Public Health - Doctoral Theses

Permanent URI for this collection


Recent Submissions

Now showing 1 - 5 of 47
  • Item
    Real-time surveillance for evidence-based responses to suicide contagion and clustering
    (University College Cork, 2022-04-01) Benson, Ruth; Arensman, Ella; Rigby, Jan; Brunsdon, Chris; Health Research Board
    Background: Although a rare phenomenon, suicide clusters are a cause for great community concern. In recent years, strong emphasis has been placed on the importance of detection and real-time surveillance of suicide clusters. Cluster detection increases understanding on the aetiology of a suicide cluster and provides the basis for targeted intervention to mitigate further contagion through the identification of linked cases and socioecological factors associated with the increased risk of clustering in the affected area or population. Policy makers and public health officials benefit from detection of suicide clusters by means of implementing targeted evidence-based interventions in identified vulnerable populations and high-risk areas in a timely manner. To date, suicide cluster detection has been largely restricted to retrospective investigations, limiting its capacity as a tool for intervention. Aim: The aim of this research was to provide a comprehensive understanding of the real-time surveillance, detection, and responses to suicide contagion and clustering. Methods: The thesis is comprised of five interrelated studies, based on both primary and secondary research. Primary data collection was conducted on coronial records of post-inquest cases of confirmed suicide or open verdicts meeting the criteria for probably suicide, as well as pre-inquest cases of suspected suicide. This research also involved quantitative Census data from the Central Statistics Office (CSO). The retrospective and prospective space-time scan statistics based on a discrete Poisson model was employed via the R software environment using the ‘rsatscan’ and ‘shiny’ packages to conduct the space-time cluster analysis and deliver the mapping and graphic components of the dashboard interface in study 3. Evidence synthesis was conducted by means of narrative review of existing literature in studies 1 and 4, and by comparative review of response to a structured questionnaire in study 2. Study 5 employs a secondary research approach to report on policy and practice implications of real-time suicide surveillance. Results: Study 1 synthesised the existing evidence on quantitative techniques to detect suicide and self-harm clusters detection, revealing that a Poisson-based scan statistical model is most effective in accurately detecting point and echo suicide clusters, while mass clusters are typically detected by a time-series regression model, albeit that limitations exist. Study 2 demonstrated more commonalities than differences in a comparison of the components and practices of real-time suicide surveillance systems internationally. Commonalities included rapid, routine surveillance based on minimal, provisional data to facilitate timely intervention and postvention efforts. Identified differences include timeliness of case submission and system infrastructure. Study 3 tested the validity of the scan statistic as a cluster detection approach inbuilt in a dashboard prototype developed to visually display real-time suicide surveillance data. Study 4 identified consistency in both increased quantity of media reports and portrayal of specific details of suicide cases to be significantly associated with suicide contagion and increased suicide rates or mass clusters. An elevated period of risk of suicide contagion has been found to take place between the first days up to the first three months following the media coverage of suicide. Study 5 demonstrated the importance of real-time suicide surveillance in the context of policy and practice, with a particular reference to public health emergencies and humanitarian crises. Conclusions: The findings of this thesis are of relevance in furthering our knowledge of monitoring, detecting, and responding to suicide clusters. Collectively, the findings from this thesis indicate that we can work more efficiently and collectively to mitigate further suicidal behaviour by utilising real-time, provisional suicide data to guide quicker action. The outcomes of this research have methodological implications in terms of suicide and self-harm cluster detection and real-time suicide surveillance. The implications of this research further extend to suicide prevention and mental health policy, clinical practice, means restriction, crisis planning and response, and media reporting of suicide.
  • Item
    The impact of maternal chronic hypertension and chronic kidney disease on the risk of adverse pregnancy outcomes and long-term cardiovascular disease: a population-based epidemiology study
    (University College Cork, 2022-08-31) Al Khalaf, Sukainah; Khashan, Ali; McCarthy, Fergus; O'Reilly, Eilis; Ministry of Health – Kingdom of Saudi Arabia
    Background and aims: The prevalence of chronic hypertension (CH) and chronic kidney disease (CKD) have increased among pregnant women in recent decades. Given the improvement in antenatal care over the last few decades, it is still unclear whether the risk of adverse pregnancy outcomes (APOs) among women with CH and/or CKD has decreased. There is limited evidence on the association between antihypertensive treatment and APOs in women with CH. Although there is evidence that women with a history of APOs have an increased risk of cardiovascular disease (CVD), it remains unclear whether pre-pregnancy hypertension and the occurrence of APOs would influence this association. The aim of this PhD project was to investigate the impact of maternal CH and/or CKD and antihypertensive treatment on the risk of APOs and long-term CVD. Structure and methods: This thesis includes eight chapters: Introduction, Methods, two systematic review articles on the impact of CH and CKD on APOs, three original research articles, and Discussion. Data from the Swedish National Registers were analysed to examine the associations between CH/CKD and the risk of APOs over the last three decades. Data from the UK CALIBER platform were used to investigate: i) the association between CH and APOs, with a focus on the role of antihypertensive treatment and control of hypertension, and ii) the associations between pre-pregnancy hypertension and subsequent diagnosis of 12 different CVDs, considering the role of APOs on these associations. Adverse pregnancy outcomes were pre-eclampsia, preterm birth, stillbirth, Caesarean section and small for gestational age (SGA). The statistical methods were done using logistic regression models for the Swedish data, logistic regression models with propensity score matching for the antihypertensive treatment analyses, while the associations between pre-pregnancy hypertension and CVD were analysed using stratified Cox models. All statistical models were adjusted for several potential confounders. Results: Systematic reviews and meta-analyses: CH was associated with 5-fold increased odds of pre-eclampsia and approximately 2-fold increased odds of stillbirth, preterm birth, and SGA, compared to women without CH. Women with treated CH (compared to untreated normotensive women) had higher odds of APOs. However, the results were inconsistent when outcomes were compared between treated and untreated women with CH; no increased odds of superimposed pre-eclampsia or other APOs were observed, except for 86% increased odds of SGA. Findings from the meta-analysis suggested that women with CKD had higher odds of pre-eclampsia, Caesarean section, preterm birth, very preterm birth, and SGA. All causes of CKD were associated with increased odds of pre-eclampsia, preterm birth, and SGA, with stronger associations in women with diabetic CKD, particularly for preterm birth [adjusted odds ratio (aOR): 4.76, (95% confidence interval (CI), 3.65–6.21)] and SGA [aOR: 4.50, (95% CI, 2.92–6.94)]. The findings according to the severity of kidney disease showed that later stages of CKD were associated with a greater odds of APOs than earlier stages. Swedish National Registers: The overall findings from this study suggested that the odds of APOs remain high in women with CH and/or CKD, and the odds persisted independent of parity, maternal age, and body mass index, among other potential confounders. No association was found between CKD and stillbirth. All causes of CKD were associated with higher odds of pre-eclampsia, emergency Caesarean section, and medically indicated preterm birth, and the ORs were higher in women with diabetic CKD, renovascular disease, and congenital kidney disease than other CKD subtypes. CALIBER studies: The results suggested a higher odds of APOs in women with CH (treated and untreated) compared to untreated normotensive women. In women with CH, those requiring treatment (versus untreated) had 17%, 25%, and 80% increased odds of superimposed pre-eclampsia, preterm birth, and fetal growth restriction (FGR), respectively. However, these results were mainly attributable to the level of blood pressure (BP) control among the treated group; as similar results were found between the untreated and treated women with CH who achieved tight control (BP<135/85 mmHg) for all assessed outcomes except for a 59% decreased odds of superimposed pre-eclampsia and a 55% increased odds of FGR. Pregnant women with CH who were prescribed methyldopa (versus β-blockers) had 43%, 59%, and 44% increased odds of superimposed pre-eclampsia, preterm birth, and very preterm birth, but 66% lower odds of FGR. No differences in outcomes were found in women who were prescribed calcium-channel-blockers (versus β-blockers) except for 94% increased odds of preterm birth. The magnitude of the associations increased with increasing BP, and the strongest associations were observed in women with severe hypertension (BP≥ 160/90 mmHg). In treated women with CH, less-tight (BP≥135/85 mmHg) versus tight (BP<135/85 mmHg) control was associated with almost 2-fold higher odds of superimposed pre-eclampsia, very preterm birth, and a 3-fold higher odds of severe hypertension. During the 20-year study period, 16,499 CVD incident were observed, of which two-thirds (66%) had occurred in young women (under 40 years). Pre-pregnancy hypertension (versus no pre-pregnancy hypertension) was associated with a 2-fold higher risk of any subsequent CVD. When the results were subclassified according to the presence of APOs, the strongest associations were found in women with pre-pregnancy hypertension and APOs across the 12 CVD; with almost a 3-fold increased risk to develop any subsequent CVD, an 8-fold increased risk of coronary heart disease, and a 10-fold increased risk of heart failure, compared to those who remained normotensive without APOs. Conclusions: This thesis indicated that CKD and CH were associated with a wide range of APOs than the general obstetric population. Therefore, multidisciplinary prenatal consultation and antenatal management should be provided for these women with close monitoring during pregnancy. If antihypertensive treatment is required, clinicians might consider tighter control during pregnancy as better outcomes were observed in women with tightly controlled hypertension. β-blockers might be superior in reducing APOs than methyldopa, with an exception for FGR, which was higher in the β-blockers group. Finally, the findings suggested strong associations between pre‐pregnancy hypertension with subsequent CVD, with a greater risk among women who had pre-pregnancy hypertension and APOs. Pre-pregnancy hypertension should be managed adequately during pregnancy to reduce the risk of APOs and subsequently reduce the risk of CVD, which emphasizes that a history of reproductive risk factors (including APOs) should be considered in screening tools for CVD beyond the postpartum period to optimize long-term cardiometabolic health in women.
  • Item
    Recognising emotional aspects of diabetes; understanding detection of diabetes distress and depression among people with type 2 diabetes mellitus in community settings
    (University College Cork, 2021-10-01) McGrath, Niamh; Kearney, Patricia M.; Mchugh, Sheena; Toomey, Elaine; Health Research Board
    Background: Diabetes distress and depression are important to address among people with type 2 diabetes (T2DM), who increasingly receive T2DM care in the community. Detection is a pre-requisite for appropriate management, but evidence indicates low detection rates among people with T2DM. Diabetes distress and depression screening have been recommended in contemporary T2DM management guidelines as a strategy to detect these issues. Little is known about detection among people with T2DM in high-income countries or about factors influencing detection at the health system and heath professional (HCP) levels in this population. Moreover, implementation of diabetes distress and depression screening of people with T2DM in the community remains poorly understood. Aim: To improve understanding of diabetes distress and depression detection and the implementation of screening among people with T2DM in community settings. Methods: This PhD thesis employed a mixed methods approach, using a convergent design. Three studies were conducted. First, a cross-sectional comparative analysis estimated the prevalence of depression detection outcomes (undiagnosed, symptomatic and diagnosed and asymptomatic and diagnosed) and using multivariable logistic regression, modelled the association between diabetes and undiagnosed (versus diagnosed) depression in three high income countries with different health systems. Second, a qualitative evidence synthesis (QES) following the best-fit framework method employed behaviour change theory, the theoretical domains framework (TDF), to identify HCP perceived barriers and enablers to diabetes distress and depression screening of people with T2DM in community settings, based on HCPs’ experiences of implementing screening. Finally, a primary qualitative study using the TDF as a data collection and analytic framework and drawing on principles of chart stimulated recall, was conducted with HCPs working in the community setting in Ireland to understand current diabetes distress and depression detection practices and prospectively identify determinants of screening implementation in the Irish context. Results: Undiagnosed depression was more prevalent in people with diabetes than without diabetes in each country with absolute rates varying by country; [Ireland: diabetes 10.1%(95%CI:7.5-12.8) vs no diabetes 7.5%(95%CI:6.8-8.2), England: diabetes 19.3%(95%CI:16.5-22.2) vs no diabetes 11.8%(95%CI:11.0-12.6), USA: diabetes 7.4%(95%CI:6.4-8.4) vs no diabetes 6.1%(95%CI:5.7-6.6)]. There was no clear pattern of association between diabetes status and undiagnosed depression across countries; Ireland: OR=0.82(95%CI:0.5-1.3), England: OR=1.47(95%CI:1.0-2.1), USA: OR=0.80(95%CI:0.7-1.0). Ten articles (9 studies) were synthesised in the QES. All articles related to depression screening and one to diabetes distress screening. Articles included GP (n=7) and nurse (practice and nurse specialist; n=9) perspectives. Fifteen factors comprising 11 barriers, 3 enablers and 1 factor which was both a barrier and enabler, in 8 TDF domains: knowledge, skills, professional role and social identity, beliefs about consequences, reinforcement, environmental context and resources, memory, attention and decision process, emotion, and 1 new domain: people with T2DM factors, were identified as influencing screening implementation. Three barriers; underestimation of depression severity, skills needed to screen, and discomfort screening were predominantly discussed by nurses. One barrier; culture and language issues and 1 enabler; availability of appropriate reimbursement, were only discussed by GP participants. Twenty-eight HCPs across Ireland: 7 GPs, 9 practice nurses, 8 diabetes nurse specialists and 4 dieticians participated in semi-structured telephone interviews. Current practice involved people with T2DM or third parties (e.g. their family members) alerting HCPs to possible mental health issues among people with T2DM, or by HCPs asking people with T2DM about mental health as a response to implicit cues by people with T2DM. Some participants had administered screening tools albeit not systematically or necessarily part of T2DM care. Fifteen factors, comprising 7 barriers, 6 enablers and 2 factors which could inhibit or enable screening were identified. These related to 8 TDF domains; knowledge, skills, beliefs about consequences, environmental context and resources, memory, attention and decision process, emotion, social influences and 1 other domain; person with T2DM factors. Lack of knowledge about which tool(s) to use by HCPs with experience of screening and lack of awareness of diabetes distress, and, limited focus on psychosocial aspects in general practice due to under-resourcing of T2DM were only discussed by GPs and practice nurses. Conclusion: Detection of diabetes distress and depression in people with T2DM is a challenge across health systems and the extent of detection as a challenge varies across countries. Screening implementation strategies should include educational materials, meetings and/or outreach activities, local opinion leaders, reminders and tailored interventions to address HCP specific determinants.
  • Item
    Attributable burden, life expectancy and income loss to non-communicable diseases in Ireland: from evidence to policy-making
    (University College Cork, 2020-12-18) Chakraborty, Shelly; Kabir, Zubair; Perry, Ivan J.; Balanda, Kevin; Health Research Board
    Introduction: This thesis is devoted, primarily, to the study of the burden of non-communicable diseases in the Republic of Ireland. Four major NCDs – cancer, cardiovascular disease (CVD), diabetes and Chronic Obstructive Pulmonary Disease (COPD) – are chosen and their risk factors as well as socio-economic impact is analysed. Burden of disease studies utilize the evidence base and quantify the health scenario and trends over time. Such estimates are relevant for a country’s policy makers to allocate the limited resources to the areas that require more attention. The Global Burden of Disease study was the most comprehensive study of its kind to draw attention to the diseases and risk factors in any country which require more attention than the others. We seek to do the same for Ireland with a special focus for NCDs. NCDs contributed to approximately 71% of deaths worldwide in 2016 (WHO). NCDs lead to early deaths, morbidities and long-term health effects. It was observed that life expectancies have increased in Ireland. It is critically important to determine whether people are living these additional years in good health or not. NCDs also contribute to considerable economic loss. These considerations form the motivating factors for this study, whose primary objectives are as follows: 1. To estimate NCD burden and attributable burden, including LE and HALE- 1990 to 2017 2. To investigate the main drivers of NCD burden changes in 1990-2017 3. To estimate the national economic loss to NCD burden 4. To estimate the future NCD burden- 2040 5. To carry out Internal validation exercises Methods: This work uses methods similar to that of the GBD study. The metrics used to quantify the burden are Disability Adjusted Life Years (DALYs), number of deaths, Years of Life Lost (YLLs) and Years Lived with Disability (YLDs). DALYs quantify the burden due to premature mortality and disability. They enable comparisons across nations, thus allowing benchmarking. Decomposition analysis is used to separate the effects of population ageing, population growth and changes in the risk factor scenario. We also quantified Gross Domestic Product (GDP) per capita lost to DALYs in Ireland in 2017 using the WHO model for projecting the Economic Costs of Ill-Health (WHO--EPIC model). Each DALY amounts to a loss of 1 to 3 units of GDP per capita (in international dollars). The burden of disease in 2030 was projected for Ireland using GBD’s forecasting methods, and the potential and avertable YLLs were calculated. The GBD study results were also validated against some national level estimates. This is done to ascertain whether, and to what extent, the data from GBD would be an appropriate evidence base to formulate national level policy decisions. The estimates were also validated using different standard populations by checking if the results varied significantly. Main Findings: This study analyzed data from GBD as well as some national level databases, and the following are some of the main findings of the analysis. Ireland ranks 11th best in the NCD related age-standardized DALYs, 2nd in the age-standardized deaths and 3rd in the age-standardized YLDs within the EU. This clearly shows that, while Ireland performs well in terms of number of deaths and YLDs, there is still a lot to be done when it comes to DALYs. Life Expectancy in the ROI has increased during 1990 and 2017, however the Healthy life expectancy (HALE) has not increased at the same pace. This means that the additional years are lived in less than ideal health. It was found that diabetes, ischemic heart disease (IHD) and stroke related death rate attributable to diets high in processed meat and sodium have increased during 1990 and 2017. Cancer and diabetes YLDs attributed to alcohol consumption have increased significantly (from 2.4 per 100,000 to 37.6 for diabetes and 15.9 per 100,000 to 25.4 for cancer). On disentangling the drivers of burden change, it was observed that except diabetes, the other three NCDs have shown more improvements in the risk factor scenario. We estimated that Ireland lost about 75.6 million international dollars (I$) in GDP per capita from all causes in 2017. An alarming result here is that, of the 75.6 million mention above, cancer alone contributed to 15 million I$. A part of this study also focused on forecasting the burden in 2040. The model predicts that IHD will be the leading cause of deaths in 2040, whereas smoking is on its way to becoming the leading cause for YLLs. In all the above-mentioned analyses, estimates based on national level data agreed with those obtained from the analysis of the GBD data. On using different standard populations the DALYs varied. Conclusion: To the best of our knowledge, this is the first comprehensive epidemiologic and economic profile of NCD burden in Ireland. It was observed that the YLDs are on the rise and diabetes is the most neglected NCD in terms of effective population interventions. In terms of the risk factors, alcohol was found to be the leading cause of NCD morbidity. It should also be mentioned that, even though estimates obtained from GBD were in agreement with the national burden estimates, this study cannot be a replacement for a sub-national BOD study.
  • Item
    Adverse pregnancy outcomes and the long-term risk of maternal kidney disease
    (University College Cork, 2021-01-06) Barrett, Peter M.; Khashan, Ali; McCarthy, Fergus; Kublickiene, Karolina; Wellcome Trust; Health Research Board
    Background: Adverse pregnancy outcomes, including hypertensive disorders of pregnancy (HDP), preterm delivery, foetal growth restriction, gestational diabetes (GDM), and pregnancy loss, have been associated with the risk of maternal chronic disease, particularly cardiovascular disease. Less is known about the long-term risk of chronic kidney disease (CKD) and end-stage kidney disease (ESKD) in women who have experienced pregnancy complications. This thesis aims to examine associations between adverse pregnancy outcomes and the risk of maternal CKD and ESKD in later life. Structure and methods: This thesis begins with an introductory chapter (Chapter 1) followed by a systematic review and meta-analysis of the published literature, based on a pre-specified protocol (Chapter 2). A detailed methods chapter outlines the data sources, study design, exposure and outcome variables, and statistical approach used in each of the original observational studies conducted for this research (Chapter 3). Four population-based cohort studies are presented, and they focus on the risk of maternal kidney disease following preterm delivery (Chapter 4), stillbirth (Chapter 5), HDP (preeclampsia and gestational hypertension) (Chapter 6) and GDM (Chapter 7) respectively. In each study, data from the Swedish national registers are used, and analyses are based on Cox proportional hazard regression models with time-dependent covariates, adjusted for a wide range of medical, obstetric, and socio-demographic factors. In Chapter 8, an updated systematic review and meta-analysis is presented to reflect newly published literature on this topic. This is followed by a discussion and interpretation of the key findings, including consideration of the public health implications arising from this work (Chapter 9). Finally, conclusions of the thesis are presented in Chapter 10. Results: (i) Updated systematic review and meta-analysis - Overall, the published literature on this topic was sparse and most meta-analyses were based on small numbers (<5) of original studies. HDP and preterm delivery were associated with higher risk of long-term kidney disease in parous women. Preeclampsia was associated with a strongly increased risk of ESKD (adjusted risk ratio (aRR) 4.90; 95% CI, 3.56-6.74) and a modest increased risk of CKD (aRR 1.73, 95% CI 1.42-2.12). Gestational hypertension was associated with a strongly increased risk of ESKD (aRR, 3.64, 95% CI, 2.34-5.66), and more modest increased risk of CKD (aRR 1.48, 95% CI 1.38-1.58). Preterm delivery was associated with an increased risk of ESKD (aRR 2.19, 95% CI 1.93-2.47), but there were too few studies to determine the risk of CKD, or to separate the effects of iatrogenic vs. spontaneous preterm deliveries. No significant association was observed between GDM and CKD (aRR 1.04, 95% CI 0.76-1.41), but this meta-analysis was based on pooled estimates from just two studies. (ii) Original population-based cohort studies - Preterm delivery was associated with increased risk of both CKD and ESKD in our study. This association was strongest in women who experienced iatrogenic preterm delivery (due to preeclampsia or small for gestational age (SGA)), but the risk persisted in women who only had spontaneous preterm deliveries compared to women who delivered at term (for CKD, aHR 1.32, 95% CI 1.25-1.39; for ESKD, aHR 1.99, 95% CI 1.67-2.38). Separately, stillbirth was alsoassociated with an increased risk of both CKD (aHR 1.26, 95% CI 1.09–1.45) and ESKD (aHR 2.25, 95% CI 1.55-3.25) compared to women who only experienced live births. These associations persisted independent of preeclampsia, SGA or congenital malformations. Preeclampsia was associated with a higher risk of CKD during follow-up (vs. no preeclampsia, aHR 1.92, 95% CI 1.83–2.03), but this risk differed by CKD subtype and was greater for hypertensive CKD, diabetic CKD, and glomerular/proteinuric CKD. Women who experienced preterm preeclampsia, recurrent preeclampsia, or preeclampsia complicated by antenatal obesity were at particularly high risk of CKD. There was also a modest risk of CKD among women who experienced gestational hypertension (vs. none, aHR 1.49, 95% CI 1.38–1.61). GDM-diagnosed women were at increased risk of CKD and ESKD overall. However, when GDM was stratified according to those who developed post-pregnancy type 2 diabetes (T2DM), the associations between GDM alone (without later T2DM) and maternal kidney disease were non-significant (for CKD, 1.11, 95% CI 0.89-1.38; for ESKD, aHR 1.58, 95% CI 0.70-3.60). By contrast, strong associations were observed with CKD and ESKD in those who had GDM followed by subsequent T2DM. Conclusion: Adverse pregnancy outcomes, specifically preeclampsia, gestational hypertension, preterm delivery and stillbirth, are associated with increased risk of maternal CKD and ESKD. These associations persisted in a nationwide cohort after controlling for a wide range of confounders. Although the relative risk of future kidney disease is highest for ESKD, associations with CKD are likely to be of greater importance from a population perspective, given the high prevalence of CKD. Women who experience adverse pregnancy outcomes may warrant systematic follow-up to prevent onset or progression of future kidney disease, but the optimal format and timing of this follow-up requires further research.