BackgroundDrug-related problems and potentially inappropriate prescribing impose a huge burden on patients and the health-care system. The most widely used tools for appropriate prescription in older adults in England and in other European countries are the Screening Tool of Older People’s Prescriptions (STOPP)/Screening Tool to Alert to the Right Treatment (START) tools. STOPP/START tools support medicines optimisation for older adults.ObjectivesTo identify, test and refine the programme theories underlying how interventions based on the STOPP/START tools are intended to work, for whom, in what circumstances and why, as well as the resource use and cost requirements or impacts.DesignA realist synthesis.SettingPrimary care, hospital care and nursing homes.PatientsPatients aged ≥ 65 years.InterventionsAny intervention based on the use of the STOPP/START tools.Review methodsDatabase and web-searching was carried out to retrieve relevant evidence to identify and test programme theories about how interventions based on the use of the STOPP/START tools work. A project reference group made up of health-care professionals, NHS decision-makers, older people, carers and members of the public was set up. In phase 1 we identified programme theories about STOPP/START interventions on how, for whom, in what contexts and why they are intended to work. We searched the peer-reviewed and grey literature to identify documents relevant to the research questions. We interviewed experts in the field in our reference group to gain input on our list of candidate context–mechanism–outcome configurations, to identify additional context–mechanism–outcome configurations and to identify additional literature and/or relevant concepts. In phase 2 we reviewed and synthesised relevant published and unpublished empirical evidence and tested the programme theories using evidence from a larger set of empirical studies.ResultsWe developed a single logic model structured around three key mechanisms: (1) personalisation, (2) systematisation and (3) evidence implementation. Personalisation: STOPP/START-based interventions are based on shared decision-making, taking into account patient preferences, experiences and expectations (mechanisms), leading to increased patient awareness, adherence, satisfaction, empowerment and quality of life (outcomes). Systematisation: STOPP/START tools provide a standardised/systematic approach for medication reviews (mechanisms), leading to changes in professional and organisational culture and burden/costs (outcomes). Evidence implementation: delivery of STOPP/START-based interventions is based on the implementation of best evidence (mechanisms), reducing adverse outcomes through appropriate prescribing/deprescribing (outcomes). For theory testing, we identified 40 studies of the impact of STOPP/START-based interventions in hospital settings, nursing homes, primary care and community pharmacies. Most of the interventions used multiple mechanisms. We found support for the impact of the personalisation and evidence implementation mechanisms on selected outcome variables, but similar impact was achieved by interventions not relying on these mechanisms. We also observed that the impact of interventions was linked to the proximity of the selected outcomes to the intervention in the logic model, resulting in a clearer benefit for appropriateness of prescribing, adverse drug events and prescription costs.LimitationsNone of the available studies had been explicitly designed for evaluating underlying causal mechanisms, and qualitative information was sparse.ConclusionsNo particular configuration of the interventions is associated with a greater likelihood of improved outcomes in given settings.Study registrationThis study is registered as PROSPERO CRD42018110795.FundingThis project was funded by the National Institute for Health Research (NIHR) Health Services and Delivery Research programme and will be published in full inHealth Services and Delivery Research; Vol. 9, No. 23. See the NIHR Journals Library website for further project information.

Background Cognitive impairment is common in older people admitted to hospital, but previous research has focused on single conditions. Objective This project sits in phase 0/1 of the Medical Research Council Framework for the Development and Evaluation of Complex Interventions. It aims to develop an understanding of current health-care outcomes. This will be used in the future development of a multidomain intervention for people with confusion (dementia and cognitive impairment) in general hospitals. The research was conducted from January 2015 to June 2018 and used data from people admitted between 2012 and 2013. Design For the review of outcomes, the systematic review identified peer-reviewed quantitative epidemiology measuring prevalence and associations with outcomes. Screening for duplication and relevance was followed by full-text review, quality assessment and a narrative review (141 papers). A survey sought opinion on the key outcomes for people with dementia and/or confusion and their carers in the acute hospital (n = 78). For the analysis of outcomes including cost, the prospective cohort study was in a medical admissions unit in an acute hospital in one Scottish health board covering 10% of the Scottish population. The participants (n = 6724) were older people (aged ≥ 65 years) with or without a cognitive spectrum disorder who were admitted as medical emergencies between January 2012 and December 2013 and who underwent a structured nurse assessment. ‘Cognitive spectrum disorder’ was defined as any combination of delirium, known dementia or an Abbreviated Mental Test score of < 8 out of 10 points. The main outcome measures were living at home 30 days after discharge, mortality within 2 years of admission, length of stay, re-admission within 2 years of admission and cost. Data sources Scottish Morbidity Records 01 was linked to the Older Persons Routine Acute Assessment data set. Results In the systematic review, methodological heterogeneity, especially concerning diagnostic criteria, means that there is significant overlap in conditions of patients presenting to general hospitals with confusion. Patients and their families expect that patients are discharged in the same or a better condition than they were in on admission or, failing that, that they have a satisfactory experience of their admission. Cognitive spectrum disorders were present in more than one-third of patients aged ≥ 65 years, and in over half of those aged ≥ 85 years. Outcomes were worse in those patients with cognitive spectrum disorders than in those without: length of stay 25.0 vs. 11.8 days, 30-day mortality 13.6% vs. 9.0%, 1-year mortality 40.0% vs. 26.0%, 1-year mortality or re-admission 62.4% vs. 51.5%, respectively (all p < 0.01). There was relatively little difference by cognitive spectrum disorder type; for example, the presence of any cognitive spectrum disorder was associated with an increased mortality over the entire period of follow-up, but with different temporal patterns depending on the type of cognitive spectrum disorder. The cost of admission was higher for those with cognitive spectrum disorders, but the average daily cost was lower. Limitations A lack of diagnosis and/or standardisation of diagnosis for dementia and/or delirium was a limitation for the systematic review, the quantitative study and the economic study. The economic study was limited to in-hospital costs as data for social or informal care costs were unavailable. The survey was conducted online, limiting its reach to older carers and those people with cognitive spectrum disorders. Conclusions Cognitive spectrum disorders are common in older inpatients and are associated with considerably worse health-care outcomes, with significant overlap between individual cognitive spectrum disorders. This suggests the need for health-care systems to systematically identify and develop care pathways for older people with cognitive spectrum disorders, and avoid focusing on only condition-specific pathways. Future work Development and evaluation of a multidomain intervention for the management of patients with cognitive spectrum disorders in hospital. Study registration This study is registered as PROSPERO CRD42015024492. Funding This project was funded by the National Institute for Health Research (NIHR) Health Services and Delivery Research programme and will be published in full in Health Services and Delivery Research; Vol. 9, No. 8. See the NIHR Journals Library website for further project information.

Background Many global health challenges may be targeted by changing people’s behaviour. Behaviours including cigarette smoking, physical inactivity and alcohol misuse, as well as certain dietary behaviours, contribute to deaths and disability by increasing the risk of cancers, cardiovascular diseases and diabetes. Interventions have been designed to change these health behaviours with a view to reducing these health risks. However, the effectiveness of these interventions has been quite variable and further information is needed to enhance their success. More information is needed about the specific processes that underlie the effectiveness of intervention strategies. Aim Researchers have developed a taxonomy of 93 behaviour change techniques (i.e. the active components of an intervention that bring about behavioural change), but little is known regarding their potential mechanisms of action (i.e. the processes through which a behaviour change technique affects behaviour). We therefore aimed to examine links between behaviour change techniques and mechanisms of action. Method First, we conducted a literature synthesis study of 277 behaviour change intervention studies, from which we extracted information on links, described by authors, between behaviour change techniques and mechanisms of action, and identified an average of 10 links per intervention report. Second, behaviour change experts (n = 105) were engaged in a three-round consensus study in which they discussed and rated their confidence in the presence/absence of ‘links’ and ‘non-links’ between commonly used behaviour change techniques (n = 61) and a set of mechanisms of action (n = 26). Ninety links and 460 ‘non-links’ reached the pre-set threshold of 80% agreement. To enhance the validity of these results, a third study was conducted that triangulated the findings of the first two studies. Discrepancies and uncertainties between the studies were included in a reconciliation consensus study with a new group of experts (n = 25). The final results identified 92 definite behaviour change technique–mechanism of action links and 465 definite non-links. In a fourth study, we examined whether or not groups of behaviour change techniques used together frequently across interventions revealed shared theoretical underpinnings. We found that experts agreed on the underlying theory for three groups of behaviour change techniques. Results Our results are potentially useful to policy-makers and practitioners in selecting behaviour change techniques to include in behaviour change interventions. However, our data do not demonstrate that the behaviour change techniques are effective in targeting the mechanism of action; rather, the links identified may be the ‘best bets’ for interventions that are effective in changing mechanisms of action, and the non-links are unlikely to be effective. Researchers examining effectiveness of interventions in either primary studies or evidence syntheses may consider these links for further investigation. Conclusion To make our results usable by researchers, practitioners and policy-makers, they are available in an online interactive tool, which enables discussion and collaboration (https://theoryandtechniquetool.humanbehaviourchange.org/); accessed 1 March 2020. This work, building on previous work to develop the behaviour change technique taxonomy, is part of an ongoing programme of work: the Human Behaviour Change Project (www.humanbehaviourchange.org/; accessed 1 March 2020). Funding This project was funded by the Medical Research Council via its Methodology Panel: ‘Developing methodology for designing and evaluating theory-based complex interventions: an ontology for linking behaviour change techniques to theory’ (reference MR/L011115/1).

Background Seven-day working in hospitals is a current priority of international health research and policy. Previous research has shown variability in delivering evidence-based clinical interventions across different times of the day and week. We aimed to identify factors influencing such variations in London hyperacute stroke units. Objectives To investigate variations in quality of acute stroke care and outcomes by day and time of admission in London hyperacute stroke units, and to identify factors influencing such variations. Design This was a prospective cohort study using anonymised patient-level data from the Sentinel Stroke National Audit Programme. Factors influencing variations in care and outcomes were studied through interview and observation data. Setting The setting was acute stroke services in London hyperacute stroke units. Participants A total of 7094 patients with a primary diagnosis of stroke took part. We interviewed hyperacute stroke unit staff (n = 76), including doctors, nurses, therapists and administrators, and 31 patients and carers. We also conducted non-participant observations of delivery of care at different times of the day and week (n = 45, ≈102 hours). Intervention Hub-and-spoke model for care of suspected acute stroke patients in London with performance standards was designed to deliver uniform access to high-quality hyperacute stroke unit care across the week. Main outcome measures Indicators of quality of acute stroke care, mortality at 3 days after admission, disability at the end of the inpatient spell and length of stay. Data sources Sentinel Stroke National Audit Programme data for all patients in London hyperacute stroke units with a primary diagnosis of stroke between 1 January and 31 December 2014, and nurse staffing data for all eight London hyperacute stroke units for the same period. Results We found no variation in quality of care by day and time of admission across the week in terms of stroke nursing assessment, brain scanning and thrombolysis in London hyperacute stroke units, nor in 3-day mortality nor disability at hospital discharge. Other quality-of-care measures significantly varied by day and time of admission. Quality of care was better if the nurse in charge was at a higher band and/or there were more nurses on duty. Staff deliver ‘front-door’ interventions consistently by taking on additional responsibilities out of hours, creating continuities between day and night, building trusting relationships and prioritising ‘front-door’ interventions. Limitations We were unable to measure long-term outcomes as our request to the Sentinel Stroke National Audit Programme, the Healthcare Quality Improvement Partnership and NHS Digital for Sentinel Stroke National Audit Programme data linked with patient mortality status was not fulfilled. Conclusions Organisational factors influence 24 hours a day, 7 days a week (24/7), provision of stroke care, creating temporal patterns of provision reflected in patient outcomes, including mortality, length of stay and functional independence. Future work Further research would help to explore 24/7 stroke systems in other contexts. We need a clearer understanding of variations by looking at absolute time intervals, rather than achievement of targets. Research is needed with longer-term mortality and modified Rankin Scale data, and a more meaningful range of outcomes. Funding This project was funded by the National Institute for Health Research (NIHR) Health Services and Delivery Research programme and will be published in full in Health Services and Delivery Research; Vol. 8, No. 34. See the NIHR Journals Library website for further project information.

BackgroundEmergency abdominal surgery is associated with poor patient outcomes. We studied the effectiveness of a national quality improvement (QI) programme to implement a care pathway to improve survival for these patients.ObjectivesThe objectives were to assess whether or not the QI programme improves 90-day survival after emergency abdominal surgery; to assess effects on 180-day survival, hospital stay and hospital readmission; and to better understand these findings through an integrated process evaluation, ethnographic study and cost-effectiveness analysis.DesignThis was a stepped-wedge cluster randomised trial. Hospitals were organised into 15 geographical clusters, and commenced the QI programme in random order over 85 weeks. Analyses were performed on an intention-to-treat basis. The primary outcome was analysed using a mixed-effects parametric survival model, adjusting for time-related effects. Ethnographic and economics data were collected in six hospitals. The process evaluation included all hospitals.SettingThe trial was set in acute surgical services of 93 NHS hospitals.ParticipantsPatients aged ≥ 40 years who were undergoing emergency abdominal surgery were eligible.InterventionThe intervention was a QI programme to implement an evidence-based care pathway.Main outcome measuresThe primary outcome measure was mortality within 90 days of surgery. Secondary outcomes were mortality within 180 days, length of hospital stay and hospital readmission within 180 days. The main economic measure was the quality-adjusted life-years.Data sourcesData were obtained from the National Emergency Laparotomy Audit database; qualitative interviews and ethnographic observations; quality-of-life and NHS resource use data were collected via questionnaires.ResultsOf 15,873 eligible patients from 93 NHS hospitals, primary outcome data were analysed for 8482 participants in the usual care group and 7374 in the QI group. The primary outcome occurred in 1393 participants in the usual care group (16%), compared with 1210 patients in the QI group (16%) [QI vs. usual care hazard ratio (HR) 1.11, 95% confidence interval (CI) 0.96 to 1.28]. No differences were found in mortality at 180 days or hospital readmission; there was a small increase in hospital stay in the QI group (HR for discharge 0.90, 95% CI 0.83 to 0.97). There were only modest improvements in care processes following QI implementation. The ethnographic study revealed good QI engagement, but limited time and resources to implement change, affecting which processes teams addressed, the rate of change and eventual success. In some sites, there were challenges around prioritising the intervention in busy environments and in obtaining senior engagement. The intervention is unlikely to be cost-effective at standard cost-effectiveness thresholds, but may be cost-effective over the lifetime horizon.LimitationsSubstantial delays were encountered in securing data access to national registries. Fewer patients than expected underwent surgery and the mortality rate was lower than anticipated.ConclusionsThere was no survival benefit from a QI programme to implement a care pathway for patients undergoing emergency abdominal surgery. The modest impact of the intervention on process measures, despite good clinician engagement, may have been limited by the time and resources needed to improve patient care.Future workFuture QI programmes must balance intervention complexity with the practical realities of NHS services to ensure that such programmes can be delivered with the resources available.Trial registrationCurrent Controlled Trials ISRCTN80682973 andThe Lancetprotocol 13PRT/7655.FundingThis project was funded by the National Institute for Health Research (NIHR) Health Services and Delivery Research programme and will be published in full inHealth Services and Delivery Research; Vol. 7, No. 32. See the NIHR Journals Library website for further project information.

Background Digital and online symptom checkers and assessment services are used by patients seeking guidance about health problems. NHS England is planning to introduce a digital platform (NHS111 Online) to operate alongside the NHS111 urgent-care telephone service. This review focuses on digital and online symptom checkers for urgent health problems. Objectives This systematic review was commissioned to provide NHS England with an independent review of previous research in this area to inform strategic decision-making and service design. Data sources Focused searches of seven bibliographic databases were performed and supplemented by phrase searching for names of symptom checker systems and citation searches of key included studies. The bibliographic databases searched were MEDLINE, EMBASE, The Cochrane Library, CINAHL (Cumulative Index to Nursing and Allied Health Literature), HMIC (Health Management Information Consortium), Web of Science and the Association of Computing Machinery (ACM) Digital Library, from inception up to April 2018. Review methods Brief inclusion criteria were (1) population – general population seeking information online or digitally to address an urgent health problem; (2) intervention – any online or digital service designed to assess symptoms, provide health advice and direct patients to appropriate services; and (3) comparator – telephone or face-to-face assessment, comparative performance in tests or simulations (studies with no comparator were included if they reported relevant outcomes). Outcomes of interest included safety, clinical effectiveness, costs or cost-effectiveness, diagnostic and triage accuracy, use of and contacts with health services, compliance with advice received, patient/carer satisfaction, and equity and inclusion. Inclusion was not restricted by study design. Screening studies for inclusion, data extraction and quality assessment were carried out by one reviewer with a sample checked for accuracy and consistency. Final decisions on study inclusion were taken by consensus of the review team. A narrative synthesis of the included studies was performed and structured around the predefined research questions and key outcomes. The overall strength of evidence for each outcome was classified as ‘stronger’, ‘weaker’, ‘conflicting’ or ‘insufficient’, based on study numbers and design. Results In total, 29 publications describing 27 studies were included. Studies were diverse in their design and methodology. The overall strength of the evidence was weak because it was largely based on observational studies and with a substantial component of non-peer-reviewed grey literature. There was little evidence to suggest that symptom checkers are unsafe, but studies evaluating their safety were generally short term and small scale. Diagnostic accuracy was highly variable between different systems but was generally low. Algorithm-based triage tended to be more risk averse than that of health professionals. Inconsistent evidence was found on effects on service use. There was very limited evidence on patients’ reactions to online triage advice. The studies showed that younger and more highly educated people are more likely to use these services. Study participants generally expressed high levels of satisfaction with digital and online triage services, albeit in uncontrolled studies. Limitations Findings from symptom checker systems for specific conditions may not be applicable to more general systems and vice versa. Studies of symptom checkers as part of electronic consultation systems in general practice were also included, which is a slightly different setting from a general ‘digital 111’ service. Most studies were screened by one reviewer. Conclusions Major uncertainties surround the probable impact of digital 111 services on most outcomes. It will be important to monitor and evaluate the services using all available data sources and by commissioning high-quality research. Future work Priorities for research include comparisons of different systems, rigorous economic evaluations and investigations of patient pathways. Study registration The study is registered as PROSPERO CRD42018093564. Funding The National Institute for Health Research Health Services and Delivery Research programme.

Background The ‘Productive Ward: Releasing Time to Care’™ programme (Productive Ward; PW) was introduced in English NHS acute hospitals in 2007 to give ward staff the tools, skills and time needed to implement local improvements to (1) increase the time nurses spend on direct patient care, (2) improve the safety and reliability of care, (3) improve staff and patient experience and (4) make structural changes on wards to improve efficiency. Evidence of whether or not these goals were met and sustained is very limited. Objective To explore if PW had a sustained impact over the past decade. Design Multiple methods, comprising two online national surveys, six acute trust case studies (including a secondary analysis of local audit data) and telephone interviews. Data sources Surveys of 56 directors of nursing and 35 current PW leads; 88 staff and patient and public involvement representative interviews; 10 ward manager questionnaires; structured observations of 12 randomly selected wards and documentary analysis in case studies; and 14 telephone interviews with former PW leads. Results Trusts typically adopted PW in 2008–9 on their wards using a phased implementation approach. The average length of PW use was 3 years (range < 1 to 7 years). Financial and management support for PW has disappeared in the majority of trusts. The most commonly cited reason for PW’s cessation was a change in quality improvement (QI) approach. Nonetheless, PW has influenced wider QI strategies in around half of the trusts. Around one-third of trusts had impact data relating specifically to PW; the same proportion did not. Early adopters of PW had access to more resources for supporting implementation. Some elements of local implementation strategies were common. However, there were variations that had consequences for the assimilation of PW into routine practice and, subsequently, for the legacies and sustainability of the programme. In all case study sites, material legacies (e.g. display of metrics data; storage systems) remained, as did some processes (e.g. protected mealtimes). Only one case study site had sufficiently robust data collection systems to allow an objective assessment of PW’s impact; in that site, care processes had improved initially (in terms of patient observations and direct care time). Experience of leading PW had benefited the careers of the majority of interviewees. Starting with little or no QI experience, many went on to work on other initiatives within their trusts, or to work in QI at regional or national level within the NHS or in the private sector. Limitations The research draws on participant recall over a lengthy period characterised by evolving QI approaches and system-level change. Conclusions Little robust evidence remains of PW leading to a sustained increase in the time nurses spend on direct patient care or improvements in the experiences of staff and/or patients. PW has had a lasting impact on some ward practices. As an ongoing QI approach continually used to make ongoing improvements, PW has not been sustained, but it has informed current organisational QI practices and strategies in many trusts. The design and delivery of future large-scale QI programmes could usefully draw on the lessons learnt from this study of the PW in England over the period 2008–18. Funding This National Institute for Health Research Health Services and Delivery Research programme.

BackgroundPolicy has been focused on reducing unnecessary emergency department attendances by providing more responsive urgent care services and guiding patients to ‘the right place’. The variety of services has created a complex urgent care landscape for people to access and navigate.ObjectivesTo describe how the public, providers and policy-makers define and make sense of urgent care; to explain how sense-making influences patients’ strategies and choices; to analyse patient ‘work’ in understanding, navigating and choosing urgent care; to explain urgent care utilisation; and to identify potentially modifiable factors in decision-making.DesignMixed-methods sequential design.SettingFour counties in southern England coterminous with a NHS 111 provider area.MethodsA literature review of policy and research combined with citizens’ panels and serial qualitative interviews. Four citizens’ panels were conducted with the public, health-care professionals, commissioners and managers (n = 41). Three populations were sampled for interview: people aged ≥ 75 years, people aged 18–26 years and East European people. In total, 134 interviews were conducted. Analyses were integrated to develop a conceptual model of urgent care help-seeking.FindingsThe literature review identified some consensus between policy and provider perspectives regarding the physiological factors that feature in conceptualisations of urgent care. However, the terms ‘urgent’ and ‘emergency’ lack specificity or consistency in meaning. Boundaries between urgent and emergency care are ill-defined. We constructed a typology that distinguishes three types of work that take place at both the individual and social network levels in relation to urgent care sense-making and help-seeking.Illness workinvolves interpretation and decision-making about the meaning, severity and management of physical symptoms and psychological states, and the assessment and management of possible risks. Help-seeking was guided bymoral work: the legitimation and sanctioning done by service users.Navigation workconcerned choosing and accessing services and relied on prior knowledge of what was available, accessible and acceptable. From these empirical data, we developed a model of urgent care sense-making and help-seeking behaviour that emphasises that work informs the interaction between what we think and feel about illness and the need to seek care (sense-making) and action – the decisions we take and how we use urgent care (help-seeking).LimitationsThe sample population of our three groups may not have adequately reflected a diverse range of views and experiences. The study enabled us to capture people’s views and self-reported service use rather than their actual behaviour.ConclusionsMuch of the policy surrounding urgent and emergency care is predicated on the notion that ‘urgent’ sits neatly between emergency and routine; however, service users in particular struggle to distinguish urgent from emergency or routine care. Rather than focusing on individual sense-making, future work should attend to social and temporal contexts that have an impact on help-seeking (e.g. why people find it more difficult to manage pain at night), and how different social networks shape service use.Future workA whole-systems approach considering integration across a wider network of partners is key to understanding the complex relationships between demand for and access to urgent care.Study registrationThis study is registered as UKCRN 32207.FundingThe National Institute for Health Research Health Services and Delivery Research programme.

BackgroundBreathlessness is a common and distressing symptom of many advanced diseases, affecting around 2 million people in the UK. Breathlessness increases with disease progression and often becomes chronic or refractory. Breathlessness-triggered services that integrate holistic assessment and specialist palliative care input as part of a multiprofessional approach have been developed for this group, offering tailored interventions to support self-management and reduce distress.ObjectivesThe aim was to synthesise evidence on holistic breathlessness services for people with advanced disease and chronic or refractory breathlessness. The objectives were to describe the structure, organisation and delivery of services, determine clinical effectiveness, cost-effectiveness and acceptability, identify predictors of treatment response, and elicit stakeholders’ evidence-based priorities for clinical practice, policy and research.DesignThe mixed-methods evidence synthesis comprised three components: (1) a systematic review to determine the clinical effectiveness, cost-effectiveness and acceptability of holistic breathlessness services; (2) a secondary analysis of pooled individual data from three trials to determine predictors of clinical response; and (3) a transparent expert consultation (TEC), comprising a stakeholder workshop and an online consensus survey, to identify stakeholders’ priorities.ResultsThirty-seven papers reporting on 18 holistic breathlessness services were included in the systematic review. Most studies enrolled people with thoracic cancer, were delivered over 4–6 weeks, and included breathing training, relaxation techniques and psychological support. Meta-analysis demonstrated significant reductions in the Numeric Rating Scale (NRS) distress due to breathlessness, significant reductions in the Hospital Anxiety and Depressions Scale (HADS) depression scores, and non-significant reductions in the Chronic Respiratory Disease Questionnaire (CRQ) mastery and HADS anxiety, favouring the intervention. Recipients valued education, self-management interventions, and expertise of the staff in breathlessness and person-centred care. Evidence for cost-effectiveness was limited and inconclusive. The responder analysis (n = 259) revealed baseline CRQ mastery and NRS distress to be strong predictors of the response to breathlessness services assessed by these same measures, and no significant influence from baseline breathlessness intensity, patient diagnosis, lung function, health status, anxiety or depression. The TEC elicited 34 priorities from stakeholders. Seven priorities received high agreement and consensus, reflecting stakeholders’ (n = 74) views that services should be person-centred and multiprofessional, share their breathlessness management skills with others, and recognise the roles and support needs of informal carers.LimitationsThe evidence synthesis draws predominantly from UK services and may not be generalisable to other settings. Some meta-analyses were restricted by reporting biases and statistical heterogeneity.ConclusionsDespite heterogeneity in composition and delivery, holistic breathlessness services are highly valued by recipients and can lead to significant improvements in the distress caused by breathlessness and depression. Outcomes of improved mastery and reduced distress caused by breathlessness are not influenced by patient diagnosis, lung function or health status. Stakeholders highlighted the need for improved access to person-centred, multiprofessional breathlessness services and support for informal carers.Future workOur research suggests that key therapeutic components of holistic breathlessness services be considered in clinical practice and models of delivery and educational strategies to address stakeholders’ priorities tested.Study registrationThis study is registered as PROSPERO CRD42017057508.FundingThe National Institute for Health Research (NIHR) Health Services and Delivery programme. Matthew Maddocks, Wei Gao and Irene J Higginson are supported by the NIHR Collaboration for Leadership in Applied Health Research and Care (CLAHRC) South London; Matthew Maddocks is supported by a NIHR Career Development Fellowship (CDF-2017-009), William D-C Man is supported by the NIHR CLAHRC Northwest London and Irene J Higginson holds a NIHR Emeritus Senior Investigator Award.

BackgroundChronic obstructive pulmonary disease (COPD) is the commonest respiratory disease in the UK, accounting for 10% of emergency hospital admissions annually. Nearly one-third of patients are re-admitted within 28 days of discharge.ObjectivesThe study aimed to evaluate the effectiveness of introducing standardised packages of care (i.e. care bundles) as a means of improving hospital care and reducing re-admissions for COPD.DesignA mixed-methods evaluation with a controlled before-and-after design.ParticipantsAdults admitted to hospital with an acute exacerbation of COPD in England and Wales.InterventionCOPD care bundles.Main outcome measuresThe primary outcome was re-admission to hospital within 28 days of discharge. The study investigated secondary outcomes including length of stay, total number of bed-days, in-hospital mortality, 90-day mortality, context, process and costs of care, and staff, patient and carer experience.Data sourcesRoutine NHS data, including numbers of COPD admissions and re-admissions, in-hospital mortality and length of stay data, were provided by 31 sites for 12 months before and after the intervention roll-out. Detailed pseudo-anonymised data on care during admission were collected from a subset of 14 sites, in addition to information about delivery of individual components of care collected from random samples of medical records at each location. Six case study sites provided data from interviews, observation and documentary review to explore implementation, engagement and perceived impact on delivery of care.ResultsThere is no evidence that care bundles reduced 28-day re-admission rates for COPD. All-cause re-admission rates, in-hospital mortality, length of stay, total number of bed-days, and re-admission and mortality rates in the 90 days following discharge were similar at implementation and comparator sites, as were resource utilisation, NHS secondary care costs and cost-effectiveness of care. However, the rate of emergency department (ED) attendances decreased more in implementation sites than in comparator sites {implementation: incidence rate ratio (IRR) 0.63 [95% confidence interval (CI) 0.56 to 0.70]; comparator: IRR 1.14 (95% CI 1.04 to 1.26) interactionp < 0.001}. Admission bundles appear to be more complex to implement than discharge bundles, with 3.7% of comparator patients receiving all five admission bundle elements, compared with 7.6% of patients in implementation sites, and 28.3% of patients in implementation sites receiving all five discharge bundle elements, compared with 0.8% of patients in the comparator sites. Although patients and carers were unaware that care was bundled, staff view bundles positively, as they help to standardise working practices, support a clear care pathway for patients, facilitate communication between clinicians and identify post-discharge support.LimitationsThe observational nature of the study design means that secular trends and residual confounding cannot be discounted as potential sources of any observed between-site differences. The availability of data from some sites was suboptimal.ConclusionsCare bundles are valued by health-care professionals, but were challenging to implement and there was a blurring of the distinction between the implementation and comparator groups, which may have contributed to the lack of effect on re-admissions and mortality. Care bundles do appear to be associated with a reduced number of subsequent ED attendances, but care bundles are unlikely to be cost-effective for COPD.Future workA longitudinal study using implementation science methodology could provide more in-depth insights into the implementation of care bundles.Trial registrationCurrent Controlled Trials ISRCTN13022442.FundingThis project was funded by the National Institute for Health Research Health Services and Delivery Research programme and will be published in full inHealth Services and Delivery Research; Vol. 7, No. 21. See the NIHR Journals Library website for further project information.