Background: This thesis focuses on the economic analyses of type‐2 diabetes complications defined as macro‐vascular (myocardial infarction, stroke, ischemic heart disease, heart failure) and micro‐vascular (amputation and eye‐related complications leading to blindness in one eye). Diabetes‐related complications are a substantial component of the overall economic, physical and psychological burden of the disease. As the efforts in treating diabetes are geared towards reducing the likelihood of complications, understanding the welfare benefits and future savings from reducing diabetes complications is paramount in determining the cost‐effectiveness of competing diabetes therapies. Aims: The thesis is divided into three essays aiming to (1) characterize changes in the health related quality of life of diabetes patients over time and assess the contributions of diabetes complications to these changes; (2) study the drivers of healthcare expenditure for people with diabetes in terms of both inpatient care and non‐inpatient resource utilization, and estimate the impacts of diabetes‐related complications on health care costs; (3) understand the role played by self‐reported quality of life in predicting mortality after controlling for clinical risk factors. Methods: This thesis uses longitudinal data to answer the questions of interest. A unifying theme across the thesis is the challenge of estimating causal parameters in a context in which there may be substantial observed and unobserved patient heterogeneity. Findings: Failing to account for patient heterogeneity, and in particular un‐measurable variation in patients’ outcomes, is likely to bias the impact of complications on quality of life and on non‐inpatient costs, as well as to confound predicted time to death. In the case of QoL, ignoring heterogeneity is likely to overestimate the impact of complications on self reported utility because the patients who will eventually experience diabetes‐related complications are already on a lower utility path compared to those who do not. In the case of both inpatient and non‐inpatient costs, patients who go on to develop complications have higher cost both pre and post complications. In the case of inpatient costs there is no evidence that unobserved patient heterogeneity matters, while in the case of non‐inpatient utilization the hypothesis of a common baseline level of utilization is rejected in the subset of patients that contribute to the FE identification. This subset however is systematically different from the sample as a whole, being predominately more likely to have complications and other causes of hospitalization. Moreover, a trade‐off occurs when we are interested in predictions; models that exploit within‐patient variation have wider confidence intervals and have thus less precision than population average models. The final substantive chapter finds that HRQoL is significantly associated with survival at the population level and that when patient specific unobserved heterogeneity is taken into account, the power of QoL to predict life expectancy increases. Neglected heterogeneity in frailty causes underestimation of both the extent of positive duration dependence and the impacts of time varying covariates.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:595977 |
| Date | January 2013 |
| Creators | Alva Chiola, Maria Liliana |
| Contributors | Gray, Alastair |
| Publisher | University of Oxford |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://ora.ox.ac.uk/objects/uuid:44d0bd47-2b6b-4c60-8fc1-961d6840650f |
Page generated in 0.0021 seconds