For artificial intelligence (AI) to help improve mental healthcare, the design of data-driven technologies needs to be fair, safe, and inclusive. Participatory design can play a critical role in empowering marginalised communities to take an active role in constructing research agendas and outputs. Given the unmet needs of the LGBTQI+ (Lesbian, Gay, Bisexual, Transgender, Queer and Intersex) community in mental healthcare, there is a pressing need for participatory research to include a range of diverse queer perspectives on issues of data collection and use (in routine clinical care as well as for research) as well as AI design. Here we propose a protocol for a Delphi consensus process for the development of PARticipatory Queer AI Research for Mental Health (PARQAIR-MH) practices, aimed at informing digital health practices and policy.

The development of PARQAIR-MH is comprised of four stages. In stage 1, a review of recent literature and fact-finding consultation with stakeholder organisations will be conducted to define a terms-of-reference for stage 2, the Delphi process. Our Delphi process consists of three rounds, where the first two rounds will iterate and identify items to be included in the final Delphi survey for consensus ratings. Stage 3 consists of consensus meetings to review and aggregate the Delphi survey responses, leading to stage 4 where we will produce a reusable toolkit to facilitate participatory development of future bespoke LGBTQI+–adapted data collection, harmonisation, and use for data-driven AI applications specifically in mental healthcare settings.

PARQAIR-MH aims to deliver a toolkit that will help to ensure that the specific needs of LGBTQI+ communities are accounted for in mental health applications of data-driven technologies. The study is expected to run from June 2024 through January 2025, with the final outputs delivered in mid-2025. Participants in the Delphi process will be recruited by snowball and opportunistic sampling via professional networks and social media (but not by direct approach to healthcare service users, patients, specific clinical services, or via clinicians’ caseloads). Participants will not be required to share personal narratives and experiences of healthcare or treatment for any condition. Before agreeing to participate, people will be given information about the issues considered to be in-scope for the Delphi (eg, developing best practices and methods for collecting and harmonising sensitive characteristics data; developing guidelines for data use/reuse) alongside specific risks of unintended harm from participating that can be reasonably anticipated. Outputs will be made available in open-access peer-reviewed publications, blogs, social media, and on a dedicated project website for future reuse.

To improve healthcare provider knowledge of Tanzanian newborn care guidelines, we developed adaptive Essential and Sick Newborn Care (aESNC), an adaptive e-learning environment. The objectives of this study were to (1) assess implementation success with use of in-person support and nudging strategy and (2) describe baseline provider knowledge and metacognition.

6-month observational study at one zonal hospital and three health centres in Mwanza, Tanzania. To assess implementation success, we used the Reach, Efficacy, Adoption, Implementation and Maintenance framework and to describe baseline provider knowledge and metacognition we used Howell’s conscious-competence model. Additionally, we explored provider characteristics associated with initial learning completion or persistent activity.

aESNC reached 85% (195/231) of providers: 75 medical, 53 nursing and 21 clinical officers; 110 (56%) were at the zonal hospital and 85 (44%) at health centres. Median clinical experience was 4 years (IQR 1–9) and 45 (23%) had previous in-service training for both newborn essential and sick newborn care. Efficacy was 42% (SD ±17%). Providers averaged 78% (SD ±31%) completion of initial learning and 7% (SD ±11%) of refresher assignments. 130 (67%) providers had ≥1 episode of inactivity >30 day, no episodes were due to lack of internet access. Baseline conscious-competence was 53% (IQR: 38%–63%), unconscious-incompetence 32% (IQR: 23%–42%), conscious-incompetence 7% (IQR: 2%–15%), and unconscious-competence 2% (IQR: 0%–3%). Higher baseline conscious-competence (OR 31.6 (95% CI 5.8 to 183.5)) and being a nursing officer (aOR: 5.6 (95% CI 1.8 to 18.1)), compared with medical officer, were associated with initial learning completion or persistent activity.

aESNC reach was high in a population of frontline providers across diverse levels of care in Tanzania. Use of in-person support and nudging increased reach, initial learning and refresher assignment completion, but refresher assignment completion remains low. Providers were often unaware of knowledge gaps, and lower baseline knowledge may decrease initial learning completion or activity. Further study to identify barriers to adaptive e-learning normalisation is needed.