In this thesis, a microsimulation model was developed using methods which were intended to overcome the main criticism of earlier models developed in Australia - that their estimation of the distribution of health benefits1 across income groups was not accurate. To determine whether the new model � called the Person Level Model of Pharmaceutical Benefits (PLM-PB) � was more accurate, two typical means-based models were also built to replicate the most commonly used methods in Australia. A comparison of the results of the three models revealed that while they produced comparable results at the aggregate when compared with administrative data, the PLM-PB was much more accurate in capturing distributional differences by beneficiary and medication type. The PLM-PB also indicated that, as anticipated, PBS benefits were more pro-poor than earlier means-based models had suggested. The PLM-PB had another important advantage in that the method also captured the variation in the use of medication and thus the subsidy received within sub-populations. As the PLM-PB was found to be more accurate than the means-based model, a multivariate analysis of the distribution of PBS subsidy across a number of socio-economic groups was undertaken as an example application of the model. It was found that health status (defined by number of recent illnesses) and concession card type were most important in explaining the amount of PBS subsidy received. This indicates that the distribution of PBS expenditure meets the policy objectives of assisting those most in need, whether need is defined as poor health or low income. 1 Benefits refer to expenditure as transfers from government to individuals rather than the general health benefits of using medication.

Background: Alcohol consumption is one of the most important factors for disease and disability in Europe. In clinical trials, nalmefene has resulted in a significant reduction in the number of heavy-drinking days (HDDs) per month and total alcohol consumption (TAC) among alcohol-dependent patients versus placebo. Methods: A microsimulation model was developed to estimate alcohol-attributable diseases and injuries in patients with alcohol dependence and to explore the clinical relevance of reducing alcohol consumption. Results: For all diseases and injuries considered, the number of events (inpatient episodes) increased with the number of HDDs and TAC per year. The model predicted that a reduction of 20 HDDs per year would result in 941 fewer alcohol-attributable events per 100,000 patients, while a reduction in intake of 3,000 g/year of pure alcohol (ethanol) would result in 1,325 fewer events per 100,000 patients. Conclusion: The potential gains of reducing consumption in alcohol-dependent patients were considerable.

Variable Speed Limits (VSL) are often used to improve traffic conditions on congested motorways. VSL can be implemented as mandatory or advisory. The objective of the thesis isto study in detail the effectiveness of VSL. The focus is on both, design parameters and conditions under which VSL are most effective. The MCS system on the E4 motorway inStockholm is used as a case study. The evaluation was conducted using empirical methods (including aggregate data from microwave sensors and other sources, and disaggregate data from a mobile study), and microscopic traffic simulation. The empirical analysis is based on before and after VSL data, including evaluation of individual measures of performance, and multivariate analysis in the form of the fundamental diagram, and speed-density relationships. The results from the empirical study are mixed with an indication that driver behavior has a strong impact on the effectiveness of the system. The microscopic traffic simulation analysis included the development of a platform for testing VSL and more generally motorway control strategies. The simulation platform was calibrated and validated with the empirical data and includes in addition to VSL, and Automatic Incident Detection (AID) system, the ALINEA ramp metering algorithm. The test-platform allows the testing of different control strategies and various combinations of control strategies, under different scenarios and in a controlled environment. The results from the simulation study indicate that driver compliance is an important factor and VSL performance quickly deteriorates as compliance rate drops. Hence, VSL should be implemented as mandatory instead of advisory. In addition, mandatory VSL can be effective both, under incident and moderately congested conditions. A combined VSL and ramp metering strategy can be most effective in reducing travel time, improving traffic conditions on the motorway. Furthermore, the results indicate that such a strategy also has the least impact on the flows entering the motorway from the ramps. / QC20100630