---
title: "Doctors Stopped Googling. Most Pharma Marketers Haven't Noticed."
date: "2026-10-22"
summary: "A new JAMA study shows 952,000 US clinicians sent 271 million questions to OpenEvidence — and what that means for how pharma thinks about HCP search."
tags: [aeo, ai-search]
---

I spend a lot of time looking at how pharma brands show up in search and AI answers. Almost all of that work has focused on two audiences: patients typing into Google, and patients asking ChatGPT. A study published in *JAMA* last week made me realize I'd been underweighting a third.

Researchers looked at clinicians using OpenEvidence, an AI tool that answers clinical questions from the medical literature. Between January 2024 and June 2026, about 952,000 verified US clinicians sent it 271 million questions. Monthly users went from a few hundred to more than 700,000.

The number that stuck with me is a different one. A third of those questions were about treatment and management, compared with about 14% for diagnosis. Clinicians aren't using it to work out what a patient has. They're using it to decide what to do next.

That's the moment pharma marketing has always tried to reach, and it now happens inside a tool most brand teams have never opened.

A few things about this feel different from the usual "AI is changing search" story.

**The people asking aren't the ones brand plans are built around.** Nurse practitioners and PAs ask more questions per month than physicians do. Early-career doctors use it far more than their share of the workforce would suggest. Many of these are HCPs a rep rarely sees.

**The answers don't come from brand websites.** OpenEvidence draws on published evidence, and this week Doximity said it's bringing UpToDate content into its own AI tool. The HCP site you spent a year optimizing may barely factor in. What matters is whether the label, the studies, and the guidelines say the right thing clearly.

**None of it shows up in your analytics.** A clinician who gets a full answer doesn't click through. If your dashboard reads HCP site traffic as a sign of interest, it's quietly missing more every month.

I don't think the answer is to panic or to rush to buy ads in these tools. The first step is simpler: ask them the questions your HCPs actually ask, read what comes back, and check whether it's accurate. Most teams I talk to haven't done that once.

I wrote up the full findings and a five-step plan for pharma teams on the PharmaForward site: [OpenEvidence Is the New HCP Search Engine. Here's What Pharma Should Do About It.](https://pharmaforward.com/thinking/openevidence-pharma-hcp-ai-search/)