Background Respiratory viruses are a leading cause of acute lower respiratory infections (ALRI) in young children. The role of seasonal influenza virus in childhood ALRI is generally underappreciated. This is because the global burden of disease due to ALRI attributable to seasonal influenza virus in children is unknown. This thesis aims to estimate the global and regional hospital admissions for seasonal influenzaassociated ALRI and the possible boundaries for influenza-associated ALRI mortality in children younger than five years. The WHO has developed guidelines for influenza surveillance using severe acute respiratory infections (SARI) sentinel surveillance network. However, data from sentinel surveillance are not routinely used in estimating disease burden in a population. This thesis also aims to provide tools for estimating influenza disease burden using data from SARI sentinel surveillance in developing country settings. Methods Incidence data for influenza-associated ALRI (from passive, hospital-based studies) were collected using a systematic review of studies published between January 1, 1995 and October 31, 2010. These data were supplemented by unpublished data from 15 population-based studies that were obtained by forming a consortium of researchers (Influenza Study Group) working in developing countries. The incidence meta-estimates were applied to global and regional population estimates for 2008 to calculate the estimated number of hospitalised influenza-associated ALRI cases that year. The possible bounds for influenza-associated mortality were estimated by combining incidence estimates with in-hospital case fatality ratios and identifying studies with population-based data for influenza seasonality and monthly ALRI mortality. The data to estimate the incidence of all-cause hospitalised ALRI were collected using a systematic literature review that was supplemented with unpublished data from 24 population-based studies that were obtained by collaborating with research sites in developing countries (Severe ALRI Working Group). The hospitalised ALRI incidence meta-estimates were applied to global and regional population estimates for 2008 to calculate the estimated number of all-cause hospitalised ALRI cases that year. Data on the proportion of hospitalised ALRI cases that were positive for influenza were collected using a systematic review of the studies published between January 1, 1995 and December 31, 2011. The meta-estimates of the proportion of hospitalised ALRI cases positive for influenza were applied to the estimated number of hospitalised ALRI cases in the year 2008 to estimate the number of hospitalised influenza-associated ALRI cases globally and for the six WHO regions using this alternative method. The tools for estimating influenza disease burden using surveillance data were developed after a literature review and a survey of 27 end-users (influenza epidemiologists) in 24 countries. Results Thirty nine studies (21 from developing and 18 from industrialised regions) satisfying the eligibility criteria, provided data on the incidence of influenza-associated hospitalised ALRI. The incidence is highest in infants in the first six months of life, both in developing as well as industrialised countries. It is estimated that the incidence of hospitalised influenza-associated ALRI in children under the age of five years was about 1.5 (95% CI 1.0 to 2.3) and 1.2 (95% CI 0.9 to 1.6) per 1000 children in developing and industrialised countries respectively. This translates to about 911,000 (95% CI 617,000 to 1.4 million) hospitalisations worldwide due to influenza-associated ALRI in children younger than five years in 2008, 93% of the cases occurring in developing countries (where 90% of the global under-5 population reside). An estimated 21,500 (based on 20 studies) to 115,000 deaths (based on only 1 study) in under-five children were attributable to influenza-associated ALRI in 2008. Incidence and mortality varied substantially from year to year in any one setting. Eighty five studies (61 from developing and 24 from industrialised) reported incidence of hospitalised ALRI in children aged 0 to 4 years. It is estimated that about 11.3 (95% CI 9.5 to 13.5) million episodes of ALRI resulting in hospitalisation occurred worldwide in children aged 0 to 4 years in 2008, 92% of these occurring in developing countries. Twenty three studies (19 from developing and 4 from industrialised) reported data on proportion of hospitalised ALRI cases testing positive for influenza using laboratory tests. The estimated proportion of influenza-positive hospitalised ALRI cases was about 5.0 (95% CI 3.6 to 7) percent and 8.4 (95% CI 4.2 to 16.7) percent in developing and industrialised countries respectively. This translates to about 772,000 (95% CI 343,000 to 1.8 million) cases of influenza-associated hospitalised ALRI in children younger than five years worldwide in the year 2008. A manual (targeted at developing countries) describing the methods to estimate the disease burden associated with seasonal influenza using the various surveillance data was developed after considering the results of the preliminary survey. An electronic tool (based on a spread sheet model) to help the end-users (epidemiologists at sentinel surveillance sites and Ministries of Health) to estimate the disease burden at local and national levels was developed as an adjunct to the manual. The manual along with the electronic tool were piloted at three different sites in two developing countries (India and Ghana) and feedback from the end-users was obtained to make the version more user-friendly. The final draft of the manual along with the tool has been submitted to the WHO for final clearance. The member states and the WHO Eastern Mediterranean Regional Office decided to adopt the manual and in the first instance estimate the influenza disease burden in 8 member states having the requisite data for undertaking disease burden estimation. Conclusions Influenza is a common pathogen identified in children with ALRI and results in a substantial burden on hospital inpatient services worldwide. There are significant gaps in published data from developing countries (especially the African and Eastern Mediterranean regions of the WHO). Sufficient data to precisely estimate the role of influenza in childhood mortality from ALRI are not presently available. Effective use of sentinel surveillance data for disease burden estimation would greatly improve the quality and precision of disease burden estimates (especially those resulting in hospitalisation). Improved disease burden estimates (particularly at the national level) would inform policy makers and national governments in formulating immunization policies for vaccinating high-risk groups, and planning annual requirements for vaccines and anti-viral drugs against seasonal influenza.

MSc (Med) Emergency Medicine, Faculty of Health Sciences, University of the Witwatersrand / Adequate exposure to paediatric pre-hospital emergency cases for students undertaking clinical learning is a key component of preparation for independent practise. Both clinical reasoning and psychomotor skills require practise in a realistic environment in order to best equip the qualifying practitioner for demands of the real world of pre-hospital emergency care. The aim of this study was to retrospectively describe the exposure of pre-hospital emergency care students in the University of Johannesburg‟s National Diploma in Emergency Medical Care programme to emergencies involving paediatric patients in the Greater Johannesburg Metropolitan area over a continuous eight year period, between 1 January 2001 and 31 December 2008. Patient care records contained in an electronic clinical learning management information system entered over the eight-year study period were analysed in order to characterise the exposure of students to paediatric emergency cases in general, and clinical skills performed during this exposure. Results showed that, with the exception of infants and children seen by first year students, median exposure to paediatric emergency cases for students in all academic years was below 50%. Exposure to emergencies involving younger patients was generally lower than that for older patients, however the acuity of patients increased with decreasing age. Exposure to most clinical skills also decreased with decreasing patient age. Opportunities for students to practise critical or invasive skills were relatively rare. Suggestions for the improvement of student exposure to paediatric emergency cases and clinical skills include a period of internship and greater utilisation of hospital-based clinical skills exposure and practice.

Background/Project Description: There is a growing interest in the contribution of late preterm (34 – 36 week gestational age (GA)) birth to neonatal morbidity and mortality. Late preterm infants have an increased incidence of both respiratory and non- respiratory complications over the first year of life. Rates of prematurity as well as morbidity/mortality in infancy are higher in lower socioeconomic status (SES) groups but how GA and SES interact is relatively unexplored. Methods/Participant Population: A retrospective cohort study was undertaken utilizing anonymized data housed at the Manitoba Centre for Health Policy (MCHP). A population-based cohort of infants born at 34 to 41 weeks of GA was assembled; individual and area-level income information was used to develop SES groups. Outcomes studied included diagnoses received during the birth hospitalisation, neonatal and post-neonatal admissions. Regression models were constructed to explore the effects of GA and SES as well as control for multiple perinatal variables. Appropriate approvals and safeguards for data privacy were maintained. Results: GA and SES exerted a gradient effect on morbidity, which persisted after controlling for multiple confounding variables. The effect of GA was strongest during the birth hospitalisation but persisted throughout the first year with increased morbidity evident with each week of decreasing GA. The detrimental association of low SES with morbidity increased in effect size throughout the first year surpassing that of GA for post-neonatal admissions. An interaction effect of maternal diabetes, respiratory morbidity and SES was suggested and merits further investigation. Neonatal stays of 3 days or longer negated the association of GA with readmission within the first 28 days; in addition shorter stay infants had the highest risks of readmission at 37 weeks as compared to the late preterm gestations. Conclusions: The consistent associations between poverty, prematurity and morbidity require both further study and attention. Attention to the neonatal health of both late preterm and term infants is important due to their large numbers and population impact. The added risk of poverty merits urgent and multifaceted interventions to lay the groundwork for healthy childhood and long-term success.

Background/Project Description: There is a growing interest in the contribution of late preterm (34 – 36 week gestational age (GA)) birth to neonatal morbidity and mortality. Late preterm infants have an increased incidence of both respiratory and non- respiratory complications over the first year of life. Rates of prematurity as well as morbidity/mortality in infancy are higher in lower socioeconomic status (SES) groups but how GA and SES interact is relatively unexplored. Methods/Participant Population: A retrospective cohort study was undertaken utilizing anonymized data housed at the Manitoba Centre for Health Policy (MCHP). A population-based cohort of infants born at 34 to 41 weeks of GA was assembled; individual and area-level income information was used to develop SES groups. Outcomes studied included diagnoses received during the birth hospitalisation, neonatal and post-neonatal admissions. Regression models were constructed to explore the effects of GA and SES as well as control for multiple perinatal variables. Appropriate approvals and safeguards for data privacy were maintained. Results: GA and SES exerted a gradient effect on morbidity, which persisted after controlling for multiple confounding variables. The effect of GA was strongest during the birth hospitalisation but persisted throughout the first year with increased morbidity evident with each week of decreasing GA. The detrimental association of low SES with morbidity increased in effect size throughout the first year surpassing that of GA for post-neonatal admissions. An interaction effect of maternal diabetes, respiratory morbidity and SES was suggested and merits further investigation. Neonatal stays of 3 days or longer negated the association of GA with readmission within the first 28 days; in addition shorter stay infants had the highest risks of readmission at 37 weeks as compared to the late preterm gestations. Conclusions: The consistent associations between poverty, prematurity and morbidity require both further study and attention. Attention to the neonatal health of both late preterm and term infants is important due to their large numbers and population impact. The added risk of poverty merits urgent and multifaceted interventions to lay the groundwork for healthy childhood and long-term success.

This thesis presents a series of longitudinal studies of HIV-outcomes in Australia and the Asia-Pacific region since highly active antiretroviral therapy (HAART) became available. The primary source of data is the Australian HIV Observational Database (AHOD). AHOD is an observational cohort of more than 2000 patients with HIV recruited via hospitals, sexual health centres and general medical practices specialising in HIV medicine. Chapter five of this thesis addresses whether there were any differences in antiretroviral therapy use and virological response that might explain the different trends in new HIV diagnosis rates between state jurisdictions in Australia in recent years. Analysis of data from cohort studies of primary and chronic HIV infection, gay community surveys and national prescription data suggest that, for the most part, antiretroviral therapy use and virological response were similar in each jurisdiction during the first decade of HAART. Chapter six describes the prevalence of, and risk factors for, an incomplete immune response despite sustained viral suppression in patients on HAART in AHOD. The clinical relevance of this phenomenon is also explored in terms of AIDS and death during follow-up. Of those with sustained viral suppression, one third of patients did not achieve immune recovery greater than 350 cells/??l in the 12-24 months after starting their first or second HAART regimen, and this was associated with a lower CD4 cell count at baseline. Chapter seven describes cause-specific mortality in patients with HIV in the Asia-Pacific region. Immunodeficiency was associated with non-AIDS and AIDS mortality, and the risk of non-AIDS mortality increased with age. Less conclusive was the relationship between country-income level and risk of death from AIDS or non-AIDS causes because of the relatively high proportion of unknown causes of death in low-income settings. Chapter eight presents hospitalisation rates, risk factors and associated diagnoses in patients with HIV in Australia. Older, sicker individuals, as indicated by markers of advanced immunodeficiency or frequency of hospitalisation, were at greater risk of hospitalisation and death in the AHOD cohort. Despite effective antiretroviral therapy, patients with HIV are currently hospitalised at higher rates than people of similar age in the general population.

Th.--Méd.--Nancy 1, 1984. N°: 245.

Research suggests that trauma plays a part in Post Traumatic Stress Disorder (PTSD) and psychosis, but it is unclear what role psychotic symptoms or hospitalisation have. Some research has been carried out on mediators and predictors of PTSD and integrating the evidence has key implications for individual and service outcomes. The two papers presented in this thesis aim to contribute to research in this area, followed by a critical review of the research, its relevance and future implications.Paper One is a systematic review of the literature investigating the prevalence of psychosis-related and hospital-related PTSD and considered what factors moderate or mediates these symptoms. Studies showed high levels of psychosis-related and hospital-related PTSD and seventeen factors that may influence the development of psychosis-related or hospital-related PTSD were explored. However incidence rates in different populations are lacking and there is a need to better explore mediating and moderating factors.Paper Two aimed to investigate the traumatic nature of psychosis and hospitalisation and their relationships with attachment and recovery styles in people with psychosis in a secure setting. The final section, the Critical Review, aimed to place the research in a wider context, considering the findings from the research, limitations of the study, highlighting important issues for services, and implications for interventions and future studies.

Background: International analyses suggest that routine maternal vaccination with seasonal trivalent influenza vaccine is cost-effective, but few studies have been done in middle- to low- income countries. Method: A decision-tree analysis was modelled for the South African public healthcare setting over one year from a payer's perspective. Direct medical costs and consequences were obtained from published literature. Incremental cost effectiveness ratios (ICERs) and univariate sensitivity analyses were then measured. Discounting was excluded due to the seasonality of influenza, limiting the time horizon to a one year period. Findings: The model predicted that to avert influenza-associated hospitalisations amongst pregnant women and their infants less than six months of age, vaccination of pregnant women was not cost-effective. This was irrespective of whether the universal vaccination or HIV-targeted approach was used. A base model simulating 100% vaccine uptake predicted that seasonal vaccination of 100,000 pregnant women results in an estimated net cost of R69,118,114.05 per neonatal influenza-associated hospitalisation averted. Similarly, the model suggested that vaccinating 100,000 pregnant women would cost R1,197,779.79 per maternal hospitalisation averted. Univariate sensitivity analyses reinforced that influenza vaccination of pregnant women was not cost-effective, except when lower incidence of maternal influenzaassociated hospitalisations associated with antenatal influenza vaccination were simulated where the targeted approach became dominant. The latter analysis predicted savings of R770,530.86 per maternal influenza-associated hospitalisation averted. Interpretation: The ICERs suggest that influenza vaccination amongst pregnant women is not cost-effective in the South African public healthcare sector compared to no vaccination, with respect to averting influenza-associated hospitalisations amongst pregnant women and their infants less than six months of age. However, these estimates should be re-evaluated, pending vaccine effectiveness studies of higher methodological quality for low- and middle- income countries and using cost inputs relevant to South African public healthcare setting. This analysis may provide preliminary information regarding the upscaling of influenza vaccination amongst pregnant women as a priority in the constraints of a limited healthcare budget and careful consideration is required regarding vaccine mobilisation amongst pregnant women. / Dissertation (MSc)--University of Pretoria, 2016. / School of Health Systems and Public Health (SHSPH) / MSc / Unrestricted

Introduction: Acute chemotherapy toxicity is common and can have negative effects for the patient and health economy and hospitalisation can be necessitated. Aims: To identify the incidence of toxicity and admission, and predictors of toxicity occurrence, severity, hospitalisation and length of stay. Method: Data was obtained from a proactive telephone assessment of acute toxicity 24 hours after administration of a first cycle of chemotherapy to patients in a large UK NHS teaching hospital. Results: 1539 patients were studied and the overall incidence of toxicity was 35.6% (530 patients). Disease site and number of chemotherapy agents given were shown to predict toxicity, with breast and upper gastrointestinal cancers having a higher likelihood of toxicity. Disease was predictive of toxicity grade, with urology, gynaecology and lung cancer patients experiencing higher grades of toxicity than other tumour sites. The rate of hospital admission due to toxicity was 13.1% (203 patients) and median length of stay 3 days (1-28). The risk of admission had some risk factors in common with toxicity. Disease and the number of drugs in the regimen affected the risk of admission, with gynaecology, head and neck and lung cancer patients and patients who received 3 drugs having a higher likelihood of admission. Predictors in the subgroups of breast, colorectal and lung cancer patients did not differ greatly from the whole population and the number of drugs was shown to be a predictor of nausea, vomiting and fatigue when explored as secondary outcomes. Conclusion: The research partly addressed the main aim and highlighted areas where further research is required. Keywords

Avhandling--Uppsala, 2001. / Bibliogr. p. 370-395. Index. Résumé en anglais.