.

When I ask Expert AI, "How should I manage a patient with heart failure with reduced ejection fraction who's intolerant to ACE inhibitors?" it doesn't just give me an answer. It shows me why it chose that answer, which UpToDate topics informed it, and what assumptions it made about the patient.

I use UpToDate Expert AI when I need clinical decision support that mirrors how an expert consultant would think through a problem. It's particularly useful for complex management questions where multiple factors need to be weighed.

The bottom line: UpToDate Expert AI is best for complex clinical reasoning and decision support with full transparency.

Doximity's acquisition of Pathway Medical brought structured medical content—guidelines, drug information, and landmark trials—directly into a platform that 80% of U.S. physicians already use.

Pathway's strength is its structured dataset. When I need quick access to a specific guideline or drug dosing information, Doximity delivers short, citation-backed responses without the noise.

I use Doximity when I need fast, reliable answers to straightforward questions. What's the current GOLD staging for COPD? What's the dosing for nintedanib in IPF? Doximity gives me the answer in seconds, properly cited, and I can move on.

The bottom line: Doximity is best for quick reference lookups—guidelines, drug information, and clinical protocols.

Let me walk you through how I actually use these tools in practice, starting with the ICU and then moving to the outpatient pulmonology clinic.

Critical care is all about rapid decision-making with incomplete information. We need tools that don't slow us down.

Morning rounds scenario:

I'm rounding on a patient with COVID-19 ARDS on day 5 of mechanical ventilation. The team is discussing whether to initiate prone positioning.

This workflow—literature synthesis → clinical reasoning → quick verification—takes maybe 5 minutes total. Without these tools, I'd be hunting through PubMed, reading full papers, and double-checking multiple UpToDate topics. That's 30+ minutes I don't have during morning rounds.

Procedure preparation scenario:

I'm about to perform a thoracentesis on a patient with prolonged mechanical ventilation and large pleural effusions.

The key advantage in the ICU is speed. These tools let me access high-quality information without breaking my workflow. I'm asking a question and getting an answer.

Outpatient pulmonology demands a different kind of workflow support. We have more time per patient, but the questions are often more nuanced and require deeper dives into the literature.

Interstitial lung disease clinic scenario:

A 62-year-old patient with newly diagnosed idiopathic pulmonary fibrosis asks about starting antifibrotic therapy. She's read about nintedanib and pirfenidone online and wants to know which one is better.

This workflow helps me have a more informed conversation with my patient. I'm not guessing or relying solely on memory—I'm synthesizing current evidence in real time and translating it into actionable guidance.

Sleep medicine scenario:

A patient with severe OSA on CPAP isn't tolerating therapy and asks about alternatives.

Again, the workflow supports a higher-quality clinical conversation. I'm not just offering alternatives—I'm explaining the evidence behind them.

Here's how I think about tool selection:

Use OpenEvidence when:

Use UpToDate Expert AI when:

Use Doximity when:

None of these tools replace clinical judgment. They're not making decisions for us.

What they do is compress the time between question and answer. Instead of spending 20 minutes hunting through literature or navigating dense reference materials, I get evidence-based guidance in seconds. That time savings compounds—over a week, a month, a year.

More importantly, these tools make me a better clinician. I'm asking more questions, exploring more evidence, and having richer conversations with patients and colleagues. I'm not cutting corners—I'm being more thorough, faster.

The key is knowing when to use which tool. OpenEvidence for literature synthesis, UpToDate Expert AI for clinical reasoning, Doximity for quick references. Each one has a role, and together they cover most of what I need in pulmonary and critical care practice.