It has been hypothesised that a trait-like vulnerability to sleep disruption exists. This has been demonstrated in response to physiological stressors such as caffeine and phase advance. From this work the Ford Insomnia Responsivity to Stress (FIRST) questionnaire was designed, which aims to specify those who are vulnerable to stress related sleep disruption. Further to this, neuroticism and emotion focused coping have been shown to characterise the insomnia population, and suggested that these constitute risk factors for the development of an insomnia syndrome. However, there has been very little work which aims to define an at-risk population and none which aims to characterise this population from both a physiological and psychological perspective. The aim of this thesis is to define the vulnerable population with regards to psychology and psychobiology. . It was hypothesised that the vulnerable group would show greater stress reactivity, physiologically, higher levels of neuroticism relative to the resilient group, lower levels of conscientiousness and a greater inclination toward rumination and worry. Over three studies measures of sleep, personality, stress perception and coping styles amongst others were taken as well as measuring, separately, 3 indices of physiological stress response: Cortisol output: Salivary free cortisol was taken whilst a sample of good sleepers completed the Trier Social Stress Test (TSST) (n=32). Results indicate that the vulnerable group show significantly greater levels of cortisol at base line (p < 0.05).This was mediated by conscientiousness (β=0.39). They were also higher in negative affect, rumination, stress and worry (p < 0.05).The vulnerable group also showed an increase in insomnia symptoms in response to real life stress. This was also related to conscientiousness (r= 0.55, p < 0.05) Cardiovascular response: Heart Rate (HR) and Cardiac Vagal Tone (CVT) were measured while participants (n=31) completed a relaxation (baseline) and stressful task. There was found to be a main effect of group on HR response to the stress task relative to baseline, but this did not maintain when psychological variables of interest were entered (n=31) into the model. Conscientiousness was related to lower CVT change, interpreted as lower CVT flexibility. Psychologically, the vulnerable group were again found to score higher on neuroticism, perceived stress and rumination relative to the resilient group (p < 0.05). Brain activation: fMRI data was collected whilst participants completed a stroop task, in which a siren indicated an increase in task difficulty (stress cue) (n=24). It was found that the vulnerable group showed significantly less activation bilaterally in the inferior parietal lobule (IPL) (p < 0.001). In the left IPL activation was mediated by neuroticism (β=0.607).There was also significantly greater activation in the left postcentral gyrus (PG) (p < 0.001), compared to the resilient group. This was mediated by FIRST score (β=-0.61). Again, the vulnerable group scored higher on measures of neuroticism and lower on conscientiousness (p < 0.05). Psychometric information gathered across the 3 studies was collapsed into one dataset (n=84). ANOVA revealed that the vulnerable group had significantly higher scores on measures of neuroticism, perceived stress, state stress, depressive feelings, depressive thinking, brooding, worry, emotion focused and problem focused coping and significantly lower scores on conscientiousness and extroversion (p < 0.05). Results indicate that the vulnerable group are higher on neuroticism across all 3 studies, and score higher on rumination and stress questionnaires in 2 of the studies. Physiological data suggests that the vulnerable group are more sensitive to stress anticipation, as opposed to showing greater reactivity to stress. It is concluded that neuroticism is a risk factor for developing insomnia and that the vulnerable population show greater physiological responses whilst anticipating stress, a phenomena which represents the interaction between personality, rumination and the physiology of the stress system.

Introduction: Alcohol related admissions to critical care are increasing. However, there is uncertainty about the impact of excessive alcohol use on the intensive care stay and recovery from critical illness. Aim: The aim of this study was to understand the impact of alcohol use disorders on the critically ill patient's journey. Settings & participants: The setting for this study was a 20 bed mixed ICU, in a large teaching hospital in Scotland. On admission patients were allocated to one of three alcohol groups: low risk; harmful/hazardous or alcohol dependency. Methods: This was a mixed methods study. An 18 month prospective observational cohort study was undertaken. In addition, 21 in depth, semi structured interviews were undertaken with patients with and without alcohol use disorders, three to seven months after discharge from critical care. Results: 580 ICU patients were screened for the presence of alcohol use disorders during the study period. 34.4% of patients were admitted with a background of alcohol misuse. ICU stay was significantly different between the three study groups, with those in the alcohol dependency group having a longer stay (p=0.01). After adjustment for all lifestyle factors which were significantly different between the groups, alcohol dependence was associated with more than a twofold increased odds of ICU mortality (OR 2.28; 95% CI 1.2-4.69, p=0.01). Four themes which impacted on recovery from ICU were identified in this patient group: psychological resilience; impact and support for activities of daily living; social support and cohesion; and the impact of alcohol use disorders on recovery. Conclusions: Alcohol related admissions account for a significant proportion of admissions to critical care and alcohol dependency is independently associated with ICU outcome. A more targeted rehabilitation pathway for all patients leaving critical care, with specific emphasis on alcohol misuse if appropriate, needs to be generated.

Guidelines generally recommend the monitoring of serum electrolyte and creatinine concentrations in patients treated with antihypertensive therapy in order to detect potential adverse reactions to treatment. However, it is not well known to what extent these guidelines are followed in primary care. I undertook a retrospective analysis of 74096 adult patients from the General Practice Research Database with newly diagnosed hypertension and prescribed a single antihypertensive agent. Baseline biochemical testing was undertaken in 31094 patients (42%) and 37365 (50%) patients had at least one biochemical monitoring test in the year after starting antihypertensive treatment. Monitoring was significantly more likely in patients treated with angiotensin-converting enzyme inhibitors than thiazide diuretics, older patients, and patients with diabetes mellitus. These patient factors were significantly associated with monitoring when multiple imputation was used to control for the potential bias introduced by missing data. In general, follow-up monitoring was infrequent, irregular, and did not change in response to events such as abnormal test results. Patients who were monitored after the initiation of antihypertensive treatment were significantly more likely to be admitted to hospital and discontinue therapy, which is likely a result of reactive instead of planned monitoring. Using propensity score methods to control for confounding, I demonstrated a decreased risk of these same adverse outcomes in patients with baseline testing, which may be because these patients were less likely to have any follow-up monitoring and not the protective effect of the baseline testing. I described several barriers to biochemical monitoring including the lack of consensus in published guidelines, uncertain responsibility for monitoring, patient nonadherence, and absence of alerts or reminders to monitor. More work is needed to improve the primary evidence base for monitoring and to improve the guidelines on the nature and frequency of monitoring for adverse drug reactions, particularly in patients at greater risk of drug-induced harm.

The globally distributed, ciprofloxacin resistant hospital associated methicillin resistant Staphylococcus aureus (MRSA) lineage epidemic (E)MRSA-15 (sequence type 22) continues to expand into new healthcare systems around the world. To further understand the evolution of EMRSA-15, which has been proposed to have emerged in the West Midlands, a large collection of contemporaneous and historical Birmingham isolates were studied alongside a collection of International genomes. Through generation of phylogentic trees this study has been able to determine that EMRSA-15 emerged in c1984 and that during this time a highly related healthcare associated ciprofloxacin sensitive ST22 population was present in Birmingham. A dominant, geographically restricted EMRSA-15 clone was also identified in Birmingham (designated the BHM clone). Analysis of the impact of fluoroquinolone use on the Birmingham EMRSA-15 phylogeny showed increasing use of fluoroquinolones (the introduction of ofloxacin) coincided with the emergence of the BHM clone. Further evidence of geographically restricted EMRSA-15 clones were found in Guernsey and Hong Kong; demonstrating localized evolution giving rise to distinct island populations. Comparison of the desiccation tolerance of geographically restricted and non-restricted EMRSA-15 clones indicated enhanced environmental survival as a possible reason for the spread and dominance of the restricted lineages.

Anaphylaxis is a severe life-threatening allergic condition which is increasing in prevalence and now affects more than 2% of UK children. Anaphylaxis management requires the avoidance of allergen triggers and preparation in readiness for an anaphylactic reaction. People with anaphylaxis and their carers carry Adrenaline Auto-Injectors which need to be administered immediately in the event of an anaphylactic reaction. But, unfortunately, many people often do not know how to use the injectors and fail to use them or fail to use them correctly. This is due in part to deficiencies in training and also to a lack of a system encouraging continuous practice and providing feedback on that practice. Pervasive healthcare research has demonstrated potential in supporting the management of chronic conditions such as diabetes. However, research into assistive technologies for the support of anaphylaxis management has been significantly neglected. Thesis results provide evidence of the potential that smartphone tools have to significantly improve adrenaline injection training skills and a positive influence on self-efficacy. In addition, the results provide insights into possible self-efficacy failings in traditional training and benefits of embedding self-efficacy theory into the design process. The thesis also shows that clinical staff expressed positive feedback after they were provided the technology for one week.

High quality population based data are essential to evaluate breast cancer. This study has improved data quality at the West Midlands Cancer Intelligence Unit through the development of a staging algorithm which significantly increased the proportion of cases with stage recorded. It has also demonstrated the benefits of combining cancer registry data with those from the NHS Breast Screening Programme (NHSBSP). This study identified stage as the most significant prognostic factor for breast cancer patients, and the favourable staging characteristics of screen detected cancers thus suggest that improved survival should result. The study also found that only 31.6% of cancers in eligible women were screen detected in the prevalent round. As 31% of breast cancers arose in women in the screening age range, only an estimated 10% of breast cancer patients benefit directly from screening. High interval cancer rates were identified, along with the need for clarification of the definitions used to identify interval cancers. This investigation identified changes in treatment over time across the region towards the King's Fund guidelines. However, treatment varied across the region and clear divergences from the guidelines were apparent. Furthermore, no association was identified between surgical caseload and survival, suggesting that the use of caseload as a proxy for specialist care may be questionable.

The thesis focused on data analysis of surface-enhanced laser desorption/ionisation time-of-flight mass spectrometry (SELDI-TOF MS) for biomarker discovery and cancer classification. It investigated quantitative measures of reproducibility and found that SELDI protein profiles are affected by sample storage and processing procedure. Two new peak alignment algorithms were proposed, one of which achieved the best performance when compared to the existing methods. The assumption of normality of SELDI protein profiles, on which the standard statistical methods are based, was examined. Normality tests and the multiple testing procedures revealed that SELDI protein profiles do not follow normal distributions, implying that it may be reliable to use non-parametric methods for detecting disease-associated proteins. A new normalisation algorithm was proposed, and was shown to give a better improvement of normality compared with the existing methods. An integrated algorithm to discover proteomic biomarkers for cancer diagnosis was proposed and applied to two published SELDI data sets. The results demonstrated that the receiver operating characteristic (ROC) curve method may be more reliable to determine the discriminatory powers of the identified biomarkers compared to Wilcoxon test. The methods for proteomic biomarker discovery presented here may be generalisable and applicable to other mass spectrometry and genomics approaches.

Alzheimer’s disease (AD) is a devastating neurodegenerative disorder which is thought to affect 26.6 million individuals worldwide. There is growing concern over a worldwide dementia epidemic that is predicted to develop over the coming decades. The evidence thus far suggests that increased levels of oxidative stress and vascular risk factors are two major contributors, amongst others, to AD development. The thesis aimed to investigate markers of oxidative stress in AD plasma. Moreover, the oxidative status of specific proteins was investigated using both hypothesis driven and proteomic approaches. Results presented in this thesis suggest that global plasma protein oxidation levels are not different when AD and control subjects are compared, but that individual plasma proteins are specific targets for oxidative modification in AD. The thesis explores different methodologies to assess oxidative changes in AD. In addition it demonstrates that emerging novel and powerful mass spectrometry techniques can be employed successfully to identify several proteins modified by oxidation, providing an initial starting point for further investigation.

Previous research has indicated that impulsivity is associated with child weight, eating behaviour and some controlling feeding practices and that there are differences in these variables between children with high (including clinically elevated) and low impulsivity levels. Few of these studies have used a range of impulsivity measures to assess this multifaceted concept. This thesis aimed to explore these relationships and differences using a range of parent-report and behavioural impulsivity measures. Three samples of children (2-4-year-olds, 7 -11 -year-olds and 5-15-year-olds) and their parents participated in three studies. Analyses indicated that impulsivity was positively associated with child weight and snack intake (Chapters Three and Five). Links between impulsivity, restriction and pressure to eat were mixed (Chapters Three and Five). Parental monitoring moderated links between impulsivity and food approach behaviour; a lack of monitoring was detrimental to child eating behaviour (Chapter Three). Observations of mealtime behaviours of parent-child dyads in which children had high vs. low impulsivity levels showed that parents of children with high impulsivity levels used more pressure to eat, while their children made more requests for food (Chapter Four). Furthermore, impulsivity, dietary restraint and stress interacted in their effects over snack intake; children high in impulsivity and dietary restraint decreased their intake under stress, while children low in dietary restraint increased their intake under stress (Chapter Six). Finally, parents and their children with and without clinically elevated impulsivity levels differed in eating and feeding behaviours (Chapter Seven). Interesting gender differences emerged throughout and the implications of the results and limitations of the individual studies are discussed in each chapter.

Spontaneous preterm birth complicates about 3% of pregnancies before 34 weeks’ gestation and 7 – 12% before 37 weeks’ gestation. It is an important issue to public health worldwide. The aim of this thesis was to identify test(s) which would predict spontaneous preterm birth in early pregnancy when women are asymptomatic and in later pregnancy when they present with symptoms of threatened preterm labour, using systematic reviews and meta-analysis. If women at risk can be identified whether early in pregnancy or when they present with threatened preterm labour, interventions can be deployed to prevent or delay birth and to improve subsequent neonatal mortality/morbidity. Initially 40,243 title and abstract citations were scrutinised, resulting in shortlist of 1,650 full articles in which 319 were included in the systematic reviews, encompassing 22 tests. The quality of studies and accuracy of tests measured with likelihood ratio (LR) was generally poor. There were only a handful of studies for most of the tests. Few tests reached LR+ point estimates >5. In asymptomatic antenatal women these were ultrasonographic cervical funnelling and length measurement, cervico-vaginal prolactin and cervico-vaginal fetal fibronectin screening for predicting spontaneous preterm birth before 34 weeks’ gestation. In this group, tests with LR- point estimates approaching <0.2 were detection of uterine contraction (by mammary stimulating test) and amniotic fluid CRP measurement. In symptomatic women with threatened preterm labour tests with LR+ point estimate >5 were absence of fetal breathing movements, cervical length measurement, amniotic fluid IL6 and IL8, serum CRP and cervico-vaginal hcg for predicting birth within 2-7 days of testing. In this group tests with LR- point estimate <0.2 were measurement of cervico-vaginal hcg, cervical length measurement, absence of fetal breathing movement, amniotic fluid IL6 and IL8, and serum CRP for predicting birth within 2 - 7 days of testing. In conclusion, no exceptional, but many promising tests for predicting spontaneous preterm birth was identified to aid the development of evidence based practice.