Using stochastic simulation modelling to study occupancy levels of decentralised admission avoidance units in Norway.

Identifying alternatives to acute hospital admission is a priority for many countries. Over 200 decentralised municipal acute units (MAUs) were established in Norway to divert low-acuity patients away from hospitals. MAUs have faced criticism for low mean occupancy and not relieving pressures on hospitals. We developed a discrete time simulation model of admissions and discharges to MAUs to test scenarios for increasing absolute mean occupancy. We also used the model to estimate the number of patients turned away as historical data was unavailable. Our experiments suggest that mergers alone are unlikely to substantially increase MAU absolute mean occupancy as unmet demand is generally low. However, merging MAUs offers scope for up to 20% reduction in bed capacity, without affecting service provision. Our work has relevance for other admissions avoidance units and provides a method for estimating unconstrained demand for beds in the absence of historical data.

Machine learning for real-time aggregated prediction of hospital admission for emergency patients.

Machine learning for hospital operations is under-studied. We present a prediction pipeline that uses live electronic health-records for patients in a UK teaching hospital's emergency department (ED) to generate short-term, probabilistic forecasts of emergency admissions. A set of XGBoost classifiers applied to 109,465 ED visits yielded AUROCs from 0.82 to 0.90 depending on elapsed visit-time at the point of prediction. Patient-level probabilities of admission were aggregated to forecast the number of admissions among current ED patients and, incorporating patients yet to arrive, total emergency admissions within specified time-windows. The pipeline gave a mean absolute error (MAE) of 4.0 admissions (mean percentage error of 17%) versus 6.5 (32%) for a benchmark metric. Models developed with 104,504 later visits during the Covid-19 pandemic gave AUROCs of 0.68-0.90 and MAE of 4.2 (30%) versus a 4.9 (33%) benchmark. We discuss how we surmounted challenges of designing and implementing models for real-time use, including temporal framing, data preparation, and changing operational conditions.

Operational analysis of school-based delivery models to vaccinate children against influenza.

Large-scale immunisation programmes against seasonal influenza are characterised by logistical challenges related to the need for vaccinating large cohorts of people in a short amount of time. Careful operational planning of resources is essential for a successful implementation of such programmes. We focused on the process of child vaccination in schools and analysed the staffing and workflow aspects of a school-aged children vaccination programme in England. Our objectives were to document vaccination processes and analyse times and costs associated with different models deployed across England. We collected data through direct non-participatory observations. Statistical data analysis enabled us to identify potential factors influencing vaccine delivery time and informed the development of a tool to simulate vaccination sessions. Using this tool, we carried out scenario analyses and explored trade-offs between session times and costs in different settings. Our work ultimately supported the local implementation of school-based vaccination.

Exploring the role of mass immunisation in influenza pandemic preparedness: A modelling study for the UK context.

The nature and timing of the next influenza pandemic is unknown. This makes it difficult for policy makers to assess whether spending money now to prepare for mass immunisation in the event of a pandemic is worthwhile. We used simple epidemiological modelling and health economic analysis to identify the range of pandemic and policy scenarios under which plans to immunise the general UK population would have net benefit if a stockpiled vaccine or, alternatively, a responsively purchased vaccine were used. Each scenario we studied comprised a combination of pandemic, vaccine and immunisation programme characteristics in presence or absence of access to effective antivirals, with the chance of there being a pandemic each year fixed. Monetarised health benefits and cost savings from any influenza cases averted were set against the option, purchase, storage, distribution, administration, and disposal costs relevant for each scenario to give a discounted net present value over 10 years for planning to immunise, accounting for the possibility that there may be no pandemic over the period considered. To support understanding and exploration of model output, an interactive visualisation tool was devised and made available online. We evaluated over 29 million combinations of pandemic and policy characteristics. Preparedness plans incorporating mass immunisation show positive net present value for a wide range of scenarios, predominantly in the absence of effective antivirals. Plans based on the responsive purchase of vaccine have wider benefit than plans reliant on the purchase and maintenance of a stockpile if immunisation can start without extensive delays. This finding is not dependent on responsively purchased vaccine being more effective than stockpiled vaccine, but rather is driven by avoiding the costs of storing and replenishing a stockpile.

A method for evaluating the cost-benefit of different preparedness planning policies against pandemic influenza.

•Our work presents a unifying method to calculate the net-benefit of different preparedness policies against different pandemic influeunza strains. Unlike previous methods, which have focused on evaluating specific strategies against specific pandemics, our method allows assessment of mass immunisation strategies in presence and absence of antiviral drugs for a large range of pandemic influenza strain characteristics and programme features. Overall, the model described here combines two parts to evaluate different preparedness planning policies against pandemic influenza.•The first part is adaptation of an existing transmission model for seasonal influenza to include generalisation across large number of pandemic influenza scenarios.•The second part is development of a tailor-made health economic model devised in collaboration with colleagues at the UK Department of Health and Social Care.

Using visualisation methods to analyse referral networks within community health care among patients aged 65 years and over.

Community health care services are considered integral to overcoming future problems in health care. However, this sector faces its own challenges, such as how to organise services to provide coordinated care given: their physical distribution, patients using multiple services, increased patient use and differing patient needs. The aim of this work was to explore, analyse and understand patterns in community referrals for patients aged 65 years and over, and their use of multiple services through data visualisation. Working with a large community provider, these methods helped researchers and service managers to investigate questions that were otherwise difficult to answer from raw data. Each map focuses on a different characteristic of community referrals: patients reusing services, concurrent uses of different services and patterns of subsequent referrals. We apply these methods to routine patient data and discuss their implications in designing of a single point of access - a service for streamlining referrals.

A tool for routine monitoring and feedback of morbidities following paediatric cardiac surgery.

Short-term survival after paediatric cardiac surgery has improved significantly over the past 20 years and increasing attention is being given to measuring and reducing incidence of morbidities following surgery. How to best use routinely collected data to share morbidity information constitutes a challenge for clinical teams interested in analysing their outcomes for quality improvement. We aimed to develop a tool facilitating this process in the context of monitoring morbidities following paediatric cardiac surgery, as part of a prospective multi-centre research study in the United Kingdom.We developed a prototype software tool to analyse and present data about morbidities associated with cardiac surgery in children. We used an iterative process, involving engagement with potential users, tool design and implementation, and feedback collection. Graphical data displays were based on the use of icons and graphs designed in collaboration with clinicians.Our tool enables automatic creation of graphical summaries, displayed as a Microsoft PowerPoint presentation, from a spreadsheet containing patient-level data about specified cardiac surgery morbidities. Data summaries include numbers/percentages of cases with morbidities reported, co-occurrences of different morbidities, and time series of each complication over a time window.Our work was characterised by a very high level of interaction with potential users of the tool, enabling us to promptly account for feedback and suggestions from clinicians and data managers. The United Kingdom centres involved in the project received the tool positively, and several expressed their interest in using it as part of their routine practice.

Erlang could have told you so-A case study of health policy without maths.

Little consideration is given to the operational reality of implementing national policy at local scale. Using a case study from Norway, we examine how simple mathematical models may offer powerful insights to policy makers when planning policies. Our case study refers to a national initiative requiring Norwegian municipalities to establish acute community beds (municipal acute units or MAUs) to avoid hospital admissions. We use Erlang loss queueing models to estimate the total number of MAU beds required nationally to achieve the original policy aim. We demonstrate the effect of unit size and patient demand on anticipated utilisation. The results of our model imply that both the average demand for beds and the current number of MAU beds would have to be increased by 34% to achieve the original policy goal of transferring 240 000 patient days to MAUs. Increasing average demand or bed capacity alone would be insufficient to reach the policy goal. Day-to-day variation and uncertainty in the numbers of patients arriving or leaving the system can profoundly affect health service delivery at the local level. Health policy makers need to account for these effects when estimating capacity implications of policy. We demonstrate how a simple, easily reproducible, mathematical model could assist policy makers in understanding the impact of national policy implemented at the local level.

Incidence and risk factors for important early morbidities associated with pediatric cardiac surgery in a UK population.

OBJECTIVE: Given excellent 30-day survival for pediatric cardiac surgery, other outcome measures are important. We aimed to study important early postoperative morbidities selected by stakeholders following a rigorous and evidenced-based process, with a view to identifying potential risk factors. METHODS: The incidence of selected morbidities was prospectively measured for 3090 consecutive pediatric cardiac surgical admissions in 5 UK centers between October 2015 and June 2017. The relationship between the candidate risk factors and the incidence of morbidities was explored using multiple regressions. Patient survival, a secondary outcome, was checked at 6 months. RESULTS: A total of 675 (21.8%) procedure episodes led to at least 1 of the following: acute neurologic event, unplanned reoperation, feeding problems, renal replacement therapy, major adverse events, extracorporeal life support, necrotizing enterocolitis, surgical infection, or prolonged pleural effusion. The highest adjusted odds ratio of morbidity was in neonates compared with children, 5.26 (95% confidence interval, 3.90-7.06), and complex heart diseases (eg, hypoplastic left heart), 2.14 (95% confidence interval, 1.41-3.24) compared with low complexity (eg, atrial septal defect, P < .001 for all). Patients with any selected morbidity had a 6-month survival of 88.2% (95% confidence interval, 85.4-90.6) compared with 99.3% (95% confidence interval, 98.9-99.6) with no defined morbidity (P < .001). CONCLUSIONS: Evaluation of postoperative morbidity provides important information over and above 30-day survival and should become a focus for audit and quality improvement. Our results have been used to initiate UK-based audit for 5 of these 9 morbidities, co-develop software for local monitoring of these morbidities, and parent information about these morbidities.

Exploring communication between parents and clinical teams following children's heart surgery: a survey in the UK.

OBJECTIVE: To explore communication between clinicians and families of children undergoing heart surgery. DESIGN: This study was part of a larger study to select, define and measure the incidence of postoperative complications in children undergoing heart surgery. Parents of children recruited to a substudy between October 2015 and December 2017 were asked to complete a questionnaire about communication during their child's inpatient stay. We explored all responses and then disaggregated by the following patient characteristics: presence of a complication, length of stay, hospital site, ethnicity and child's age. This was a descriptive study only. SETTING: Four UK specialist hospitals. RESULTS: We recruited 585 children to the substudy with 385 responses (response rate 66%).81% of parents reported that new members of staff always introduced themselves (18% sometimes, 1% no). Almost all parents said they were encouraged to be involved in decision-making, but often only to some extent (59% 'yes, definitely'; 37% 'to some extent'). Almost two-thirds of parents said they were told different things by different people which left them feeling confused (10% 'a lot'; 53% 'sometimes'). Two-thirds (66%) reported that staff were definitely aware of their child's medical history (31% 'to some extent'). 90% said the operation was definitely explained to them (9% 'to some extent') and 79% that they were definitely told what to do if they were worried after discharge (17% 'to some extent').Parents of children with a complication tended to give less positive responses for involvement in decision-making, consistent communication and staff awareness of their child's medical history. Parents whose children had longer stays in hospital tended to report lower levels of consistent communication and involvement in decision-making. CONCLUSIONS: Our results emphasise the need for consistent communication with families, particularly where complications arise or for children who have longer stays in the hospital.

Are we prepared for the next influenza pandemic? Lessons from modelling different preparedness policies against four pandemic scenarios.

In the event of a novel influenza strain that is markedly different to the current strains circulating in humans, the population have little/no immunity and infection spreads quickly causing a global pandemic. Over the past century, there have been four major influenza pandemics: the 1918 pandemic ("Spanish Flu"), the 1957-58 pandemic (the "Asian Flu"), the 1967-68 pandemic (the "Hong Kong Flu") and the 2009 pandemic (the "Swine flu"). To inform planning against future pandemics, this paper investigates how different is the net-present value of employing pre-purchase and responsive- purchased vaccine programmes in presence and absence of anti-viral drugs to scenarios that resemble these historic influenza pandemics. Using the existing literature and in discussions with policy decision makers in the UK, we first characterised the four past influenza pandemics by their transmissibility and infection-severity. For these combinations of parameters, we then projected the net-present value of employing pre-purchase vaccine (PPV) and responsive-purchase vaccine (RPV) programmes in presence and absence of anti-viral drugs. To differentiate between PPV and RPV policies, we changed the vaccine effectiveness value and the time to when the vaccine is first available. Our results are "heat-map" graphs displaying the benefits of different strategies in pandemic scenarios that resemble historic influenza pandemics. Our results suggest that immunisation with either PPV or RPV in presence of a stockpile of effective antiviral drugs, does not have positive net-present value for all of the pandemic scenarios considered. In contrast, in the absence of effective antivirals, both PPV and RPV policies have positive net-present value across all the pandemic scenarios. Moreover, in all considered circumstances, vaccination was most beneficial if started sufficiently early and covered sufficiently large number of people. When comparing the two vaccine programmes, the RPV policy allowed a longer timeframe and lower coverage to attain the same benefit as the PPV policy. Our findings suggest that responsive-purchase vaccination policy has a bigger window of positive net-present value when employed against each of the historic influenza pandemic strains but needs to be rapidly available to maximise benefit. This is important for future planning as it suggests that future preparedness policies may wish to consider utilising timely (i.e. responsive-purchased) vaccines against emerging influenza pandemics.

Addressing the challenges of knowledge co-production in quality improvement: learning from the implementation of the researcher-in-residence model.

The concept of knowledge co-production is used in health services research to describe partnerships (which can involve researchers, practitioners, managers, commissioners or service users) with the purpose of creating, sharing and negotiating different knowledge types used to make improvements in health services. Several knowledge co-production models have been proposed to date, some involving intermediary roles. This paper explores one such model, researchers-in-residence (also known as 'embedded researchers').In this model, researchers work inside healthcare organisations, operating as staff members while also maintaining an affiliation with academic institutions. As part of the local team, researchers negotiate the meaning and use of research-based knowledge to co-produce knowledge, which is sensitive to the local context. Even though this model is spreading and appears to have potential for using co-produced knowledge to make changes in practice, a number of challenges with its use are emerging. These include challenges experienced by the researchers in embedding themselves within the practice environment, preserving a clear focus within their host organisations and maintaining academic professional identity.In this paper, we provide an exploration of these challenges by examining three independent case studies implemented in the UK, each of which attempted to co-produce relevant research projects to improve the quality of care. We explore how these played out in practice and the strategies used by the researchers-in-residence to address them. In describing and analysing these strategies, we hope that participatory approaches to knowledge co-production can be used more effectively in the future.

A method for evaluating and comparing immunisation schedules that cover multiple diseases: Illustrative application to the UK routine childhood vaccine schedule.

BACKGROUND: In the UK, the childhood immunisation programme is given in the first 5 years of life and protects against 12 vaccine-preventable diseases. Recently, this programme has undergone changes with addition of vaccination against Meningitis B from September 2015 and the removal of the primary dose of protection against Meningitis C from July 2016. These hanges have direct impact on the associated diseases but in addition may induce indirect effects on the vaccines that are given simultaneously or later in the programme. In this work, we developed a novel formal method to evaluate the impact of vaccination changes to one aspect of the programme across an entire vaccine programme. METHODS: Firstly, we combined transmission modelling (for four diseases) and historic data synthesis (for eight diseases) to project, for each disease, the disease burden at different levels of effective coverage against the associated disease. Secondly, we used a simulation model to determine the vector of effective coverage against each disease under three variations of the current childhood schedule. Combining these, we calculated the vector of disease burden across the programme under different scenarios, and assessed the direct and indirect effects of the schedule changes. RESULTS: Through illustrative application of our novel framework to three scenarios of the current childhood immunisation programme in the UK, we demonstrated the feasibility of this unifying approach. For each disease in the programme, we successfully quantified the residual disease burden due to the change. For some diseases, the change was indirectly beneficial and reduced the burden, whereas for others the effect was adverse and the change increased the disease burden. CONCLUSIONS: Our results demonstrate the potential benefit of considering the programme-wide impact of changes to an immunisation schedule, and our framework is an important step in the development of a means for systematically doing so.

Parents' Experience and Views of Vaccinating Their Child against Influenza at Primary School and at the General Practice.

The purpose of this study was to gain an in-depth understanding of parents' experience and views of vaccinating their four to six-year-old child against influenza at school and at the general practice (GP). A cross-sectional qualitative study was conducted between March-June 2016 with parents of children in Reception and Year 1 in four randomly selected schools in Bury, Leicestershire, and Surrey, England. Twenty-five outreach forms were completed and returned, and seven interviews were conducted. Interview transcripts were coded by theme in NVivo (version 11, QSR International Pty Ltd., Melbourne, Australia). The primary reason parents gave for vaccinating their child was to prevent their child from contracting influenza. Parents' perceived benefits of vaccinating in schools were to avoid the inconvenience of having to take their child to the GP, and that their child would behave better at school. Parents viewed that accompanying their child for the vaccination at school would undermine the convenience and peer-pressure advantages of the school as a venue. No parents expressed concern about their child being too young to be vaccinated in school. This research suggests that the school is a desirable venue for childhood influenza vaccination, both from the parents' view and given that influenza vaccination coverage is higher when delivered through schools than GPs.

Analysing barriers to service improvement using a multi-level theory of innovation: the case of glaucoma outpatient clinics.

The development and implementation of innovation by healthcare providers is understood as a multi-determinant and multi-level process. Theories at different analytical levels (i.e. micro and organisational) are needed to capture the processes that influence innovation by providers. This article combines a micro theory of innovation, actor-network theory, with organisational level processes using the 'resource based view of the firm'. It examines the influence of, and interplay between, innovation-seeking teams (micro) and underlying organisational capabilities (meso) during innovation processes. We used ethnographic methods to study service innovations in relation to ophthalmology services run by a specialist English NHS Trust at multiple locations. Operational research techniques were used to support the ethnographic methods by mapping the care process in the existing and redesigned clinics. Deficiencies in organisational capabilities for supporting innovation were identified, including manager-clinician relations and organisation-wide resources. The article concludes that actor-network theory can be combined with the resource-based view to highlight the influence of organisational capabilities on the management of innovation. Equally, actor-network theory helps to address the lack of theory in the resource-based view on the micro practices of implementing change.

Interventional treatments and risk factors in patients born with hypoplastic left heart syndrome in England and Wales from 2000 to 2015.

OBJECTIVE: To describe the long-term outcomes, treatment pathways and risk factors for patients diagnosed with hypoplastic left heart syndrome (HLHS) in England and Wales. METHODS: The UK's national audit database captures every procedure undertaken for congenital heart disease and updated life status for resident patients in England and Wales. Patients with HLHS born between 2000 and 2015 were identified using codes from the International Paediatric and Congenital Cardiac Code. RESULTS: There were 976 patients with HLHS. Of these, 9.6% had a prepathway intervention, 89.5% underwent a traditional pathway of staged palliation and 6.4% of infants underwent a hybrid pathway. Patients undergoing prepathway procedures or the hybrid pathway were more complex, exhibiting higher rates of prematurity and acquired comorbidity. Prepathway intervention was associated with the highest in-hospital mortality (34.0%).44.6% of patients had an off-pathway procedure after their primary procedure, most frequently stenting or dilation of residual or recoarctation and most commonly occurring between stage 1 and stage 2.The survival rate at 1 year and 5 years was 60.7% (95% CI 57.5 to 63.7) and 56.3% (95% CI 53.0 to 59.5), respectively. Patients with an antenatal diagnosis (multivariable HR (MHR) 1.63 (95% CI 1.12 to 2.38)), low weight (<2.5 kg) (MHR 1.49 (95% CI 1.05 to 2.11)) or the presence of an acquired comorbidity (MHR 2.04 (95% CI 1.30 to 3.19)) were less likely to survive. CONCLUSION: Treatment pathways among patients with HLHS are complex and variable. It is essential that the long-term outcomes of conditions like HLHS that require serial interventions are studied to provide a fuller picture and to inform quality assurance and improvement.

Informing decisions on the purchase of equipment used by health services in response to incidents involving hazardous materials.

Accidents involving release of chemical, biological, radiological or nuclear substances may prompt the need to decontaminate exposed casualties prior to further medical treatment. Health service workers who carry out decontamination procedures wear protective suits to avoid direct contact with contaminants. We developed an analytical framework based on queueing theory to inform UK Department of Health's decisions on the stock of protective suits that ambulance services and hospitals with emergency departments in England should hold. Our aim was to ensure that such allocation gave an accepted degree of resilience to locally identified hazards. Here we give an overview of our work and describe how we incorporated information in the public domain about local hazards with expert opinion about the patterns of demand for decontamination associated with different types of incident. We also give an account of how we worked with decision makers to inform national guidance on this topic.

A systematic literature review of operational research methods for modelling patient flow and outcomes within community healthcare and other settings.

An ambition of healthcare policy has been to move more acute services into community settings. This systematic literature review presents analysis of published operational research methods for modelling patient flow within community healthcare, and for modelling the combination of patient flow and outcomes in all settings. Assessed for inclusion at three levels - with the references from included papers also assessed - 25 "Patient flow within community care", 23 "Patient flow and outcomes" papers and 5 papers within the intersection are included for review. Comparisons are made between each paper's setting, definition of states, factors considered to influence flow, output measures and implementation of results. Common complexities and characteristics of community service models are discussed with directions for future work suggested. We found that in developing patient flow models for community services that use outcomes, transplant waiting list may have transferable benefits.