Introduction
The purpose of this study was to investigate trends and patterns in research papers presented at the annual Health Services Research and Pharmacy Practice conference in the UK. An average of around 50 abstracts are published each year at this conference, attracting most of the principal research centres in the UK and, increasingly, European centres of research in this field. It is argued that this annual collection of peer reviewed papers is generally representative of the output for pharmacy practice research. It could thus be used to chart the progress of the research discipline, in terms of evolving methodology, research topics and presentation style. A historical review could be informative about the changing pace of active research and provide indicators for future trends. We have used a content analytical method, incorporating validity checks, to establish these patterns since the inception of the HSR and Pharmacy Practice conference.

Methods and sample
The sample was all the available abstracts published in the HSR and Pharmacy Practice conference books since 1995. This amounted to 228 peer reviewed abstracts, representing a census of available published research. Each abstract was copied and randomised with a unique identification code. A coding frame was constructed which was able to objectively describe variables such as type of project, type of methodology and statistics, sample sizes and author collaboration, among others. Using this frame, the content of each abstract was independently coded by two analysts, blinded to the year of publication and the other analyst's codes. These two sets of coded variables were cross-tabulated and a kappa value for each variable was computed. The kappa statistic is a measure of non-chance agreement between judges, and was used in this study to reject any variables where independent coding agreement was not present, thus providing statistical confidence in the variables entered for quantitative analysis. In addition to this, an analytic critique based on structural linguistics was used to investigate active verb usage in research titles.

Results and Discussion
The outcome of the quantitative analysis will be presented. The principal results will be based on the overall census data, coupled with comparative analysis by year of publication, anonymised research centre, type of methodology employed and research topic. The authors will provide evidence of historical trends since 1995, which includes the analysis of the style of language used by researchers over this period. The results will provide a basis on which informed discussion of research trends and future prospects can be based.