Search

Abstract Background It has been argued that ethics review committees—e.g., Research Ethics Committees, Institutional Review Boards, etc.— have weaknesses in reviewing big data and artificial intelligence research. For instance, they may, due to the novelty of the area, lack the relevant expertise for judging collective risks and benefits of such research, or they may exempt it from review in instances involving de-identified data. Main body Focusing on the example of medical research databases we highlight here ethical issues around de-identified data sharing which motivate the need for review where oversight by ethics committees is weak. Though some argue for ethics committee reform to overcome these weaknesses, it is unclear whether or when that will happen. Hence, we argue that ethical review can be done by data access committees, since they have de facto purview of big data and artificial intelligence projects, relevant technical expertise and governance knowledge, and already take on some functions of ethical review. That said, like ethics committees, they may have functional weaknesses in their review capabilities. To strengthen that function, data access committees must think clearly about the kinds of ethical expertise, both professional and lay, that they draw upon to support their work. Conclusion Data access committees can undertake ethical review of medical research databases provided they enhance that review function through professional and lay ethical expertise.

Abstract Artificial intelligence (AI) is often cited as a possible solution to current issues faced by healthcare systems. This includes the freeing up of time for doctors and facilitating person-centred doctor-patient relationships. However, given the novelty of artificial intelligence tools, there is very little concrete evidence on their impact on the doctor-patient relationship or on how to ensure that they are implemented in a way which is beneficial for person-centred care. Given the importance of empathy and compassion in the practice of person-centred care, we conducted a literature review to explore how AI impacts these two values. Besides empathy and compassion, shared decision-making, and trust relationships emerged as key values in the reviewed papers. We identified two concrete ways which can help ensure that the use of AI tools have a positive impact on person-centred doctor-patient relationships. These are (1) using AI tools in an assistive role and (2) adapting medical education. The study suggests that we need to take intentional steps in order to ensure that the deployment of AI tools in healthcare has a positive impact on person-centred doctor-patient relationships. We argue that the proposed solutions are contingent upon clarifying the values underlying future healthcare systems.

Many legal, ethical, and regulatory frameworks allow de-identified health data to be shared for research without patients’ opt-in consent. However, there may be public concerns about this practice, as people may feel they should have some say in how such data is used. This paper introduces the concept of the “inalienability of de-identified data,” to describe a key assumption underlying that public concern and preference. The assumption, derived from ethnographic research with public and professional stakeholders in AI driven medical image analysis over the past two years, refers to a sense of felt ownership over de-identified health data, even where the subject has been obscured as referent and no clear legal rights of data ownership otherwise exist. The concept is important to medical ethics because it underpins public expectations regarding the rights people should have over the sharing of medical data (including expectations for consent). We note that where those expectations go counter to current legal and bioethical frameworks for de-identified data sharing, they provide a challenge for public support of big data and artificial intelligence driven health research.

Abstract Background As the use of AI becomes more pervasive, and computerised systems are used in clinical decision-making, the role of trust in, and the trustworthiness of, AI tools will need to be addressed. Using the case of computational phenotyping to support the diagnosis of rare disease in dysmorphology, this paper explores under what conditions we could place trust in medical AI tools, which employ machine learning. Methods Semi-structured qualitative interviews (n = 20) with stakeholders (clinical geneticists, data scientists, bioinformaticians, industry and patient support group spokespersons) who design and/or work with computational phenotyping (CP) systems. The method of constant comparison was used to analyse the interview data. Results Interviewees emphasized the importance of establishing trust in the use of CP technology in identifying rare diseases. Trust was formulated in two interrelated ways in these data. First, interviewees talked about the importance of using CP tools within the context of a trust relationship; arguing that patients will need to trust clinicians who use AI tools and that clinicians will need to trust AI developers, if they are to adopt this technology. Second, they described a need to establish trust in the technology itself, or in the knowledge it provides—epistemic trust. Interviewees suggested CP tools used for the diagnosis of rare diseases might be perceived as more trustworthy if the user is able to vouchsafe for the technology’s reliability and accuracy and the person using/developing them is trusted. Conclusion This study suggests we need to take deliberate and meticulous steps to design reliable or confidence-worthy AI systems for use in healthcare. In addition, we need to devise reliable or confidence-worthy processes that would give rise to reliable systems; these could take the form of RCTs and/or systems of accountability transparency and responsibility that would signify the epistemic trustworthiness of these tools. words 294.

Computational phenotyping (CP) technology uses facial recognition algorithms to classify and potentially diagnose rare genetic disorders on the basis of digitised facial images. This AI technology has a number of research as well as clinical applications, such as supporting diagnostic decision-making. Using the example of CP, we examine stakeholders’ views of the benefits and costs of using AI as a diagnostic tool within the clinic. Through a series of in-depth interviews (n ​= ​20) with: clinicians, clinical researchers, data scientists, industry and support group representatives, we report stakeholder views regarding the adoption of this technology in a clinical setting. While most interviewees were supportive of employing CP as a diagnostic tool in some capacity we observed ambivalence around the potential for artificial intelligence to overcome diagnostic uncertainty in a clinical context. Thus, while there was widespread agreement amongst interviewees concerning the public benefits of AI assisted diagnosis, namely, its potential to increase diagnostic yield and enable faster more objective and accurate diagnoses by up skilling non specialists and thereby enabling access to diagnosis that is potentially lacking, interviewees also raised concerns about ensuring algorithmic reliability, expunging algorithmic bias and that the use of AI could result in deskilling the specialist clinical workforce. We conclude that, prior to widespread clinical implementation, on-going reflection is needed regarding the trade-offs required to determine acceptable levels of bias and conclude that diagnostic AI tools should only be employed as an assistive technology within the dysmorphology clinic.

Abstract Digital pathology – the digitalisation of clinical histopathology services through the scanning and storage of pathology slides – has opened up new possibilities for health care in recent years, particularly in the opportunities it brings for artificial intelligence (AI)‐driven research. Recognising, however, that there is little scholarly debate on the ethics of digital pathology when used for AI research, this paper summarises what it sees as four key ethical issues to consider when deploying AI infrastructures in pathology, namely, privacy, choice, equity, and trust. The themes are inspired from the authors' experience grappling with the challenge of deploying an ethical digital pathology infrastructure to support AI research as part of the National Pathology Imaging Cooperative (NPIC), a collaborative of universities, hospital trusts, and industry partners largely located across the North of England. Though focusing on the UK case, internationally, few pathology departments have gone fully digital, and so the themes developed here offer a heuristic for ethical reflection for other departments currently making a similar transition or planning to do so in the future. We conclude by promoting the need for robust public governance mechanisms in AI‐driven digital pathology.

Plain English summary How might the public be authentically involved in decisions about medical data sharing for artificial intelligence (AI) research? In this paper, we highlight three ways in which public views are used to improve such decisions, namely, through lay representation on data access committees, through patient and public involvement groups, and through a variety of public engagement events we call “citizen forums.” Though each approach has common strengths and weaknesses, we argue that they are unable to support each other due to a lack of proper integration. We therefore propose combining them so that they work in a more coordinated way. The combined model, we argue, could be useful for improving the trustworthiness of big data and AI related medical research in the future.

Abstract Digital Pathology (DP) is a platform which has the potential to develop a truly integrated and global pathology community. The generation of DP data at scale creates novel challenges for the histopathology community in managing, processing, and governing the use of these data. The current understanding of, and confidence in, the legal and ethical aspects of DP by pathologists is unknown. We developed an electronic survey (e‐survey), comprising 22 questions, with input from the Royal College of Pathologists (RCPath) Digital Pathology Working Group. The e‐survey was circulated via e‐mail and social media (Twitter) through the RCPath Digital Pathology Working Group network, RCPath Trainee Committee network, the Pathology image data Lake for Analytics, Knowledge and Education (PathLAKE) digital pathology consortium, National Pathology Imaging Co‐operative (NPIC), local contacts, and to the membership of both The Pathological Society of Great Britain and Ireland and the British Division of the International Academy of Pathology (BDIAP). Between 14 July 2020 and 6 September 2020, we collected 198 responses representing a cross section of histopathologists, including individuals with experience of DP research. We ascertained that, in the UK, DP is being used for diagnosis, research, and teaching, and that the platform is enabling data sharing. Our survey demonstrated that there is often a lack of confidence and understanding of the key issues of consent, legislation, and ethical guidelines. Of 198 respondents, 82 (41%) did not know when the use of digital scanned slide images would fall under the relevant legislation and 93 (47%) were ‘Not confident at all’ in their interpretation of consent for scanned slide images in research. With increasing uptake of DP, a working knowledge of these areas is essential but histopathologists often express a lack of confidence in these topics. The need for specific training in these areas is highlighted by the findings of this study.

Abstract Background Equipoise and role conflict have been previously identified as important factors in professionals’ engagement with trials, inducing behaviours which can impact on recruitment. We explored these phenomena as potential explanations for the low levels of involvement of teenagers and young adults (TYA) with cancer in clinical trials in oncology. Methods We report findings from interviews with 30 purposively sampled direct-care professionals involved in delivering cancer care and/or facilitating clinical trials in Scotland. We undertook qualitative descriptive analysis, focussed on identifying key issues and themes. Results Interviewees largely identified as clinician–researchers and portrayed oncology as a specialty in which research was integral to care. They saw their primary responsibility as ensuring patients received the best treatment, but asserted that, in general, trials provided a vehicle for optimal care. Role conflict in its traditional form was rarely evident; however, other tensions were manifest. Professionals found the significant time costs of delivering trials difficult to reconcile with the increasing pressures on clinical services. They felt a responsibility to make prudent choices about the trials with which to engage. Guided by utilitarian principles, these choices were oriented towards benefiting the largest number of patients. This favoured trials in high volume diseases; as TYA tend to have rarer forms of cancer, professionals’ support for—and TYA’s access to—relevant trials was, by default, more limited. Conclusions Neither lack of individual equipoise nor experiences of traditional forms of role conflict accounted for the low levels of involvement of TYA with cancer in clinical trials. However, prominent tensions around the management of scarce resources provided an alternative explanation for TYA’s limited access to cancer trials. The prevailing approach to decision-making about whether and which trials to support was recognised as contributing to inequalities in access and care. Professionals’ choices, however, were made in the context of scarcity, and structured by incentives and sanctions understood by them as signalling governmental priorities. A franker discussion of the extent and distribution of the costs and benefits of trials work is needed, for change to be achieved.

This paper is based upon interviews with twenty-seven women and men who have an inherited risk of developing gastric cancer and have had their stomach removed as a preventative measure. We describe what happens when bodily processes – digestion - are disrupted by the removal of the stomach. Interviewees' who had undergone prophylactic total gastrectomy experienced changes to the lived experience of hunger and appetite. The interviewees' accounts of life post-surgery suggested that private sensations of hunger (i.e. internalised feelings of hunger) disappeared following gastrectomy and in most cases never reappeared. The majority of interviewees described an alternative sense of hunger as developing over time. This externalised or disembodied hunger was described as triggered by a range of extracorporeal or external bodily criteria rather than internal sensations. We argue that this externally motivated desire for food, generated from without rather than within, serves the same purpose as internally generated feelings of hunger - it encourages eating, which sustains the body. Interviewees reported not only having to learn when to consume food following PTG, but also what to eat by adapting to what they perceived as their body's food tolerances and changing tastes. We argue that these changes in the lived experience of hunger and appetite may affect individuals' ability to participate in commensal relationships following gastrectomy, and that this may have negative consequences for identity.

Background: Around 60,000 babies are born preterm (prior to 37 weeks’ gestation) each year in the UK. There is little evidence on the optimal birth mode (vaginal or caesarean section). Objective: The overall aim of the CASSAVA project was to determine if a trial to define the optimal mode of preterm birth could be carried out and, if so, determine what sort of trial could be conducted and how it could best be performed. We aimed to determine the specific groups of preterm women and babies for whom there are uncertainties about the best planned mode of birth, and if there would be willingness to recruit to, and participate in, a randomised trial to address some, but not all, of these uncertainties. This project was conducted in response to a Heath Technology Assessment programme commissioning call (17/22 ‘Mode of delivery for preterm infants’). Methods: We conducted clinician and patient surveys (n = 224 and n = 379, respectively) to identify current practice and opinion, and a consensus survey and Delphi workshop (n = 76 and n = 22 participants, respectively) to inform the design of a hypothetical clinical trial. The protocol for this clinical trial/vignette was used in telephone interviews with clinicians (n = 24) and in focus groups with potential participants (n = 13). Results: Planned sample size and data saturation was achieved for all groups except for focus groups with participants, as this had to be curtailed because of the COVID-19 pandemic and data saturation was not achieved. There was broad agreement from parents and health-care professionals that a trial is needed. The clinician survey demonstrated a variety of practice and opinion. The parent survey suggested that women and their families generally preferred vaginal birth at later gestations and caesarean section for preterm infants. The interactive workshop and Delphi consensus process confirmed the need for more evidence (hence the case for a trial) and provided rich information on what a future trial should entail. It was agreed that any trial should address the areas with most uncertainty, including the management of women at 26–32 weeks’ gestation, with either spontaneous preterm labour (cephalic presentation) or where preterm birth was medically indicated. Clear themes around the challenges inherent in conducting any trial emerged, including the concept of equipoise itself. Specific issues were as follows: different clinicians and participants would be in equipoise for each clinical scenario, effective conduct of the trial would require appropriate resources and expertise within the hospital conducting the trial, potential participants would welcome information on the trial well before the onset of labour and minority ethnic groups would require tailored approaches. Conclusion: Given the lack of evidence and the variation of practice and opinion in this area, and having listened to clinicians and potential participants, we conclude that a trial should be conducted and the outlined challenges resolved. Future work: The CASSAVA project could be used to inform the design of a randomised trial and indicates how such a trial could be carried out. Any future trial would benefit from a pilot with qualitative input and a study within a trial to inform optimal recruitment. Limitations: Certainty that a trial could be conducted can be determined only when it is attempted. Trial registration: Current Controlled Trials ISRCTN12295730. Funding: This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 25, No. 61. See the NIHR Journals Library website for further project information.

The concept of ‘digital phenotyping’ was originally developed by researchers in the mental health field, but it has travelled to other disciplines and areas. This commentary draws upon our experiences of working in two scientific projects that are based at the University of Oxford’s Big Data Institute – The RADAR-AD project and The Minerva Initiative – which are developing algorithmic phenotyping technologies. We describe and analyse the concepts of digital biomarkers and computational phenotyping that underlie these projects, explain how they are linked to other research in digital phenotyping and compare and contrast some of their epistemological and ethical implications. In particular, we argue that the phenotyping paradigm in both projects is grounded on an assumption of ‘objectivity’ that is articulated in different ways depending on the role that is given to the computational/digital tools. Using the concept of ‘affordance’, we show how specific functionalities relate to potential uses and social implications of these technologies and argue that it is important to distinguish among them as the concept of digital phenotyping is increasingly being used with a variety of meanings.

Abstract Background While a growing body of research has explored why people take part in clinical trials, this research has not considered how people’s understandings, motivations and agendas might influence their conduct during a trial. This is an important area of enquiry because it is now widely recognised that an intervention might lead to different clinical outcomes when delivered as part of a trial than when implemented in routine clinical practice; however, the reasons for this are not fully understood. Methods/design We interviewed 24 individuals who took part in a trial of an innovative health technology under development for people with type 1 diabetes which automatically regulates blood glucose: the closed-loop system. Participants were interviewed following randomisation to a closed-loop and at trial closeout. Results Participants provided complex agendas for taking part in which altruistic and self-interested considerations were often inseparable. Many described belonging to a wider diabetes community and being beneficiaries of others’ participation in research and how this had given rise to attendant citizenship obligations. Participants also shared the excitement and pride they experienced from contributing to research which situated them at the forefront of technological innovation and enabled them to present themselves to others, by virtue of their trial participation, as ambassadors of hope and research pioneers. Given their desire to support the progression of a potentially life-changing technology, and be part of that innovation, participants, at follow-up, described having made extra effort during the trial. Specifically, participants described having been more focused on their diabetes management to help create conditions in which the closed-loop could work most effectively to optimize their blood glucose control. Conclusions Our findings contribute a new dimension to understandings of trial effects; specifically, we argue that, to aid interpretation of trial outcomes, participants’ understandings and motivations for participation need to be considered. We highlight the potential pertinence of our findings in the contemporary era of bio-citizenship where, increasingly, people are driving research agendas and see themselves as co-producers of knowledge. We also recommend a new concept be introduced into the literature—‘the altruselfish agenda’—to recognise potential inseparability of self-interested and altruistic motivations. Trial registration ClinicalTrials.gov, NCT02523131. Registered on 14 August 2015.

BackgroundRetained placenta following vaginal delivery is a major cause of postpartum haemorrhage. Currently, the only effective treatments for a retained placenta are the surgical procedures of manual removal of placenta (MROP) and uterine curettage, which are not universally available, particularly in low- and middle-income countries. The objective of the trial was to determine whether sublingual nitroglycerin spray was clinically effective and cost-effective for medical treatment of retained placenta following vaginal delivery.Methods and findingsA randomised, placebo-controlled, double-blind trial was undertaken between October 2014 and July 2017 at 29 delivery units in the UK (Edinburgh, Glasgow, Manchester, Newcastle, Preston, Warrington, Chesterfield, Crewe, Durham, West Middlesex, Aylesbury, Furness, Southampton, Bolton, Sunderland, Oxford, Nottingham [2 units], Burnley, Chertsey, Stockton-on-Tees, Middlesborough, Chester, Darlington, York, Reading, Milton Keynes, Telford, Frimley). In total, 1,107 women with retained placenta following vaginal delivery were recruited. The intervention was self-administered 2 puffs of sublingual nitroglycerin (800 μg; intervention, N = 543) or placebo spray (control, N = 564). The primary clinical outcome was the need for MROP, assessed at 15 minutes following administration of the intervention. Analysis was based on the intention-to-treat principle. The primary safety outcome was measured blood loss between study drug administration and transfer to the postnatal ward or other clinical area. The primary patient-sided outcomes were satisfaction with treatment and side-effect profile, assessed by questionnaires pre-discharge and 6 weeks post-delivery. Secondary clinical outcomes were measured at 5 and 15 minutes after study drug administration and prior to hospital discharge. There was no statistically significant or clinically meaningful difference in need for MROP by 15 minutes (primary clinical outcome, 505 [93.3%] for nitroglycerin versus 518 [92.0%] for placebo, odds ratio [OR] 1.01 [95% CI 0.98-1.04], p = 0.393) or blood loss (1,000 ml: nitroglycerin, 119 [22.2%], versus placebo, 87 [15.5%]; ordinal OR 1.14 [95% CI 0.88-1.48], p = 0.314) or satisfaction with treatment (nitroglycerin, 288 [75.4%], versus placebo, 303 [78.1%]; OR 0.87 [95% CI 0.62-1.22], p = 0.411) or health service costs (mean difference [£] 55.3 [95% CI -199.20 to 309.79]). Palpitations following drug administration were reported more often in the nitroglycerin group (36 [9.8%] versus 15 [4.0%], OR 2.60 [95% CI 1.40-4.84], p = 0.003). There were 52 serious adverse events during the trial, with no statistically significant difference in likelihood between groups (nitroglycerin, 27 [5.0%], versus placebo, 26 [4.6%]; OR 1.13 [95% CI 0.54-2.38], p = 0.747). The main limitation of our study was the low return rate for the 6-week postnatal questionnaire. There were, however, no differences in questionnaire return rates between study groups or between women who did and did not have MROP, with the patient-reported use of outpatient and primary care services at 6 weeks accounting for only a small proportion (approximately 5%) of overall health service costs.ConclusionsIn this study, we found that nitroglycerin is neither clinically effective nor cost-effective as a medical treatment for retained placenta, and has increased side effects, suggesting it should not be used. Further research is required to identify an effective medical treatment for retained placenta to reduce the morbidity caused by this condition, particularly in low- and middle-income countries where surgical management is not available.Trial registrationISRCTN.com ISRCTN88609453 ClinicalTrials.gov NCT02085213.

Background: Retained placenta is associated with postpartum haemorrhage and can lead to significant maternal morbidity if untreated. The only effective treatment is the surgical procedure of manual removal of placenta, which is costly, requires skilled staff, requires an operative environment and is unpleasant for women. Small studies suggest that glyceryl trinitrate may be an effective medical alternative. Objective: To determine the clinical effectiveness and cost-effectiveness of sublingual glyceryl trinitrate spray compared with placebo in reducing the need for manual removal of placenta in women with retained placenta after vaginal delivery following the failure of current management. Design: A group-sequential randomised double-blind placebo-controlled trial with a cost-effectiveness analysis. Setting: There were 29 obstetric units in the UK involved in the study. Participants: There were 1107 women (glyceryl trinitrate group, n = 543; placebo group, n = 564) randomised between October 2014 and July 2017. Interventions: Glyceryl trinitrate spray was administered to 541 women in the intervention group, and a placebo was administered to 563 women in the control group. Main outcome measures: Four primary outcomes were defined: (1) clinical – the need for manual removal of placenta, (2) safety – measured blood loss, (3) patient sided – satisfaction with treatment and side effects and (4) economic – cost-effectiveness of both treatments using the UK NHS perspective. Secondary clinical outcomes included a > 15% decrease in haemoglobin level, time from randomisation to delivery of placenta in theatre, the need for earlier manual removal of placenta than planned, increase in heart rate or decrease in blood pressure, requirement for blood transfusion, requirement for general anaesthesia, maternal pyrexia, and sustained uterine relaxation requiring additional uterotonics. Results: No difference was observed between the glyceryl trinitrate group and the control group for the placenta remaining undelivered within 15 minutes of study treatment (93.3% vs. 92%; odds ratio 1.01, 95% confidence interval 0.98 to 1.04; p = 0.393). There was no difference in blood loss of > 1000 ml between the glyceryl trinitrate group and the control group (22.2% vs. 15.5%; odds ratio 1.14, 95% confidence interval 0.88 to 1.48; p = 0.314). Palpitations were more common in the glyceryl trinitrate group than in the control group after taking the study drug (9.8% vs. 4.0%; odds ratio 2.60, 95% confidence interval 1.40 to 4.84; p = 0.003). There was no difference in any other measures of patient satisfaction between the groups. There was no difference in costs to the health service between groups (mean difference £55.30, 95% confidence interval –£199.20 to £309.79). Secondary outcomes revealed that a fall in systolic or diastolic blood pressure, or an increase in heart rate, was more common in the glyceryl trinitrate group than in the control group (odds ratio 4.9, 95% confidence interval 3.7 to 6.4; p

Abstract Background Genomic and biosocial research data about individuals is rapidly proliferating, bringing the potential for novel opportunities for data integration and use. The scale, pace and novelty of these applications raise a number of urgent sociotechnical, ethical and legal questions, including optimal methods of data storage, management and access. Although the open science movement advocates unfettered access to research data, many of the UK’s longitudinal cohort studies operate systems of managed data access, in which access is governed by legal and ethical agreements between stewards of research datasets and researchers wishing to make use of them. Amongst other things, these agreements aim to respect the reasonable expectations of the research participants who provided data and samples, as expressed in the consent process. Arguably, responsible data management and governance of data and sample use are foundational to the consent process in longitudinal studies and are an important source of trustworthiness in the eyes of those who contribute data to genomic and biosocial research. Methods This paper presents an ethnographic case study exploring the foundational principles of a governance infrastructure for Managing Ethico-social, Technical and Administrative issues in Data ACcess (METADAC), which are operationalised through a committee known as the METADAC Access Committee. METADAC governs access to phenotype, genotype and ‘omic’ data and samples from five UK longitudinal studies. Findings Using the example of METADAC, we argue that three key structural features are foundational for practising responsible data sharing: independence and transparency; interdisciplinarity; and participant-centric decision-making. We observe that the international research community is proactively working towards optimising the use of research data, integrating/linking these data with routine data generated by health and social care services and other administrative data services to improve the analysis, interpretation and utility of these data. The governance of these new complex data assemblages will require a range of expertise from across a number of domains and disciplines, including that of study participants. Human-mediated decision-making bodies will be central to ensuring achievable, reasoned and responsible decisions about the use of these data; the METADAC model described in this paper provides an example of how this could be realised.

The clinical utility of computational phenotyping for both genetic and rare diseases is increasingly appreciated; however, its true potential is yet to be fully realized. Alongside the growing clinical and research availability of sequencing technologies, precise deep and scalable phenotyping is required to serve unmet need in genetic and rare diseases. To improve the lives of individuals affected with rare diseases through deep phenotyping, global big data interrogation is necessary to aid our understanding of disease biology, assist diagnosis, and develop targeted treatment strategies. This includes the application of cutting-edge machine learning methods to image data. As with most digital tools employed in health care, there are ethical and data governance challenges associated with using identifiable personal image data. There are also risks with failing to deliver on the patient benefits of these new technologies, the biggest of which is posed by data siloing. The Minerva Initiative has been designed to enable the public good of deep phenotyping while mitigating these ethical risks. Its open structure, enabling collaboration and data sharing between individuals, clinicians, researchers and private enterprise, is key for delivering precision public health.

Background: Insulin is generally administered to people with type 1 diabetes mellitus (T1DM) using multiple daily injections (MDIs), but can also be delivered using infusion pumps. In the UK, pumps are recommended for patients with the greatest need and adult use is less than in comparable countries. Previous trials have been small, of short duration and have failed to control for training in insulin adjustment. Objective: To assess the clinical effectiveness and cost-effectiveness of pump therapy compared with MDI for adults with T1DM, with both groups receiving equivalent structured training in flexible insulin therapy. Design: Pragmatic, multicentre, open-label, parallel-group cluster randomised controlled trial, including economic and psychosocial evaluations. After participants were assigned a group training course, courses were randomly allocated in pairs to either pump or MDI. Setting: Eight secondary care diabetes centres in the UK. Participants: Adults with T1DM for > 12 months, willing to undertake intensive insulin therapy, with no preference for pump or MDI, or a clinical indication for pumps. Interventions: Pump or MDI structured training in flexible insulin therapy, followed up for 2 years. MDI participants used insulin analogues. Pump participants used a Medtronic Paradigm® VeoTM (Medtronic, Watford, UK) with insulin aspart (NovoRapid, Novo Nordisk, Gatwick, UK). Main outcome measures: Primary outcome – change in glycated haemoglobin (HbA1c) at 2 years in participants whose baseline HbA1c was ≥ 7.5% (58 mmol/mol). Key secondary outcome – proportion of participants with HbA1c ≤ 7.5% at 2 years. Other outcomes at 6, 12 and 24 months – moderate and severe hypoglycaemia; insulin dose; body weight; proteinuria; diabetic ketoacidosis; quality of life (QoL); fear of hypoglycaemia; treatment satisfaction; emotional well-being; qualitative interviews with participants and staff (2 weeks), and participants (6 months); and ICERs in trial and modelled estimates of cost-effectiveness. Results: We randomised 46 courses comprising 317 participants: 267 attended a Dose Adjustment For Normal Eating course (132 pump; 135 MDI); 260 were included in the intention-to-treat analysis, of which 235 (119 pump; 116 MDI) had baseline HbA1c of ≥ 7.5%. HbA1c and severe hypoglycaemia improved in both groups. The drop in HbA1c% at 2 years was 0.85 on pump and 0.42 on MDI. The mean difference (MD) in HbA1c change at 2 years, at which the baseline HbA1c was ≥ 7.5%, was –0.24% [95% confidence interval (CI) –0.53% to 0.05%] in favour of the pump (p = 0.098). The per-protocol analysis showed a MD in change of –0.36% (95% CI –0.64% to –0.07%) favouring pumps (p = 0.015). Pumps were not cost-effective in the base case and all of the sensitivity analyses. The pump group had greater improvement in diabetes-specific QoL diet restrictions, daily hassle plus treatment satisfaction, statistically significant at 12 and 24 months and supported by qualitative interviews. Limitation: Blinding of pump therapy was not possible, although an objective primary outcome was used. Conclusion: Adding pump therapy to structured training in flexible insulin therapy did not significantly enhance glycaemic control or psychosocial outcomes in adults with T1DM. Research priority: To understand why few patients achieve a HbA1c of

BACKGROUND: A newly introduced cell-free fetal DNA sequencing based non-invasive prenatal testing (DNA-NIPT) detects Down syndrome with sensitivity of 99% at early gestational stage without risk of miscarriage. Attention has been given to its public health implications; little is known from consumer perspectives. This qualitative study aimed to explore women's motivations for using, and perceptions of, DNA-NIPT in Hong Kong. METHODS AND FINDINGS: In-depth interviews were conducted with 45 women who had undertaken DNA-NIPT recruited by purposive sampling based on socio-demographic and clinical characteristics. The sample included 31 women identified as high-risk from serum and ultrasound based Down syndrome screening (SU-DSS). Thematic narrative analysis examined informed-decision making of the test and identified the benefits and needs. Women outlined a number of reasons for accessing DNA-NIPT: reducing the uncertainty associated with risk probability-based results from SU-DSS, undertaking DNA-NIPT as a comprehensive measure to counteract risk from childbearing especially at advanced age, perceived predictive accuracy and absence of risk of harm to fetus. Accounts of women deemed high-risk or not high-risk are distinctive in a number of respects. High-risk women accessed DNA-NIPT to get a clearer idea of their risk. This group perceived SU-DSS as an unnecessary and confusing procedure because of its varying, protocol-dependent detection rates. Those women not deemed high-risk, in contrast, undertook DNA-NIPT for psychological assurance and to reduce anxiety even after receiving the negative result from SU-DSS. CONCLUSIONS: DNA-NIPT was regarded positively by women who chose this method of screening over the routine, less expensive testing options. Given its perceived utility, health providers need to consider whether DNA-NIPT should be offered as part of universal routine care to women at high-risk for fetal aneuploidy. If this is the case, then further development of guidelines and quality assurance will be needed to provide a service suited to patients' needs.