The objective of this thesis was to identify statistical issues that are commonly associated with evaluations of services for older people with a view to establishing the most appropriate methods of addressing them. This goal was achieved in two stages. In the first stage, a comprehensive literature review of studies that have reported such evaluations on populations of older people in the UK was conducted. The second stage involved demonstrating approaches for dealing with these issues on a dataset drawn from largest evaluation of intermediate care done and published in the UK to date. The approaches were adapted from the studies reported in the literature review and where appropriate, from other sources. This thesis identified a number of statistical issues including those associated with distributional characteristics of variables, missing data and the need to predict utility outcome measures from non-utility ones. Robust approaches of dealing with these problems were demonstrated. The results obtained underlined the importance of avoiding erroneous results and conclusions by applying methods with a sound theoretical background.