Comments on: Use of artificial intelligence to draft a mini-HTA report on a new medical device belonging to class IIb-III

The article recently published by Messori et al. [1] described an artificial intelligence algorithm, based on the subscription version of ChatGPT, that generates the initial draft of a mini-HTA report for medical devices belonging to risk classes IIb or III. The process managed by ChatGPT begins with a standard request form compiled by the applicant institution, proceeds by performing two separate PubMed searches on safety and effectiveness, and finally generates a text containing a few paragraphs for each of the 5 remaining EUnetHTA domains (i.e., domains 5 to 9). The initial request is a PDF file; the final text is a Word DOCX file. In operational terms, ChatGPT acts by developing a script written in Python and then running this script to execute the various steps of this procedure. The final HTA report is managed as a confidential document with access restricted only to the applicants.

We have investigated this entire process in detail and identified the underlying Python script. This script offers a significant advantage: it can be run on a stand-alone basis, thereby eliminating the need for a paid ChatGPT subscription. In fact, the Python application is freely available online and can be installed on any computer with a Windows operating system at no cost. The installation instructions (in English or in Italian) and the specific Python script—which generates the exact same mini-HTA report as ChatGPT—are available from the authors upon request.

Other information

This article includes supplementary material

Corresponding author:

Andrea Messori

email: andrea.messori@osservatorioinnovazione.net

Disclosures

Conflict of interest: The authors declare no conflict of interest.

Financial support: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.