Explore the vast range of titles available.

Background Public and patient involvement and engagement (PPIE) is intended to shape research priorities and improve relevance and impact. However, implementing PPIE in complex fields such as artificial intelligence (AI) and big data health research presents specific challenges. This study explores the issues and barriers to meaningful PPIE using the AI‐Multiply project as a case example. Methods AI‐Multiply is a large, interdisciplinary UK‐based research project using AI and routine health data to investigate trajectories of multiple long‐term conditions and polypharmacy. PPIE was embedded across all five work packages. We used a mixed‐methods approach, drawing on CUBE framework surveys, PPIE feedback forms and impact logs to evaluate involvement. Data were analysed thematically using a ‘follow‐a‐thread’ approach to identify key issues across sources. Results Three themes were identified: (1) differing priorities—public contributors prioritised person‐centred outcomes, while researchers focused on data‐driven healthcare metrics, often constrained by data availability; (2) movement on both sides—both researchers and contributors expressed early apprehension, but mutual trust and integration developed over time; and (3) the importance of established guidance—many issues raised echoed longstanding PPIE guidance on clarity, feedback and facilitation. Conclusion While AI and data‐specific challenges exist, many PPIE issues in this context relate to applying existing good practice in complex projects. Strong PPIE leadership, early expectation‐setting and consistent facilitation are critical for success. Findings will inform the development of practical tools to support involvement in data‐driven research. Patient or Public Contribution Public contributors with lived experience of multiple long‐term conditions contributed to the interpretation of data and co‐authored this manuscript.

Patient and Public Involvement and Engagement (PPIE) is critical in the development and application of Artificial Intelligence (AI) in healthcare research to ensure that outcomes align with patients’ and the public’s needs. However, current PPIE practices often limit involvement to reactive tasks such as reviewing documents and providing plain English summaries. Whilst important, this approach can sideline PPIE from influencing key research decisions. Consequently, PPIE interactions often fail to adequately reach and influence everyday decision makers. On AI and big data research projects, these decisions are often made by Early Career Researchers (ECRs) who play a vital role in the day-to-day research process. After realising these limitations, and to address them, the NIHR-funded AI MULTIPLY consortium introduced twice-monthly \"ECRs meet PPIE\" sessions. These sessions began in May 2024 and enabled ECRs to present and discuss work in progress and gain targeted input from PPIE members during early phases of research, such as research direction, data and variable selection. By integrating PPIE at this stage, the project aimed to improve the relevance and impact of the healthcare research but also provide ECRs with essential skills in public engagement. At time of writing, 12 sessions have been conducted. Through ethnographic observations integrated with internal surveys, the findings show how the sessions were developed, overcame challenges, and helped to embed PPIE contributors’ voices into an AI-in-healthcare project. Based on our findings we have identified 5 recommendations for other large interdisciplinary consortia to strengthen the contribution of PPIE to everyday decision-making in research. Plain English Summary Patient and Public Involvement and Engagement (PPIE) helps make healthcare research meet patient needs, but it is sometimes limited to tasks like reviewing documents or writing plain English summaries. This can keep patients away from key decisions made early in the research process. Early Career Researchers (ECRs), who are at the start of their careers, often make important decisions like choosing which data to include, but they usually have little interaction with patients or the public. This can mean that research might not look at what patients feel is important, especially when using complicated tools like artificial intelligence (AI). To try and improve this, the NIHR-funded AI MULTIPLY project introduced twice-monthly \"ECRs meet PPIE\" sessions. These sessions let ECRs share their work with PPIE members early on, getting input on shaping the research direction and choosing data. This early involvement helps our research to meet patient needs. As well as this, the ECRs gained valuable skills in public engagement, communication, and professional development. Since starting in May 2024, we have had 12 sessions, with positive feedback. By observing the sessions and asking people to fill out surveys, positives and challenges of meetings were discovered. We were able to show how ECRs report improved confidence, insights, and skills, while PPIE contributors feel empowered and valued for their role in shaping research. This approach shows how PPIE can link researchers and the public, supporting patient-centred AI innovations in healthcare.

Digital health interventions are increasingly being used as a supplement or replacement for face-to-face services as a part of predictive prevention. They may be offered to those who are at high risk of cardiovascular disease and need to improve their diet, increase physical activity, stop smoking, or reduce alcohol consumption. Despite the popularity of these interventions, there is no overall summary and comparison of the effectiveness of different modes of delivery of a digital intervention to inform policy. This review aims to summarize the effectiveness of digital interventions in improving behavioral and health outcomes related to physical activity, smoking, alcohol consumption, or diet in nonclinical adult populations and to identify the effectiveness of different modes of delivery of digital interventions. We reviewed articles published in the English language between January 1, 2009, and February 25, 2019, that presented a systematic review with a narrative synthesis or meta-analysis of any study design examining digital intervention effectiveness; data related to adults (≥18 years) in high-income countries; and data on behavioral or health outcomes related to diet, physical activity, smoking, or alcohol, alone or in any combination. Any time frame or comparator was considered eligible. We searched MEDLINE, Embase, PsycINFO, Cochrane Reviews, and gray literature. The AMSTAR-2 tool was used to assess review confidence ratings. We found 92 reviews from the academic literature (47 with meta-analyses) and 2 gray literature items (1 with a meta-analysis). Digital interventions were typically more effective than no intervention, but the effect sizes were small. Evidence on the effectiveness of digital interventions compared with face-to-face interventions was mixed. Most trials reported that intent-to-treat analysis and attrition rates were often high. Studies with long follow-up periods were scarce. However, we found that digital interventions may be effective for up to 6 months after the end of the intervention but that the effects dissipated by 12 months. There were small positive effects of digital interventions on smoking cessation and alcohol reduction; possible effectiveness in combined diet and physical activity interventions; no effectiveness for interventions targeting physical activity alone, except for when interventions were delivered by mobile phone, which had medium-sized effects; and no effectiveness observed for interventions targeting diet alone. Mobile interventions were particularly effective. Internet-based interventions were generally effective. Digital interventions have small positive effects on smoking, alcohol consumption, and in interventions that target a combination of diet and physical activity. Small effects may have been due to the low efficacy of treatment or due to nonadherence. In addition, our ability to make inferences from the literature we reviewed was limited as those interventions were heterogeneous, many reviews had critically low AMSTAR-2 ratings, analysis was typically intent-to-treat, and follow-up times were relatively short. PROSPERO International Prospective Register of Systematic Reviews CRD42019126074; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=126074.

Including public contributors in the development of artificial intelligence (AI) systems in healthcare research is growing, however, traditional methods of participation fail to engage people from minoritised groups. This work explores how we can utilise art-based methods to involve the perspectives of those not previously included in AI development. We collaborated with a East London-based organisation to involve people not previously included in research to contribute to a study on multiple long-term conditions (MLTCs) and polypharmacy. Patient and public involvement and engagement (PPIE) contributors all had lived experience of MLTCs and represented a range of different ages, genders, socio-demographic backgrounds and multilingual abilities. We ran a series of six workshops that used different visual arts methods; ceramics, collage, body mapping and AI-generated images, to create research priorities and to inform AI development. The arts-based methods served as a platform for communication which supported PPIE contributors to develop multiple research priorities, for example the impact of the lack of routine appointments on MLTCs. Through these workshops PPIE contributors also highlighted concepts that are important to consider during AI model development, such as utilising local housing data and considering bias. Visual images and art helped to facilitate different forms of communication, whilst being fun and engaging and provided a way to make abstract AI concepts more tangible whilst building AI literacy. Arts-based methods were a useful tool to make involvement in research more accessible for under-represented communities in the development of AI tools in healthcare research. There is a need for more inclusive participatory approaches as the use of AI in healthcare and research increases. Working with staff and interpreters from a local community-based charity, Social Action for Health, we invited 22 PPIE contributors from under-represented communities in Tower Hamlets who had no previous experience of PPIE research. PPIE contributors developed the research priorities for a large academic consortia and helped create a community art exhibition to highlight their artwork. Additionally, two experienced PPIE contributors from the wider AI-Multiply study assisted with the preparation of this manuscript.

This study examined sex-based differences in associations of vascular risk factors with incident cardiovascular events in the UK Biobank. Baseline participant demographic, clinical, laboratory, anthropometric, and imaging characteristics were collected. Multivariable Cox regression was used to estimate independent associations of vascular risk factors with incident myocardial infarction (MI) and ischaemic stroke for men and women. Women-to-men ratios of hazard ratios (RHRs), and related 95% confidence intervals, represent the relative effect-size magnitude by sex. Among the 363 313 participants (53.5% women), 8470 experienced MI (29.9% women) and 7705 experienced stroke (40.1% women) over 12.66 [11.93, 13.38] years of prospective follow-up. Men had greater risk factor burden and higher arterial stiffness index at baseline. Women had greater age-related decline in aortic distensibility. Older age [RHR: 1.02 (1.01-1.03)], greater deprivation [RHR: 1.02 (1.00-1.03)], hypertension [RHR: 1.14 (1.02-1.27)], and current smoking [RHR: 1.45 (1.27-1.66)] were associated with a greater excess risk of MI in women than men. Low-density lipoprotein cholesterol was associated with excess MI risk in men [RHR: 0.90 (0.84-0.95)] and apolipoprotein A (ApoA) was less protective for MI in women [RHR: 1.65 (1.01-2.71)]. Older age was associated with excess risk of stroke [RHR: 1.01 (1.00-1.02)] and ApoA was less protective for stroke in women [RHR: 2.55 (1.58-4.14)]. Older age, hypertension, and smoking appeared stronger drivers of cardiovascular disease in women, whereas lipid metrics appeared stronger risk determinants for men. These findings highlight the importance of sex-specific preventive strategies and suggest priority targets for intervention in men and women.

Electronic healthcare records (EHR) contain a huge wealth of data that can support the prediction of clinical outcomes. EHR data is often stored and analysed using clinical codes (ICD10, SNOMED), however these can differ across registries and healthcare providers. Integrating data across systems involves mapping between different clinical ontologies requiring domain expertise, and at times resulting in data loss. To overcome this, code-agnostic models have been proposed. We assess the effectiveness of a code-agnostic representation approach on the task of long-term microvascular complication prediction for individuals living with Type 2 Diabetes. Our method encodes individual EHRs as text using fine-tuned, pretrained clinical language models. Leveraging large-scale EHR data from the UK, we employ a multi-label approach to simultaneously predict the risk of microvascular complications across 1-, 5-, and 10-year windows. We demonstrate that a code-agnostic approach outperforms a code-based model and illustrate that performance is better with longer prediction windows but is biased to the first occurring complication. Overall, we highlight that context length is vitally important for model performance. This study highlights the possibility of including data from across different clinical ontologies and is a starting point for generalisable clinical models.

Developing artificial intelligence (AI) tools for healthcare is a multiple disciplinary effort, bringing data scientists, clinicians, patients and other disciplines together. In this paper, we explore the AI development workflow and how participants navigate the challenges and tensions of sharing and generating knowledge across disciplines. Through an inductive thematic analysis of 13 semi-structured interviews with participants in a large research consortia, our findings suggest that multiple disciplinarity heavily impacts work practices. Participants faced challenges to learn the languages of other disciplines and needed to adapt the tools used for sharing and communicating with their audience, particularly those from a clinical or patient perspective. Large health datasets also posed certain restrictions on work practices. We identified meetings as a key platform for facilitating exchanges between disciplines and allowing for the blending and creation of knowledge. Finally, we discuss design implications for data science and collaborative tools, and recommendations for future research.