InformGen: An AI Copilot for Accurate and Compliant Clinical Research Consent Document Generation
Zifeng Wang, Junyi Gao, Benjamin Danek, Brandon Theodorou, Ruba Shaik, Shivashankar Thati, Seunghyun Won, Jimeng Sun
arXiv, 2 April 2025
Open Access
Abstract
Leveraging large language models (LLMs) to generate high-stakes documents, such as informed consent forms (ICFs), remains a significant challenge due to the extreme need for regulatory compliance and factual accuracy. Here, we present InformGen, an LLMdriven copilot for accurate and compliant ICF drafting by optimized knowledge document parsing and content generation, with humans in the loop. We further construct a benchmark dataset comprising protocols and ICFs from 900 clinical trials. Experimental results demonstrate that InformGen achieves near 100% compliance with 18 core regulatory rules derived from FDA guidelines, outperforming a vanilla GPT-4o model by up to 30%. Additionally, a user study with five annotators shows that InformGen, when integrated with manual intervention, attains over 90% factual accuracy, significantly surpassing the vanilla GPT-4o model’s 57%-82%. Crucially, InformGen ensures traceability by providing inline citations to source protocols, enabling easy verification and maintaining the highest standards of factual integrity.

Co-designing the consent process of critical care trials with patients and the public: Project protocol
Timo Tolppa, Arishay Hussaini, Vrindha Pari, Nikhat Ahmed, Arjen Dondorp, Shehla Farooq, Madiha Hashmi, Monaza Khan, Adnan Masood, Saima Saleem, Zahyd Shuja, Marianne Vidler, Shahnaz Zaman, Srinivas Murthy
Wellcome Open Research, 31 March 2025
Abstract
Background
Informed consent processes in critical care trials can be overwhelming for patients and their families as decisions about participation need to be made under distressing circumstances. Existing consent processes have been developed for and by White Western populations, making them less relevant for racialized groups from low- and middle-income countries. One solution is involving patients, their relatives and members of the public from diverse backgrounds in co-designing informed consent processes.
Methods
This project aims to co-design the consent materials and processes for two ongoing critical care trials with an existing Patient and Public Involvement and Engagement group at Ziauddin University in Pakistan. Using experience-based co-design and participatory action research, the process will follow six stages for each trial: (i) Introduction to trial and consent, (ii) Review of existing materials, (iii) Process mapping & recommendations, (iv) Creation of new materials, (v) Vetting, and (vi) Finalization. The final two steps involve collaboration with clinicians, trial teams, sponsors, and members of ethics review boards.
Expected Outcomes
The co-design process will produce new consent materials, aligned with patient and substitute decision-maker needs, that can be implemented to improve the conduct of two active trials. Deliverables include an informed consent process map, recommendations for revising consent materials, and guidelines on patient-centred formats. Additionally, training patient and public members in co-design will build capacity and enable the group to contribute to the design of consent processes of future trials.
Conclusion
Collaboration between patients, researchers and the public has the potential to promote ethical conduct of critical care trials in Pakistan and elsewhere by supporting patient-centered informed decision-making. This co-design process represents the first step towards achieving this goal, with future work evaluating the impact of the new consent materials and processes on trial enrolment rates, participant diversity and consent experiences.

Data sharing in child and adolescent psychiatry research: Key challenges (and some potential solutions)
Case Study
Beth Oakley, Alexandra Lautarescu, Tony Charman, Christopher Chatham, Eva Loth, Christian Beckmann, Thomas Bourgeron, Florence Campana, Rosie Holt, Eliza Eaton, Pierre Violland, Katrien Van den Bosch, Siofra Heraty, Scott Wagers, Jan Buitelaar, Declan Murphy, Amy Goodwin, Emily Jones
Open Research Europe, 31 March 2025
Open Access
Abstract
Background
The field of biomedical research is entering a new era, in which public data sharing is increasingly the norm. There are many advantages of embracing data sharing initiatives, including tackling the replication crisis through enhanced transparency and publication of null findings, facilitating global collaborations to accelerate research progress, enhancing cost-effectiveness by reducing duplication of efforts, and making scientific advances more accessible to the public. However, there are also several crucial ethical and logistical challenges that must be addressed to maximise the benefits of data sharing and minimise risks. The potential, and increasingly recognised, risks of unregulated data sharing (e.g., data reidentification, misuse, and lack of representativeness due to variability in who agrees to share data) have also been exemplified by high profile data breaches and directly clash with efforts to make research more robust, accessible, and global.
Methods/Results
Here, we narratively outline current challenges for data sharing from the perspective of child and adolescent psychiatry, one area where they may be particularly acute. For example, child and early adolescent research often requires caregivers to consent on behalf of a minor – increasing the responsibility of researchers to consider how the science of today may evolve into the future (when those individuals are no longer minors). We use data from our research consortium Autism Innovative Medicines Study – 2 – Trials (AIMS-2-TRIALS; https://www.aims-2-trials.eu/) to illustrate the points raised in this perspective piece.
Conclusions
We also propose some potential solutions to begin to address current challenges for data sharing, focusing on key priorities, including shared control of data curation between researcher and participant communities and equity of access by research groups to the tools and resources needed to conduct responsible and sustainable data sharing.

Improving shared decision‑making between paediatric haematologists, children with sickle cell disease and their parents: an observational post-intervention study
Ricardo Orlando Wijngaarde, Samantha C. Gouw, Dirk T. Ubbink
Research Square, 31 March 2025
Abstract
Background
Children with sickle cell disease (SCD) suffer from a chronic disease that can lead to serious co-morbidity and impacts their quality of life. During the course of their disease, a variety of health-related decisions need to be made for and by SCD-patients, depending on their age and health status, together with their parents and paediatric haematology clinicians. Shared decision-making (SDM) may improve health outcomes of chronically ill children but is still not commonly applied. We assessed the level of SDM among paediatric haematologists after the introduction of SDM-interventions.
Methods
An observational post-intervention study was conducted in a paediatric outpatient clinic of a university hospital. After an SDM consultation training of the three paediatric haematologists and introduction of SDM-supporting tools for both paediatricians and (parents of) patients with SCD, two evaluators independently and objectively analysed the level of patient involvement in decision-making from audio-recordings of the consultations using the OPTION-5 instrument. SDM-Q-9 and SDM-Q-Doc questionnaires were used to measure the level of SDM as perceived by patients/parents and paediatricians, respectively. Scores were expressed as a percentage, ranging from 0% (no SDM observed) to 100% (exemplary level of SDM).
Results
Participants were 9 female and 9 male patients between 3 months and 17 years old, with a mean age of 7.6 years (SD 5.5). Eighteen consultations (six per paediatrician) in which a decision was to be made about SCD treatment options were analysed. Median OPTION-5 score was 50 (Interquartile Range [IQR] 40–65%). Median SDM-Q-9 and SDM-Q-Doc scores were 73% (IQR 52.2–91) and 62.2% (IQR 55.6–71.1), respectively.
Conclusion
After the introduction of SDM training and tools, paediatric haematologists reached a moderately good level of SDM. This level had doubled as compared to the baseline level, as assessed in a previous study.