PP19 - Initial Exploration Of Generative Artificial Intelligence - Identified Tasks, Team Perspectives, And Early Implementation Insights

Sep 29, 2026

Artificial intelligence in health technology assessment report development is addressed through automation, standardization, and responsible training for generative AI integration. Conducted at the Agency for Health Quality and Assessment of Catalonia (AQuAS), the study reviewed technical literature and leading tools (e.g., ChatGPT, Copilot), then surveyed the evaluation team to capture knowledge, attitudes, perceived utility, and adoption barriers. Results supported practical best practices, with high engagement and strong satisfaction.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages1

Reading Time3 min

AccessFree to download

Introduction

Methods

Results

Conclusions

PP18 Effectiveness And Safety Of Long-Acting Agents-Based Dual Therapy In HIV-1/AIDS Patients: A Network Meta-Analysis Of Observational Studies Daniela Gorski ([danielagorski@ufpr.br](danielagorski@ufpr.br)), Bruno Gonçalves Ferreira, Helena Hiemisch Lobo Borba, Fernanda Stumpf Tonin and Roberto Pontarolo Introduction: Most treatment regimens for people living with HIV (PLHIV) require the intake of multiple daily pills, which can pose challenges to adherence. As an alternative, long-acting (LA) dual regimens have gained attention for their potential to simplify treatment. Given the importance of evaluating the long-term effects of LA regimens, this study aimed to synthesize evidence on their effectiveness and safety. Methods: A systematic review was performed following the Cochrane Collaboration recommendations and the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) guidelines, with searches in PubMed, Scopus, and Web of Science (July 2024)(PROSPERO-CRD42024570754) . Longitudinal studies evaluating any dual-therapy antiretroviral regimen with LA agents (islatravir, cabotegravir, lenacapavir, rilpivirine, albuvirtide) for PLHIV were included. For each outcome of interest, data were pooled using bivariate meta-analysis (discontinuations due to adverse events) and network meta-analysis (NMA) (virologic failure [VF]), with results presented as risk ratios (RR) with 95 percent credibility intervals (CrIs); heterogeneity was assessed using the I-squared statistic. The surface under the cumulative rating curve analysis (SUCRA) was also conducted for the NMA. Results: Thirty-two studies (2016 to 2024) evaluating 11 treatments were included. Four cohorts (n=2,570) were assessed in the NMA of VF, where the rilpivirine/dolutegravir (RPV/DTG) regimen was superior to non-LA regimens dolutegravir/lamivudine (RR 􀀁1.60, 95% CrI: 􀀁4.08, 0.91) and dolutegravir/doravirine (RR 􀀁0.42, 95% CrI: 􀀁1.68, 0.68), with SUCRA probabilities of 9.6 percent, 31 percent, and 59.5 percent, respectively. The triplet regimens dolutegravir plus tenofovir/emtricitabine and dolutegravir plus abacavir/lamivudine had the highest probabilities of leading to CD4+ count increases at six to 24 months (65.6 to 85.7% and 65.8 to 84.4%, respectively), followed by RPV/DTG (59 to 60%) and dolutegravir/lamivudine (46 to 53%) . However, RPV/DTG regimens were considered safer than dolutegravir/lamivudine (RR 0.53, 95% CrI: 0.32, 0.89) regarding discontinuation due to adverse events. Conclusions: This updated synthetized evidence suggested that the RPV/DTG regimen is a promising alternative for simplifying the treatment of PLHIV, demonstrating comparable effectiveness to the dolutegravir/lamivudine and dolutegravir/doravirine regimens, while offering better tolerability. However, considerations of drug accessibility and cost must be carefully weighed in the decisionmaking process in different settings. PP19 Initial Exploration Of Generative Artificial Intelligence: Identified Tasks, Team Perspectives, And Early Implementation Insights Maria Dolors Estrada Sabadell, Joan Segur-Ferrer, Roland Pastells-Peiró, Jessica Ruiz-Baena and Rosa Maria Vivanco-Hidalgo (rvivanco@gencat.cat) Introduction: Artificial intelligence (AI) in health technology assessment (HTA) report development addresses three needs: accelerating processes via automation, standardizing tools for quality and consistency, and training to responsibly integrate generative AI. Conducted at the Agency for Health Quality and Assessment of Catalonia (AQuAS), this study identified tasks that generative AI can enhance and evaluated the team’s perceptions, knowledge, attitudes, and experience with these tools. Methods: A review of technical literature and an exploration of leading tools, including ChatGPT and Copilot, were conducted to identify relevant functionalities and their potential applications in HTA. Additionally, an online survey was designed to assess the team’s knowledge and use

FAQs

What needs does generative AI target in HTA report development at AQuAS?

It targets accelerating processes via automation, standardizing tools for quality and consistency, and training teams to responsibly integrate generative AI.

Which tools and approaches were used to explore generative AI functionality?

The study reviewed technical literature and explored leading tools such as ChatGPT and Copilot, then used an online survey to assess knowledge and use across multiple sections.

What were the main findings on team adoption and usability?

Participants developed an initial best-practices guide, reported high satisfaction, used the tools for daily tasks (including text drafting, information retrieval, and PDF reading), and nonusers expressed willingness to adopt them.

Sep 29, 2026

Upload by River Wang