The Role of ChatGPT in Reducing Preoperative Anxiety in Patients Planning to Undergo Rhinoplasty

Dündar G, Çelikal Ö, Çiçek MM.

PubMed 42298163 Publisher (DOI)Free full text (PMC)

What this paper says

In a randomised trial of 134 rhinoplasty patients, talking to ChatGPT about their concerns lowered pre-surgery anxiety scores more than reading a standard society information sheet.

Overview

The authors ran a single-blind randomised controlled trial. The control group (n = 66) read the Turkish Society of Facial Plastic Surgery's rhinoplasty patient information document; the study group (n = 68) discussed their surgical concerns with ChatGPT-3.5. Anxiety was measured with the State-Trait Anxiety Inventory before and after.

Sections of note

Baseline state-anxiety medians were similar: 36 (control) versus 39 (ChatGPT), p = 0.573.
After the intervention, control fell to a median of 31 and the ChatGPT group to 28 (both p < 0.001).
Median change was -6.0 with ChatGPT versus -4 with the document (p = 0.035).
Level of Evidence I.
The authors note the chatbot was a generic model, not one optimised for medical use.

What it means for a patient

Both a good written information sheet and an AI conversation reduced anxiety before surgery in this trial.
The interactive chat gave a modestly larger reduction, about two points more on the scale.
The trial measured anxiety only, not whether the information given was accurate.

Why this paper matters

It is a Level I randomised trial, uncommon in this field, testing a practical use of consumer AI in pre-surgical education. Limits: single centre, immediate post-intervention measurement only, no accuracy check of chatbot content.

Terms

Randomised controlled trial: patients are assigned to groups by chance so the groups are comparable.
Single-blind: the patients did not know which intervention was under test.
State-Trait Anxiety Inventory (STAI): a questionnaire separating current (state) anxiety from general (trait) anxiety.
Median: the middle value in a set of scores.

Summary written by rhinoplasty.cc from the abstract, 2026-09-08; not medical advice. The authors' own abstract follows.

Abstract

Objective: This study investigates the impact of an artificial intelligence (AI) chatbot (ChatGPT-3.5, OpenAI) on preoperative anxiety among patients planning rhinoplasty.

Design: In this single-blind, randomized controlled, prospective study, 134 patients were divided into control (n = 66) and study groups (n = 68). The control group read the rhinoplasty patient information document from the Turkish Society of Facial Plastic Surgery, while the study group interacted with ChatGPT, discussing their surgical concerns. Anxiety scores were measured using the State-Trait Anxiety Inventory (STAI) before and after the interventions.

Results: Initial State Anxiety Inventory (SAI) scores showed no significant difference between the groups, with medians of 36 (min-max: 21-60) for the control group and 39 (min-max: 21-58) for the study group (p = 0.573). Post-intervention, the control group showed a significant reduction in SAI scores to a median of 31 (min-max: 21-55; p < 0.001), with a more pronounced decrease in the ChatGPT group to a median of 28 (min-max: 20-53; p < 0.001). The median change in SAI scores was significantly greater in the ChatGPT group (- 6.0; min-max: - 24-6) compared to the control group (-4; min-max: - 16-8; p = 0.035).

Conclusion: Both traditional patient information and interaction with ChatGPT reduced preoperative anxiety in rhinoplasty patients, with ChatGPT proving more effective. Despite promising results, the limitation lies in the non-optimized nature of generic AI chatbots for specific medical uses, suggesting a potential for more specialized AI platforms in healthcare.

Level Of Evidence I: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .

Abstract as indexed by PubMed; the article is open access (PubMed Central).