Clinicians, patients and policy makers lack access to accurate, real time information on new treatments for treating cancer. This is because such a large amount of information is continuously generated, and it is too complicated to be manually analysed in a timely fashion. This is sometimes referred to as a health 'infodemic'. Information analysed to create clinical evidence (known as systematic reviews) quickly goes out of date, and national bodies responsible for appraising new treatments such as the National Institute for Clinical Excellence are unable to keep up. It is increasingly hard to detect misinformation published within medical literature, and an increasing number of papers have to be withdrawn after publication. INDICATE is a deep learning tool for the autonomous generation of systematic reports and analysis of both structured and unstructured data from published literature on cancer. It has been developed through a collaboration between Imperial College London and Amazon Web Services, NICE and the British Medical Journal (BMJ). The aim is to develop a methodology for the real time analysis of healthcare infodemics that can be used to autonomously create clinical guidance and identify misinformation. This project will build on previous work to develop AI methodologies that automate how we search for medical literature and it will intelligently support peer reviewers as they appraise and assess the quality of research papers. This work has three main goals: 1. To develop a tool for detecting research fraud. 2. To asses if our AI tools can speed up the creation of NICE guidance. 3. To develop autonomous summary reports of clinical evidence of breast cancer treatment that could be used by medical publishers. The study group will work with clinicians, researchers and NICE to define and prioritise critical questions that require answering and to refine the user interface for the system. Moreover, we will prospectively validate the performance of the system to determine the accuracy and performance of its reporting mechanism. The validated data generated by this study will form the basis of a phase II study that scales the number of cancer types and the trial of the technology in a real world clinical environment.

Adolescence is a formative phase of life which shapes the health and well-being of an individual for adulthood. In South Africa, adolescents are experiencing large and growing burdens of disease with high rates of communicable diseases (like HIV and TB), pregnancy, obesity, mental illness and trauma. Government policies to support adolescent healthcare have been widely implemented, however, challenges remain in how such policies are operationalised to strengthen adolescent healthcare: primary care clinics report low attendance of adolescents and school health reviews show low service provision. The Knowledge Translation Unit in Cape Town has developed the PACK Adolescent guide and training programme to support primary care clinicians (nurses and doctors) with tools to meet the needs of adolescents. The aim of this study is to understand the status of adolescent healthcare in primary care clinics and schools in order to refine and adapt PACK Adolescent to meet the needs of adolescents and the challenges faced by clinicians in providing care to this underserved grouped. Once refined, we shall pilot and evaluate PACK Adolescent in schools and clinics to see what more we need to do to optimise PACK Adolescent and what we need to alter about the delivery of adolescent healthcare to enable its use in primary care clinics and schools in other settings. From the outset, the design of this study involved the input of adolescents themselves, primary care clinicians and policymakers in health and education sectors during a stakeholder workshop. Throughout the grant's duration, stakeholder input and co-production will characterise the study through theory of change workshops and an advisory board. The study seeks to assess adolescent primary health care in six clinics and twelve schools in the Western Cape province of South Africa through a facility questionnaire to determine various aspects of adolescent healthcare services and the clinical characteristics of adolescents using these services. We shall observe the processes that adolescents follow to receive health care, including their consultations with clinicians. Focus group discussions with school and clinic managers, clinicians, community-based organisations, policymakers and adolescents will provide a broad understanding of existing adolescent health care. We shall then refine the PACK Adolescent guide and training programme following a theory of change workshop (a workshop intended to develop a system of ideas to explain how we think change happens in the area we want to address, and how we intend to work to influence these changes) with key stakeholders. The adapted intervention will then be piloted and evaluated in the six clinics and twelve linked schools (primary and secondary). Through observations of the use of PACK Adolescent and resultant changes to the clinical pathway for adolescents as well as interviews with adolescents, clinic clinicians and school health nurses we will see how PACK Adolescent meets the needs of adolescents and clinicians along with the challenges to its use. A survey questionnaire completed by clinicians in schools and clinics along with routine health data will help us to assess its impact. The outcomes of the study will provide an understanding of adolescent primary healthcare across schools and clinics. It will allow us to identify the minimum system strengthening components needed to support adolescent healthcare in an integrated way in both health and education sectors. We shall share these outcomes with participating schools and clinics, as well as the Departments of Health and Education in the province and nationally. They will inform the design of a future study evaluating whether PACK Adolescent leads to improved outcomes for young people.

There are important gaps in the primary care delivery of child health in low and middle income countries: there is little user-friendly guidance for health workers seeing children over the age of 5 or those with increasingly common long term conditions like asthma, and preventive care (like vaccinations and growth checks) is not smoothly integrated with sick-child services, making for disjointed care for children, their caregivers and health workers. The Knowledge Translation Unit in Cape Town has developed a health systems intervention for adult primary care (Practical Approach to Care Kit - PACK) that has become entrenched in health services throughout South Africa and is being piloted elsewhere in Africa and in South America. PACK comprises a clinical guide, a training package and work at policy and management level to prepare the system and the health worker for its implementation. The unit has conducted rigorous implementation science studies showing that PACK improves care and patient outcomes. The reasons for PACK's success are that it takes a comprehensive, simple approach to clinical care while tackling the systems issues that make improving primary care difficult. A partnership with the British Medical Journal is helping take PACK to a global audience (www.pack.bmj.com). In response to the gaps described above and to requests from those using PACK Adult, the KTU has now developed the first version of the PACK Child clinical guide in collaboration with local government in the Western Cape province of South Africa, and plans to implement it alongside PACK Adult in several countries. Before rollout, however, we need to work out two things: One, how best to implement the PACK Child programme in a system that has multiple other programmes and priorities - and limited capacity - and two, how best to evaluate whether PACK Child does indeed improve the care and health of children. This Foundation Grant will support the KTU to: - Develop and pilot the PACK Child health systems intervention package. - Design the research protocol to evaluate the PACK Child health systems intervention. - Establish a PACK Child Advisory Board. Each of these activities will draw on stakeholders from policy makers to nurses and doctors to children themselves to ensure that the PACK Child health systems intervention and its evaluation speaks to the needs of those who will use and benefit from it.

We conducted two studies that demonstrated the need to improve the way trial participants are told about potential trial benefits and harms. In the first, we found that half of the participants in trials who take a placebo treatment (like a sugar pill) report having a negative 'side effect.' 1 in 5 of the participants who took a placebo dropped out due to an intervention 'side effect.' There are many reasons for this, including negative expectations. A trial participant might be warned about a possible side effect in a way that caused them to expect, and then actually experience this side effect. Negative side effects among patients taking placebos were more common in pain-related, cancer, and mental health trials. For example, in a trial of aspirin or sulfinpyrazone for treating chest pain, some patients were told the drug might cause stomach pain, and others were not. The patients who were told about experiencing stomach pain were six times more likely to withdraw from the trial because of stomach pain. In our second study, we found that patient information leaflets (PILs) do not always tell trial participants what they understand or want. We looked at 33 PILs and found that the way information about harms was shared did not seem to follow any logical pattern. Most of them had more information about harms than benefits, and some did not mention benefits at all. This seems to be because researchers feel they have to disclose every problem as part of ethical informed consent. Our background research also highlighted an ethical issue. The way in which information about harms of participating in trials is shared can actually cause harm. Therefore, presenting information about trial harms in a scary way (especially if benefits are not mentioned) could be unethical if it causes unnecessary harm. Patient representatives and other stakeholders assist in developing PILs. However, there is currently no guidance specifically on how they should reflect about the way potential benefits and harms are presented to avoid worrying trial participants unnecessarily. We propose to develop a method for presenting benefit and harm information within PILs that rigorously considers patient, doctor and research ethics committee member views. We will then test to see whether they reduce side effects and improve trial recruitment rates. We will achieve our aim in five steps. 1. We will survey stakeholders to understand their views about how the information about trial participation harms and benefits should be communicated in PILs. The information should provide facts and at the same time not cause 'information-induced harm.' The stakeholders will include patients, research ethics committee members, clinicians, medico-legal experts, regulators, and clinical trial managers. They will be chosen based on their experience with one of the conditions that our background research identified as being sensitive to information-induced harm (see 'Objectives'). 2. We will then work with our advisory group to identify principles from the stakeholder interviews. The group will include applicants, patient representatives, and REC members. 3. Using the principles, we will design PILs for five trials. We will call these 'PrinciPILs'. 4. We will then compare PrinciPILs with standard PILs in the trials. We will test whether patients exposed to PrinciPILs had fewer negative side effects. 5. Finally, we will develop and disseminate guidance to relevant stakeholders so that those providing information about potential trial or treatment benefits and harms can generate PrinciPILs. The main outputs from our study will be a report on effects of using PrinciPILs and guidelines designing them. We expect our findings will be useful outside trials and in clinical practice, where clinicians also need to present information about harms and benefits of interventions to patients.