<b>Background:</b> The Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach offers a transparent framework for rating the certainty of evidence in systematic reviews. Concerns, however, have been raised that use of GRADE beyond biomedical interventions frequently downgrades the 'best evidence possible' for many complex interventions. This DPhil thesis aims to (1) further investigate the challenges of using GRADE in systematic reviews of complex interventions, (2) explore how the GRADE approach can be advanced to address these challenges, and (3) inform the write-up and dissemination of a new GRADE guidance for complex interventions. <b>Methods:</b> To address the broad aims of this thesis a range of methodological approaches were employed, primarily drawing on the best-practice techniques for developing research reporting guidelines (see Chapter 2). First, a systematic literature review method was used to establish whether an adequate system already exists for rating the certainty of evidence for complex interventions and informing the need for a new guidance (Chapter 3). Further consultation with experts, including semi-structured interviews with review authors and GRADE methodologists, provided a nuanced understanding of the challenges of applying GRADE in reviews of complex interventions and suggestions for advancing the guidance on GRADE (Chapter 4). Agreement around these suggestions was explored in a Delphi-based online expert panel (Chapter 5), and the content of the new GRADE guidance for complex interventions was discussed indepth in a three-day expert meeting held in Oxford in May 2017 (Chapter 6). <b>Results:</b> The systematic literature review identified a few systems attempting to modify GRADE for public health interventions; however, there was little reporting of rigorous procedures in the development and dissemination of these systems. Qualitative interviews captured differences in views on GRADE use between review authors and GRADE methodologists. Specifically, GRADE methodologists found it critical to consider GRADE from the beginning of the review process and exercise judgment in GRADE ratings. Review authors, on the other hand, often thought of GRADE as an 'annoying add-on' at the end of the review process and felt challenged by the need to be more interpretative with evidence and sift through many publications on GRADE. Suggestions were made to enhance the GRADE guidance. No significant disagreement was found in the online expert panel on any domain of evidence, and the expert meeting provided further insights into the content of the new GRADE guidance for complex interventions. Participants agreed that the new guidance should specify the meaning of the construct of 'certainty of evidence' for complex interventions, consider revisions of the initial categorisation of evidence based on study design, and better assess the coherence of the causal pathway of complex interventions. <b>Conclusion:</b> This thesis work consolidates up-to-date methodological knowledge on reviewing complex interventions by providing critical examination of the existing approaches and new insights. In transparent reporting of the research phases, it informs development of a new GRADE guidance on rating the certainty of evidence in systematic reviews of complex interventions.

Randomised trials of complex interventions are typically designed, conducted, and analysed as if they are drug trials. Although there are many parallels there are also a number of important distinctions, which are seldom considered when designing individual trials. One of these concerns the involvement of therapists in delivering psychotherapy. Systematic reviews and meta-analyses provide an opportunity for exploring the full range and complexity of issues encountered in realistically complex situations. The first objective of the thesis was therefore to develop a conceptual framework for understanding the role of the therapist in trial designs. It was addressed by a review of the psychotherapy and statistical literatures structured according to the broad concepts of precision, internal and external validity and refined on the basis of a systematic methodological review of Cochrane reviews meta-analysing trials involving psychotherapy. The second objective was then to review, adapt, illustrate and compare methods for meta-analysing psychotherapy trials with nested designs. Methods for meta-analysing ICC estimates, absolute and standardised mean differences were adapted to allow for heteroscedasticity between treatments at the therapist- and patient- levels. These were illustrated using the example of counselling in primary care, with comparisons being made between aggregate and one-step approaches to the meta-analysis of individual-patient-data.It was argued that the therapist has two roles in randomised trials. Firstly, they are one component of a multi-component intervention, and are thus a potential treatment variable. Second, the nesting of patients within therapists creates an additional level in the design, so the therapist is also an experimental unit. The inability to conceal or randomise allocations leads to observational components within the trial design and to heteroscedasticity which deserves more attention. Characterising complex interventions, like psychotherapy, with more than one treatment variable could facilitate greater understanding of their components, how they interact, which are important, to what extent, and for whom. It also brings what is currently referred to as process research into the remit of trials, enabling a more complete evaluation of the causal effects. The broad concept of multiple experimental units makes cluster-randomised, longitudinal, multi-centre, crossover, therapist- and group-based intervention trials special cases of a more general class of multilevel trial. All involve clustering effects; their nature and the appropriate statistical model varying according to the design. Methods were proposed for the meta-analysis of continuous outcome data for two-level nested designs. A general approach was adopted, where possible, to incorporate methods covering cluster-randomised trials and the Behrens-Fisher problem. It was clear that this is a relatively untouched methodological area in need of further exploration. For the same reasons as it became necessary to summarise clinical research, it is recommended that systematic methodological reviews be carried out on a larger scale in future.

Yes / The prevalence of dementia is rising worldwide and many people will die with the disease. Symptoms towards the end of life may be inadequately managed and informal and professional carers poorly supported. There are few evidence-based interventions to improve end-of-life care in advanced dementia. To develop an integrated, whole systems, evidence-based intervention that is pragmatic and feasible to improve end-of-life care for people with advanced dementia and support those close to them. Design: A realist-based approach in which qualitative and quantitative data assisted the development of statements. These were incorporated into the RAND/UCLA appropriateness method to achieve consensus on intervention components. Components were mapped to underlying theory of whole systems change and the intervention described in a detailed manual. Setting/participants: Data were collected from people with dementia, carers and health and social care professionals in England, from expert opinion and existing literature. Professional stakeholders in all four countries of the United Kingdom contributed to the RAND/UCLA appropriateness method process. Results: A total of 29 statements were agreed and mapped to individual, group, organisational and economic/political levels of healthcare systems. The resulting main intervention components are as follows: (1) influencing local service organisation through facilitation of integrated multi-disciplinary care, (2) providing training and support for formal and informal carers and (3) influencing local healthcare commissioning and priorities of service providers. Conclusion: Use of in-depth data, consensus methods and theoretical understanding of the intervention components produced an evidence-based intervention for further testing in end-of-life care in advanced dementia.

Yes / Recruiting isolated older adults to clinical trials is complex, time-consuming and difficult. Previous studies have suggested querying existing databases to identify appropriate potential participants. We aim to compare recruitment techniques (general practitioner (GP) mail-outs, community engagement and clinician referrals) used in three randomised controlled trial (RCT) studies assessing the feasibility or effectiveness of two preventative interventions in isolated older adults (the Lifestyle Matters and Putting Life In Years interventions). Methods: During the three studies (the Lifestyle Matters feasibility study, the Lifestyle Matters RCT, the Putting Life In Years RCT) data were collected about how participants were recruited. The number of letters sent by GP surgeries for each study was recorded. In the Lifestyle Matters RCT, we qualitatively interviewed participants and intervention facilitators at 6 months post randomisation to seek their thoughts on the recruitment process. Results: Referrals were planned to be the main source of recruitment in the Lifestyle Matters feasibility study, but due to a lack of engagement from district nurses, community engagement was the main source of recruitment. District nurse referrals and community engagement were also utilised in the Lifestyle Matters and Putting Life In Years RCTs; both mechanisms yielded few participants. GP mail-outs were the main source of recruitment in both the RCTs, but of those contacted, recruiting yield was low (< 3%). Facilitators of the Lifestyle Matters intervention questioned whether the most appropriate individuals had been recruited. Participants recommended that direct contact with health professionals would be the most beneficial way to recruit. Conclusions: Recruitment to the Lifestyle Matters RCT did not mirror recruitment to the feasibility study of the same intervention. Direct district nurse referrals were not effective at recruiting participants. The majority of participants were recruited via GP mail-outs, which may have led to isolated individuals not being recruited to the trials. Further research is required into alternative recruitment techniques, including respondent-driven sampling plus mechanisms which will promote health care professionals to recruit vulnerable populations to research. / Medical Research Council (grant number G1001406); Sheffield Health and Social Research Consortium; National Institute for Health Research Public Health Research programme (project number 09/ 3004/01)

Yes / Recruiting isolated older adults to clinical trials is complex, time-consuming and difficult. Previous studies have suggested querying existing databases to identify appropriate potential participants. We aim to compare recruitment techniques (general practitioner (GP) mail-outs, community engagement and clinician referrals) used in three randomised controlled trial (RCT) studies assessing the feasibility or effectiveness of two preventative interventions in isolated older adults (the Lifestyle Matters and Putting Life In Years interventions). Methods: During the three studies (the Lifestyle Matters feasibility study, the Lifestyle Matters RCT, the Putting Life In Years RCT) data were collected about how participants were recruited. The number of letters sent by GP surgeries for each study was recorded. In the Lifestyle Matters RCT, we qualitatively interviewed participants and intervention facilitators at 6 months post randomisation to seek their thoughts on the recruitment process. Results: Referrals were planned to be the main source of recruitment in the Lifestyle Matters feasibility study, but due to a lack of engagement from district nurses, community engagement was the main source of recruitment. District nurse referrals and community engagement were also utilised in the Lifestyle Matters and Putting Life In Years RCTs; both mechanisms yielded few participants. GP mail-outs were the main source of recruitment in both the RCTs, but of those contacted, recruiting yield was low (< 3%). Facilitators of the Lifestyle Matters intervention questioned whether the most appropriate individuals had been recruited. Participants recommended that direct contact with health professionals would be the most beneficial way to recruit. Conclusions: Recruitment to the Lifestyle Matters RCT did not mirror recruitment to the feasibility study of the same intervention. Direct district nurse referrals were not effective at recruiting participants. The majority of participants were recruited via GP mail-outs, which may have led to isolated individuals not being recruited to the trials. Further research is required into alternative recruitment techniques, including respondent-driven sampling plus mechanisms which will promote health care professionals to recruit vulnerable populations to research. / The Lifestyle Matters RCT was funded by the Medical Research Council (grant number G1001406); Sheffield Health and Social Research Consortium; National Institute for Health Research Public Health Research programme (project number 09/ 3004/01)

Yes / In this article, we present an exemplar of the initial theory-building phase of theory-driven evaluation for the PARTNERS2 project, a collaborative care intervention for people with experience of psychosis in England. Initial theory-building involved analysis of the literature, interviews with key leaders and focus groups with service users. The initial programme theory was developed from these sources in an iterative process between researchers and stakeholders (service users, practitioners, commissioners) involving four activities: articulation of 442 explanatory statements systematically developed using realist methods; debate and consensus; communication; and interrogation. We refute two criticisms of theory-driven evaluation of complex interventions. We demonstrate how the process of initial theory-building made a meaningful contribution to our complex intervention in five ways. Although time-consuming, it allowed us to develop an internally coherent and well-documented intervention. This study and the lessons learnt provide a detailed resource for other researchers wishing to build theory for theory-driven evaluation. / This research was funded by a UK NIHR Programme Grant (RP-PG-0611- 20004) and the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care South West Peninsula (NIHR CLAHRC South West Peninsula).

Contexte : L'insuffisance cardiaque (IC) est un syndrome aux étiologies variées, et hétérogène dans ses présentations cliniques, dont l'incidence augmente avec l'âge et dont le pronostic reste sombre. Les facteurs pronostiques dans l'IC ont été largement décrits dans la littérature, en particulier dans l'IC à fraction d'éjection réduite (ICFER), à plus ou moins long terme. Des interventions thérapeutiques médicamenteuses, chirurgicales, ou complexes sont recommandées pour traiter ce syndrome. Objectifs : Les objectifs de ce travail étaient d'évaluer la valeur pronostique des marqueurs de fibrose chez les patients hypertendus, de décrire la survie et d'identifier les facteurs pronostiques chez les patients en IC, à très long terme chez les ICFER, et à 1 an chez les IC à fraction d'éjection préservée (ICFEP), et de mesurer l'impact sur les hospitalisations de 2 interventions complexes recommandées. Méthodes : Une première étude de cohorte a été menée chez des patients hypertendus, avec un recueil à l'inclusion des dosages sériques des marqueurs du renouvellement de la matrice extracellulaire cardiaque (MRMEC) et un suivi pendant 6 ans. Deux autres études de cohorte ont permis d'identifier les facteurs pronostiques de décès à 15 ans chez des patients ICFER (EPICAL), et à 1 an chez les ICFEP (Odin). Enfin, les données du PMSI ont permis de mesurer l'impact pronostique de la prise en charge dans une unité spécialisée (UTIC) et d'une prise en charge de type disease management (ICALOR). Résultats : La première étude a mis en évidence la valeur pronostique des MRMEC chez des patients hypertendus. L'étude EPICAL a montré que l'âge avancé, le diabète, l'insuffisance rénale chronique, l'ancienneté de l'IC, les antécédents de décompensation, la tachycardie, une fraction d'éjection réduite, et une hyponatrémie étaient associés à une mortalité plus élevée à 15 ans chez les ICFER. L'étude ODIN a montré la valeur pronostique négative de facteurs sociaux comme le fait de vivre seul ou la limitation des activités, chez les ICFEP. Enfin, les prises en charge en UTIC et par le réseau ICALOR étaient associées à une diminution des hospitalisations pour IC, mesurable à l'échelle populationnelle. Conclusion : Ces travaux permettent d'envisager de nouvelles pistes de prévention primaire et secondaire, afin de diminuer l'incidence, la morbidité et la mortalité liées à l'IC / Context: Heart failure (HF) is a heterogeneous syndrome with various aetiologies. HF incidence increases with age, and the prognosis remains poor. Prognostic factors have been widely described in the literature, especially in HF with reduced ejection fraction with short-term or mid-term follow-up. Medications, surgical interventions, and complex interventions are part of the current guidelines. Objectives: To assess the prognostic value of fibrosis markers in hypertensive patients; to describe the survival and to detect prognostic factors, first at 15 years in HF patients with reduced ejection fraction (HFREF), and second at one year in HF patients with preserved ejection fraction (HFPEF); and to assess the prognostic impact of 2 complex interventions on HF hospitalisations. Methods: Cardiac extracellular matrix serum markers (CEMSM) were assessed at baseline in a first cohort of hypertensive patients followed-up for 6 years. Two other cohort studies were used to detect prognostic factors associated with very-long term mortality in HFREF (EPICAL), and with 1-year mortality in HFPEF (Odin). And the national diagnostic related group database was used to assess the prognostic impact of a HF unit (UTIC) and a disease management programme (ICALOR). Results: The first cohort study highlighted the prognostic value of CEMSM in hypertensive patients. In EPICAL study, older age, diabetes mellitus, chronic renal failure, time from HF onset, history of hospitalisations for worsening HF, tachycardia, a low left ventricular ejection fraction, and hyponatraemia were associated with a higher mortality in HFREF patients. In Odin study, we highlighted the negative prognostic impact of social factors, such as living alone or daily activity limitation, in HFPEF patients. Both UTIC and ICALOR were associated with a decrease in HF hospitalisations compatible with a population impact. Conclusion: These studies offer new insights for primary and secondary prevention strategies that could eventually lead to decrease HF incidence, HF morbidity, and HF mortality

Introduction: Medication adherence is vital to ensuring optimal patient outcomes, particularly amongst older adults prescribed polypharmacy. However, complex interventions aimed at improving adherence have shown only limited effectiveness. To maximise effectiveness, the Medical Research Council (MRC) supports the use of both evidence and theory in developing interventions. Feasibility and pilot testing is then recommended to optimise interventions in advance of definitive trials. The aim of this research was to develop a novel complex intervention (using evidence and theory) to improve adherence in older adults (prescribed polypharmacy) and to test the feasibility of delivering this in community pharmacies. Methods: The presented research models the MRC complex intervention framework and focuses on development and feasibility testing phases. Firstly, a systematic review was conducted to address an identified evidence gap in relation to theory-based adherence interventions previously delivered to older adults prescribed polypharmacy. Qualitative research was then conducted to explore older patients’ adherence behaviour and identify determinants (barriers, facilitators) to target for change. Using the Theoretical Domains Framework (TDF) as a lens, key domains were selected for targeting and mapped to behaviour change techniques (BCTs) using established methods. These BCTs formed the basis of a complex intervention that was delivered to older patients by community pharmacists (CPs) as part of a small-scale feasibility study. In addition to exploring older patients’ adherence behaviours, further research focused on CPs’ clinical behaviour in relation to providing medication adherence support (MAS). The qualitative TDF-based methods used in the patient study were extended and a mixed methods (qualitative, quantitative) approach was used to identify determinants influencing CPs’ behaviour. Key target domains were identified and mapped to BCTs that could be directed at CPs (e.g. in a training package) to improve future implementation of the patient intervention. Results: The systematic review found that adherence interventions delivered to older patients prescribed polypharmacy were rarely based on theory, supporting the need for further research. The qualitative research conducted with older patients identified eight key domains (e.g. ‘Beliefs about consequences’, Memory, attention and decision process’) that could be targeted and these domains were mapped to 11 BCTs (e.g. ‘Prompts/cues’, ‘Self-monitoring’) which formed the basis of a complex intervention. The feasibility study demonstrated that the intervention was highly acceptable to both patients and CPs but some modifications were suggested. It also highlighted the need for additional research that focuses on CPs’ behaviour (i.e. MAS provision). Findings from the mixed methods study on CPs’ behaviour led to the identification of seven key domains that could be targeted for change (e.g. ‘Skills’, ‘Motivation and goals’). Eighteen BCTs were then selected for inclusion in a training package (e.g. ‘Demonstration of the behaviour’) or for delivery alongside the patient intervention in future research (e.g. ‘Rewards/incentives’) to improve implementation. Discussion/Conclusion: The MRC framework served as a useful guide for developing a complex intervention to improve adherence in older patients prescribed polypharmacy. This systematic theory-based approach that involved explicitly linking theoretical domains to intervention components (BCTs) will aid future replication and understanding of how the intervention aims to bring about behaviour change. Aside from targeting patients’ adherence behaviours, this research emphasised the importance of exploring the behaviours of intervention providers (i.e. CPs) to enhance implementation. Future research will involve pilot testing a refined version of the patient intervention and CP training package to establish if a definitive trial of effectiveness (e.g. randomised controlled trial) is warranted.

BACKGROUND:Integrative medicine (IM) is a patient-centered, healing-oriented clinical paradigm that explicitly includes all appropriate therapeutic approaches whether they originate in conventional or complementary medicine (CM). While there is some evidence for the clinical and cost-effectiveness of IM practice models, the existing evidence base for IM depends largely on studies of individual CM therapies. This may in part be due to the methodological challenges inherent in evaluating a complex intervention (i.e., many interacting components applied flexibly and with tailoring) such as IM.METHODS/DESIGN:This study will use a combination of observational quantitative and qualitative methods to rigorously measure the health and healthcare utilization outcomes of the University of Arizona Integrative Health Center (UAIHC), an IM adult primary care clinic in Phoenix, Arizona. There are four groups of study participants. The primary group consists of clinic patients for whom clinical and cost outcomes will be tracked indicating the impact of the UAIHC clinic (n=500). In addition to comparing outcomes pre/post clinic enrollment, where possible, these outcomes will be compared to those of two matched control groups, and for some self-report measures, to regional and national data. The second and third study groups consist of clinic patients (n=180) and clinic personnel (n=15-20) from whom fidelity data (i.e., data indicating the extent to which the IM practice model was implemented as planned) will be collected. These data will be analyzed to determine the exact nature of the intervention as implemented and to provide covariates to the outcomes analyses as the clinic evolves. The fourth group is made up of patients (n=8) whose path through the clinic will be studied in detail using qualitative (periodic semi-structured interviews) methods. These data will be used to develop hypotheses regarding how the clinic works.DISCUSSION:The US health care system needs new models of care that are more patient-centered and empower patients to make positive lifestyle changes. These models have the potential to reduce the burden of chronic disease, lower the cost of healthcare, and offer a sustainable financial paradigm for our nation. This protocol has been designed to test whether the UAIHC can achieve this potential.TRIAL REGISTRATION:Clinical Trials.gov NCT01785485.

Background: To improve health in the population, public health interventions must be successfully implemented within organisations, requiring behaviour change in health service providers as well as in the target population group. Such behavioural change is seldom easily achieved. The purpose of this study was to examine the outcomes of a child health promotion programme (The Salut Programme) on professionals' self-reported health promotion practices, and to investigate perceived facilitators and barriers for programme implementation. Methods: A before-and-after design was used to measure programme outcomes, and qualitative data on implementation facilitators and barriers were collected on two occasions during the implementation process. The sample included professionals in antenatal care, child health care, dental services and open pre-schools (n=144 pre-implementation) in 13 out of 15 municipalities in a Swedish county. Response rates ranged between 81% and 96% at the four measurement points. Results: Self-reported health promotion practices and collaboration were improved in all sectors at follow up. Significant changes included: 1) an increase in the extent to which midwives in antenatal care raised issues related to men's violence against women, 2) an increase in the extent to which several lifestyle topics were raised with parents/clients in child health care and dental services, 3) an increased use of motivational interviewing (MI) and separate 'fathers visits' in child health care 4) improvements in the supply of healthy snacks and beverages in open pre-schools and 5) increased collaboration between sectors. Main facilitators for programme implementation included cross-sectoral collaboration and sector-specific work manuals/questionnaires for use as support in everyday practice. Main barriers included high workload, and shortage of time and staff. Conclusion: This multisectoral programme for health promotion, based on sector-specific intervention packages developed and tested by end users, and introduced via interactive multisectoral seminars, shows potential for improving health promotion practices and collaboration across sectors. Consideration of the key facilitators and barriers for programme implementation as highlighted in this study can inform future improvement efforts.