Objectives: This study has been conducted to determine the levels of inappropriate use of the A&E for conditions that could be treated by GPs, the nature of the morbidity pattern of those conditions, the reasons why primary care services were not being utilized, and also examined the validity (i.e. sensitivity and specificity) of patient classifications undertaken by nurses at the time of admission within this local context. Study design and setting: A cross sectional study was conducted over a one year period and subjects were randomly selected from four A&E departments located across the four principle geographic regions of Hong Kong by stratified, two-stage sampling. Main outcome measure: The gold standard in differentiating true emergency cases and GP cases was based on a retrospective record review conducted independently by a panel of emergency physicians. A random sub-sample of those classified as GP cases was interviewed and compared to a matched (via morbidity status) sample of primary care patients who had attended a hospitals' GOPC in order to determine factors distinguishing these two patient groups. Multiple Logistic Regression was used to distinguish the difference between GP cases and matched GOPC primary care patients on significance and odds ratios of the variables. The morbidity pattern according to ICPC was tabulated and analysed for the 'true' A&E cases and non-urgent cases. Sensitivity, specificity and positive predictive values were computed for both non-weighted and weighted conditions. Results: The level of GP cases was found to be 57% with a significant higher proportion of patients in younger age group, and late evening. The morbidity pattern of those top 10 diagnoses of non-urgent cases was very similar to the Hong Kong general practice morbidity pattern for self limiting conditions. Closure of the clinic was the main reason for GP cases attending A&E. Other major reasons were deterioration of symptoms, GPs' inability to diagnose efficiently and patients' wish to continue medical treatment in the same hospital. Affordability was the most pronounced reason for utilising the GOPC, but did not apply to the A&E GP patients. The most accurate weighted nurses' triage classification had the average sensitivity of 75%, specificity of 65.7%, and positive predictive value of 54%. The most accurate weighted patients' self-triage classification yielded a sensitivity of 43.3%, specificity of 49.2%, and a positive predictive value of 38.6%. Conclusion: The reasons for high level of utilisation of A&E services are complex and reflect problems of delivery of GP services. There is an urgent need for the GPs to set up a network system to provide out of hours services, and also for a better interfacing between primary and secondary care, and between public and private sectors, so that patients can be referred back to GPs. The design and measures chosen for this study will help provide A&E policy makers and planners with relevant information for better addressing practical solutions. / Albert Lee. / Source: Dissertation Abstracts International, Volume: 73-01, Section: B, page: . / Thesis (M.D.)--Chinese University of Hong Kong, 2004. / Includes bibliographical references (leaves 137-151). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [201-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Appendix 4 in Chinese.