Asthma and Chronic Obstructive Pulmonary Disease (COPD) are two of the most common chronic respiratory diseases causing a major burden of ill health to populations across the world. Respiratory medication prescribing can be used as an indicator of air pollution effect on asthma and COPD, capturing patients with any severity of disease from mild to severe. In contrast, the traditional indicators of asthma exacerbation, such as hospital admissions and emergency room visits, only capture events of patients who suffer severe symptoms. In this study, I aimed to develop statistical models for assessing the spatiotemporal patterns of salbutamol prescribing in relation to air quality, in a primary health care setting. Salbutamol represents 93% of short-acting β2-agonists, which are prescribed for quick-relief of symptoms and acute exacerbations to individuals that suffer from asthma or COPD. I analysed salbutamol medication (approximately 67 billion Average-Daily-Quantities) prescribed by 64 GP practices in Newcastle and North Tyneside Primary Care Trusts, Northeast England, in 2002-2006. I used a mixed-effects model suitable for data that are not independent in time or space. My study found ambient Particulate Matter (PM10) concentrations to have a significant relationship to salbutamol prescribing in primary care. An increase of 10μg/m3 in ambient PM10 concentrations was associated with an increase of 1% in salbutamol prescribing. Income deprivation and average age of patients registered per practice also had a significant relationship with salbutamol prescribing. The findings showed that the variation of salbutamol prescribing was subject not only to health needs caused by deprivation and air quality, but also random effects that were practice specific, such as facilities within the practice or experience and prescribing pattern of practitioners. Overall, the findings demonstrated that respiratory prescribing in primary care can be used as indicator of air pollution effect on asthma and COPD, increasing the scope of its use for health surveillance in the future.

Title: Daily air pollution levels and asthma; exploring the influence of time-activity patterns on personal exposure in Al Jubail industrial city, Saudi Arabia Introduction: Air pollution is a known risk factor for adverse cardio-respiratory health effects. In the last few years, epidemiological studies have provided evidence that exposure to air pollution can aggravate symptoms in asthmatic patients. Some epidemiological studies have used ambient air pollution levels based on fixed-site monitoring (FSM) data to evaluate the short-term effects of ambient air pollution levels on asthma-related emergency department visits (AEDv) using time-series analysis. In the recent past, technology has greatly improved, making it possible to carry out personal monitoring of indoor and outdoor microenvironments (ME). While the existing literature on time-activity patterns (TAP) and ME exposures for populations in the USA and Europe keeps on growing, little research on this topic has been carried out in the Middle East. This study was designed to (i) investigate the statistical association between exposure to air pollution and AEDv, and (ii) identify factors that influence personal exposure in different ME in Jubail Industrial City, Saudi Arabia. Methods: Daily number of AEDv, air pollution levels (particulate matter (PM2.5 and PM10), sulphur dioxide (SO2), carbon monoxide (CO) nitrogen oxides (NOx)) and weather variables (temperature and relative humidity) were obtained from the Royal Commission of Al Jubail Industrial City for the period between 2007 and 2011. Data were analysed using a time series approach, which involved application of a generalised linear model (GLM). Relative risks (RRs) were estimated using Poisson regression, while controlling weather variables, day of the week and holiday indicator for lag times of 0 - 7 days. RRs and 95% confidence intervals (CIs) in AEDv were calculated with each increment of inter-quartile range (IQR) change of each pollutant. Furthermore, to explore the influence of different ME on personal exposure levels, 27 students aged between 16-18 years were recruited and asked to record their detailed movements using a time-activity diary at 15-minute intervals over a period of 24 hours. The students were asked to carry a small backpack containing ii a personal air monitor to measure their personal exposure to PM2.5, and a GPS device to help identify ME including travelling, outdoors, at school and at home. Results: The association between AEDv and change in the quantity of SO2, NO2, PM2.5 and PM10 remained positive and statistically significantly after adjustment in the multi-pollutants model. The RR (in percent %) of AEDv increased by 5.4% (95% CI: 2.4, 8.5) at lag 2 for SO2 , 3.4% (95% CI: 0.8, 6.1) at lag 3 for NO2, 4.4% (95% CI: 2.4, 6.6) for PM2.5 and 2.2% (95% CI: 1.3, 3.2) for PM10 at lag 0 per IQR change in pollutants, 2.0ppb, 7.6ppb, 36μg/m3 and 140μg/m3 respectively. No significant associations between AEDv and CO were found. The time activity diary revealed that most of the students’ time was spent indoors (88.6%). The FSM median (IQR) PM2.5 level, 51.0μg/m3 (34.0-74.2), was significantly higher than personal median PM2.5 level, which was 30.0μg/m3 (20.9-42.4) (Wilcoxon p-value<0.001). Total personal outdoor median concentration of PM2.5, 44.4μg/m3 (31.1-59.5), was significantly higher than total personal indoor concentration, which was 28.3μg/m3 (19.2-40.2) (p-value<0.001). There was a significant but weak correlation between FSM and personal monitor PM2.5 levels when indoors (Spearman’s rank correlation=0.228, p-value < 0.001 n=544), but not at outdoors microenvironment. Conclusion: Current levels of ambient air pollution were associated with AEDv in Al Jubail. While there appear to be similarities between TAP in this small population sample from the Middle East and Europe/USA, the exposure levels in this industrial city appear to be very high, compared to the WHO air quality guidelines. The validity of FSM data as a proxy for personal exposure to PM needs to be characterised so that the exposure error associated with this proxy measure is better understood.

Background Community-acquired pneumonia (CAP) is one of the commonest causes of adult hospitalisation in sub-Saharan Africa, but recent data describing its epidemiology, microbial aetiology and outcome are limited. Focusing particularly on Malawi, the overall aim of this thesis was to describe the aetiology and outcome of CAP in sub-Saharan African to determine the key predictors of mortality. Methods Firstly, a systematic review of studies of CAP in adults in sub-Saharan Africa was performed to describe CAP aetiology, estimate the mortality rate and identify risk factors associated with death. Secondly, a prospective observational study of adults hospitalised with clinically diagnosed CAP to Queen Elizabeth Central Hospital, Blantyre, Malawi was completed to describe microbial aetiology using modern diagnostic modalities, determine outcome and identify prognostics factors. Thirdly, having identified in preliminary analyses of the prospective cohort that hypoxaemia was an independent risk factor for mortality, a study of the effectiveness of supplemental oxygen delivery by oxygen concentrator to correct hypoxaemia in adults with suspected CAP was performed. Results In both the systematic review and the prospective cohort the predominant burden of hospitalised CAP was in young (average age 38 and 35, respectively) and HIV-positive (52% and 78%) patients with limited chronic cardiovascular and pulmonary comorbidity. Streptococcus pneumoniae (27% and 21%) and Mycobacterium tuberculosis (19% and 23%) were the most commonly identified causes. The overall mortality rate for hospitalised patients in the systematic review was 9.5%, but data describing prognostic factors were limited. In the prospective cohort (n=459), death by day 30 occurred in 14.6% and was associated with: male sex (aOR 2.57); pre-presentation symptom duration (aOR 1.11 per day increase); inability to stand (aOR 4.28); heart rate (aOR 1.02 per beat/minute rise); oxygen saturations (aOR 0.95 per % rise); white cell count (aOR 0.91 per 109/L rise); haemoglobin (aOR 0.90 per g/dL rise). A newly derived four parameter mortality risk prediction tool based on male sex, oxygen saturations < 90%, inability to stand and heart rate ≥125 /min predicted 30-day mortality with reasonable accuracy (area under the receiver-operating curve (AUROC) 0.79) whilst existing tools performed poorly (CURB65: AUROC 0.60; SMRT-CO: AUROC 0.66). Hypoxaemia was corrected in 86.4% (n=59) of adults with suspected CAP with supplemental oxygen at standard flow-rate of 5 litres/minute. Failure to attain normoxaemia was associated with a more than four-fold increase in the risk of death (RR 4.25). Conclusions The major burden of hospitalised CAP in low-resource, sub-Saharan African settings is seen in young and HIV-positive adults, many of whom have TB. Extrapolating CAP assessment and treatment algorithms from well-resourced settings where the epidemiology and aetiology of disease is very different is flawed. If validated, locally derived severity assessment tools may provide a rational basis on which to stratify CAP management. Strategies to increase early detection and treatment of TB and to improve supportive care, in particular the correction of hypoxaemia, hold considerable promise for improving CAP outcomes and should be evaluated in clinical trials.

Seasonal influenza is an acute, contagious respiratory infection that causes considerable morbidity and mortality each year. The Chief Medical Officer for England recommends that healthcare workers have a seasonal influenza vaccination in an attempt to protect both patients and NHS staff. Despite current recommendations and campaigns targeted at improving vaccination uptake, many healthcare workers do not have a seasonal influenza vaccination. It is clear that more research is necessary to fully understand the vaccination decision of healthcare workers. This mixed methods thesis employed a range of novel methodological approaches to understanding the influences on the seasonal influenza vaccination decision by medical students and junior doctors. Social network analysis is a well-established research approach that looks at individuals in the context of their social connections. I used an outbreak simulation model to investigate to assess whether an individual’s risk of infection could be linked with their position in the social network. Expanding further on this, the auto-logistic regression model was applied to social network data to predict an individual’s likelihood of vaccinating given the behaviour of their peers. Finally, a qualitative approach was used to explore the factors informing vaccination decisions. Findings gathered throughout this programme of work were synthesised together to produce a more detailed evaluation of seasonal influenza vaccination amongst medical students and junior doctors. These have been disseminated widely, particularly to occupational health practitioners and the wider academic community – demonstrating that this public health research has impact in practice. By gaining a better understanding of the social effects on influenza vaccination it will be possible to improve seasonal influenza vaccination uptake by healthcare workers, in turn better protecting patients and staff.

The aim of this thesis was to understand the role of epistemic communities in the global response to Severe Acute Respiratory Syndrome (SARS), through the theoretical lens of Social Constructivism, in order to explain the extent to which ideational factors and normative power constructed through them contributed to the shaping of Global Health Governance (GHG) surrounding infectious disease outbreaks. The methodology of this thesis combined key informants interviews with archival document‐based research. Archival research consisted of gathering and reviewing government documents, publications of intergovernmental organizations, media reports, press briefings, and policy papers. Discourse analysis was employed to systematically examine the wide range of data gathered. The thesis explored how different discourses have driven the changes in public health reasoning and practice, in the form of prioritisation of certain actions in the global response to SARS at various points in time. In order to show the ideational shifts over time, the SARS story was divided into three key phases in terms of the progress of the outbreak. This thesis finds that the global response to the SARS outbreak over three phases was shaped by contestation among various discourses, which framed the perceived priority issues and policy responses pursued. These ideas did not simply arise as governing norms, but ideational success occurred as a result of collective advancement by actors who were coalesced around particular policy ideas. The thesis provided an account of the interplay between policy ideas and key actors, in the form of epistemic communities, and how epistemic communities served as key sources of advice to policy making during the SARS outbreak. The thesis demonstrates that in many ways, the GHG of SARS mirrored, the GHG of other global health issues in terms of framing of issues and the actors in the formation of and justification for interests in global health. By illustrating the origins and significance of the multiplicity of ideas shaping collective action on SARS, this thesis underscores that governance response in terms of policy outcomes is the product of reconciling health with a plethora of competing priorities, and political economic goals via social construction of reality. The thesis considered the implications of the findings for conceptual understanding of GHG of infectious diseases, and for strengthening policies and practices to address the global infectious disease outbreaks.

Asthma is a disease characterised by airway inflammation, which is predominantly eosinophilic. Recent developments in the technique of sputum induction have provided a safe non-invasive method of measuring airway inflammation that can be applied to wide populations of patients with asthma and other airway diseases. This thesis explores the use of induced sputum to measure airway inflammation in the assessment and management of adults with asthma. It provides the first evidence that the use of this technique in the management of asthma leads to improved patient outcomes. I have employed induced sputum to assess lower airway inflammation in a large population of patients with symptomatic mild to moderate asthma and have demonstrated considerable heterogeneity of the inflammatory response. I have identified a population of patients with isolated neutrophilic airway inflammation and have provided evidence that such patients respond poorly to inhaled corticosteroid treatment. I have described a management strategy directed at normalising the sputum eosinophil count, as well as controlling symptoms and peak flow readings. I have shown that this management strategy leads to a dramatic reduction in severe asthma exacerbations and prevents hospital admissions compared to a traditional clinical approach and that eosinophilic inflammation is an important risk factor for severe asthma exacerbations. Finally, I report that amongst patients with asthma who remain symptomatic despite low dose inhaled corticosteroids, high dose inhaled corticosteroids and long acting -agonists have contrasting effects on symptoms, lung function and the sputum eosinophil count, suggesting that there is a dissociation between eosinophilic airway inflammation, day-to-day symptoms and variable airflow obstruction in asthma. These findings suggest that the regular monitoring of airway inflammation may be required for optimal asthma management.

The aims of this study were to explore the lived experience of people with COPD and the views of healthcare professionals involved in the care of patients with COPD. The research question asked how health providers are meeting the psychosocial needs of people with COPD. Recent literature suggests that some patients with COPD are leaving primary care consultations with unmet psychosocial needs and that healthcare providers report being unwilling to promote behaviour change as they perceive it could damage their ongoing relationships with their patients. Data were collected from semi-structured interviews carried out with nine people with COPD and ten healthcare professionals (HCPs). The interviews were transcribed verbatim and the data analysed using Attride-Stirling’s (2001) model of thematic network analysis. Examples of themes deduced from the findings of the COPD group were loss and lifespan health. Those from the HCP group included attitudes and patient care. Both groups yielded a global theme of individuality. Findings from the study suggest that COPD patients are happy with their experience of healthcare although they valued prompt, accessible care in an emergency more highly than routine review appointments. They did not indicate that their psychosocial needs were met in routine consultations although they reported that some of these needs were met during pulmonary rehabilitation. HCPs perceived that they provided good care but that there were barriers to introducing psychosocial issues into routine appointments. In a time of change in patient demographics resulting in an increased number of older people with long-term conditions, this study adds to the body of knowledge in this field by exploring the lived experience of both people with COPD and of HCPs. The global theme of individuality for each group supports the need for person-centred care in the healthcare system in order to meet individuals’ psychosocial needs. Recommendations include; increased provision of pulmonary rehabilitation courses and training for HCPs in order for them to be more aware of the psychosocial needs of patients attending routine appointments.

INTRODUCTION: Asthma is a chronic inflammatory disorder of the airways punctuated by acute exacerbations. Coagulation and fibrinolysis pathways are increasingly recognised in the asthmatic inflammatory milieu and markers of fibrin turnover represent potential asthma biomarkers targets. A biomarker in plasma or urine predicting exacerbation could enable early intervention to attenuate or prevent exacerbation. STUDY DESIGN: A study in 2 parts comparing exacerbation and stable asthma was performed. Part 1 compared adults with exacerbations requiring hospital treatment to clinical recovery. Part 2 prospectively followed a cohort of adult moderate to severe asthmatics when well until exacerbation and through to clinical recovery. METHODS: Plasma was analysed by ELISA for markers of fibrin turnover alongside inflammatory and cellular markers, and comparison was made between exacerbation and recovery. Urine was analysed by latex agglutination for fibrin(ogen) degradation products (FDPs). Symptoms were recorded and a score was derived to distinguish asthma exacerbation from baseline and recovery states. Comparison was made between exacerbation and recovery using multivariate statistical analysis. KEY FINDINGS: Prospectively, the detection of FDP in urine significantly increased the likelihood of subsequent asthma exacerbation in the following 7-14 days between 4 and 6 times that of a negative urine FDP test. FDP +7 days (OR 4.35 (95% CI 1.12, 16.9) p=0.03), FDP +10 days (OR 5.50 (95% CI 1.38, 21.9) p=0.02), FDP +14 days (OR 6.07 (95% CI 1.43, 25.8) p=0.02). Plasma D-dimer was significantly increased from baseline to exacerbation (median 0.07 μgFEU/ml (95% CI -0.01, 0.15) p=0.02). Plasma eosinophils (median 10*9/l (95% CI 0.10, 0.29) p=0.02) and eotaxin (median 12 pg/ml (95% CI 3, 22) p<0.001) increased significantly from exacerbation to recovery, and PF4 significantly increased from exacerbation to follow up (median 9 μg/ml (95 % CI 2, 17), p=0.002). Positive correlations were observed at exacerbation between eosinophils and eotaxin (p<0.001), platelets and PF4 (p=0.02) and platelets and eosinophils (p=0.01). Clinical symptoms and reliever use increased in the 10 days preceding exacerbation and reduced during recovery (p<0.001). CONCLUSIONS: This is the largest study to date examining coagulation and fibrinolysis in acute exacerbation of asthma. Detection of urine FDP in moderate to severe asthma suggests up to 6-fold increased exacerbation risk in the subsequent 14 days. The potential to translate these findings into a point-of-care test for use by patients at home warrants further study. Positive correlations between eosinophils, eotaxin, platelets and platelet factor 4 (PF4) alongside increased PF4, eosinophils and eotaxin in plasma after a period of recovery and treatment support eosinophil and platelet migration into the lung during asthma exacerbation. The symptom score used in this study provides a simple objective measure of asthma exacerbation for use in future research or clinical practice but requires prospective validation.

Health monitoring devices are changing the way we treat, manage and understand chronic health conditions by continuously collecting data about medication use, symptoms, vital signs and a patient’s surrounding environment. Asthma is the most common chronic condition worldwide and has symptoms that include wheezing, coughing and shortness of breath - all typically treated with inhaled medication. By taking a daily ‘preventer’ inhaler a patient should be able to control their asthma and remain relatively symptom free. However, adherence to preventer medication is often poor, making patients prone to severe symptoms and asthma attacks. This leads to avoidable healthcare costs, mainly through preventable hospitalisations and wasted medication. Electronic monitoring devices (EMDs) are the most accurate method currently available for recording inhaler use. Through rigorous testing they have been shown to be sufficiently accurate and reliable for use in clinical practise. Early signs also suggest that they may improve inhaler use when a ringtone is used to remind the patient when a new dose is due. However, no research to date has considered the attitudes of patients with asthma as well as healthcare providers towards these devices. Human factors in healthcare is a now established area of research, with an international standard (ISO 62366) meaning that medical device developers are required to design for a usable and error-free experience. This creates clear scope for assessing the perceptions of patients and healthcare providers towards EMDs for use in the management of asthma. This was investigated in this thesis using three separate but related research studies. Two studies assessed the attitudes of patients and healthy volunteers, whilst the other analysed the opinions of healthcare providers and stakeholders in asthma care. The first study assessed the attitudes of a sample of adolescents with asthma towards an exemplar EMD – the SmartTrack. Asthma is most prevalent in adolescents, adherence is notoriously poor, and they are often overlooked in medical device design - making assessing their views a priority. The participants used a SmartTrack device for one month and completed questionnaires and interviews on their opinions towards important issues including being monitored and the device’s appearance, social acceptability and practicality. For the second study, a Delphi survey method was used to collect the opinions of stakeholders in asthma care including respiratory consultants, nurses and GPs towards EMDs. They were asked to state the key benefits they thought these devices could have, as well as the key potential costs and barriers for their introduction. Additional rounds of surveys were then used to assess which points they felt were most important and should be prioritised in the future development of these devices. EMDs developed new capabilities over the course of this thesis, meaning that the third study was used to investigate attitudes towards location and activity data, as both were beginning to be integrated with data on inhaler use. Two workshops and a technology trial were conducted with a sample of healthy adolescents from a local sixth form. Participants had their location and activity tracked and then had this data presented back to them. They were asked for their opinions on the usefulness of this data, as well as any potential risks associated with recording it. The findings from all three studies were then brought together to determine the requirements of EMDs going forward. A systems model of asthma care was developed to firstly consider the routes through which EMDs could be purchased, as well as the impacts EMDs could have on various points of the asthma care process, both for the patient and the healthcare provider. The requirements for EMDs were then also related back to this model, to help outline their importance and relevance to the asthma care pathway. From the research it was determined that developers need to reduce the size of EMDs, as well as integrate the ability to monitor inhalation and technique. Commissioners need to work with clinical researchers to identify the types of patients who would benefit most from an EMD, in order to reduce risks of over- purchasing. Furthermore, researchers need to work with healthcare standards bodies to establish how the vast quantities of data produced by EMDs can be successfully integrated into the clinical care process. This thesis has three key contributions. Firstly, it introduces human factors research methods to the domain of EMDs for asthma care. Secondly, it provides a set of requirements for EMDs to help ensure that these devices can be successfully introduced and used in clinical care. Lastly, it supplements the literature on human factors methods being applied to healthcare and provides a new example of where user-focused research has been used to elicit requirements for a medical device.

Background: Poor adherence to inhaled corticosteroids (ICS) is known as the main cause for therapeutic failure in asthma treatment and associated morbidity. Adherence is complex and can have many causes, which will vary between conditions, treatments and patients. To improve adherence, it is vital to understand what effects a patients adherence, so appropriate interventions can be developed and targeted, both for the patients who would benefit most and at the most important points in treatment. Very few studies have characterised the variables associated with poor adherence and how these differences may change over time, and the most appropriate methodology for investigating this relations have not previously been defined. Aims and objectives: The aim of the PhD study was to investigate what characteristics associated with a patient’s adherence to ICS, and to investigate whether these relationships change over time using a large primary care dataset. The objectives included the development of a longitudinal measure of asthma patients’ adherence to ICS, then to investigate the time dependent relationship between adherence and other available patient variables by trialling a number of different methods. In addition, the effect of adherence on clinical outcome in asthma was tested, since counter intuitively many studies have not previously found a clear relationship between the variables. Methods: A retrospective longitudinal study using a large cohort was conducted using primary care data from the Clinical Practice Research Datalink (with Hospital Episodes Statistics data) between 1997 and 2010. Asthma patients aged between 12 and 65 years, without a diagnosis of chronic obstructive pulmonary disease were included in the study cohort. ICS prescriptions were extracted and used to calculate the annual prescription possession ratio (PPR). Several definitions of the PPR measure were tested to develop a proxy measure to represent adherence. Variables related to clinical outcomes and other characteristics were also identified for each patient in the cohort. A two-way analysis was conducted to compare the relationship between adherence and each patient variable with time, and then four methods were used to further investigate the relationship between adherence and patient exacerbations including; (1) comparing adherence in the year before and after an exacerbation; (2) descriptively exploring the clinical outcomes associated with different adherence levels; (3) identifying the relative risk of an exacerbation associated with adherence defined by different cut off levels of PPR; and (4) descriptively exploring the effect of adherence on outcome and outcome on adherence over time. Finally, the available variables associated with adherence (including previous adherence and clinical outcomes) were analysed in a dynamic panel model to understand explore the effect of variables on patients’ adherence to ICS which allows for the feedback effects of previous adherence and clinical outcome and the effect of time on adherence. Results: Many patient variables were found to effect adherence. When modelling the effect of patient variables on adherence, adherence was found to be lower in younger patients (+0.11%/year), patients with fewer years in the study (+0.25%/year), with more severe asthma (step 5 patients had a 3.32% lower PPR than step 2 patients), with good control (5.21% lower), with lower previous adherence (-0.51% per % PPR), and who had not previously experienced an exacerbation (0.87% lower compared with patients who had experienced no primary car exacerbation and 1.45% lower for those who experienced no secondary care exacerbation). Adherence increased with patient year, consistently across most subgroups, with the following exceptions; the 20-25 year old age group had lower initial adherence (53.9%) than the younger patients (58.3%), patients registered in the East Midlands had the lowest adherence (57.7%), but increased over time to become the highest (90.7%) and in the first year of the study the adherence for patients treated at step 2 of the guidelines was the lowest (57.5%) but it increased over time to become the second highest step (85.7%). Conclusion: This longitudinal follow-up study using electronic patient records over time was useful to identify the effect of multiple patient variables on adherence. The main characteristics associated with poor adherence were the characteristics that we would associate with better health, or less severe asthma. Therefore, the interventions to improve adherence or to review the appropriateness of treatment should be developed to target younger patients, early on in treatment before they have experienced an exacerbation of their asthma symptoms. The PPR measure developed was useful to measure changes in adherence over time, as a measure of the maximum amount of medicine that the patient had available to them, expressed as a percentage of their recorded prescribed dose. However there are some important limitations to the PPR measure including most importantly that adherence must be measured against a routinely prescribed daily dose of ICS and medicine prescribing and not medicine taking is measured, meaning that adherence is likely to be overestimated. The methods used to analyse the adherence measure had not previously been used to assess adherence in asthma. By using the results from each analysis method, information about different parts of the relationship between adherence and other patient variables including their exacerbation risk and time could be combined, which uniquely allowed the longitudinal measurement and analysis of adherence in asthma over extended study duration.