Self-rated health, chronic diseases and health service utilisation in Hong Kong

Introduction Self-rated health (SRH) is a widely used indicator of health service utilisation and reflects self-perceived objective health condition. Poorer non-comparative SRH was shown to be related to higher inpatient and outpatient utilisation in Western and elderly populations. Little is known about how healthcare utilisation relates to SRH in non-Western settings, such as Hong Kong and in adult populations. The association of age- and time- comparative SRH with healthcare utilisation is also unclear. This study aimed to assess the association of three types of SRH (non-, age- and time- comparative SRH) with inpatient and outpatient utilisation in Hong Kong‟s general populaion. Methods Data were derived from 2011 Thematic Household Survey (THS), covering 23,892 non-institutional residents aged 20 and above. The study adopts Andersen‟s Behavioral Model of Health Service Use for the analytical framework. Healthcare utilisation was measured by inpatient use during the past year and outpatient use (including General Outpatient Clinic (GOPC) and Specialist Outpatient Clinic (SOPC)) during the past month, in terms of ever-use and the amount of use (bed-days and number of outpatient visits). SRH was measured with a 5-point Likert Scale: non-comparative SRH from “Excellent” to “Poor”; age- and time- comparative from “much worse” to “much better”. Logistic regression and zero-truncated negative binomial/ Poisson regression were applied to examine the association of SRH and chronic diseases with healthcare utilisation in the public and private sector separately as per the Andersen behavioral model. Results “Fair/ poor” non-comparative SRH was associated with higher inpatient and outpatient utilisation. The association was not significant for hospital bed-days. Similarly, age-comparative SRH was associated with inpatient (except private bed-days) and outpatient utilisation (except the number of SOPC visits). “Worse/ much worse” time-comparative SRH was associated with higher healthcare utilisation, but the relationship was less clear for private hospitalisation. The presence of cancer, cardiovascular diseases, diabetes, lower respiratory diseases, and musculoskeletal diseases were associated with higher healthcare utilisation, with stronger association observed for ever-use than the amount of use. The relationships between musculoskeletal diseases and inpatient utilisation, between cardiovascular diseases and diabetes and the number of private outpatient visits, and between lower respiratory diseases and GOPC utillisation were not significant. Conclusions The present study suggests SRH to be a useful health indicator of health service utilisation. All three SRH measures were associated with health service utilisation and no marked differences were observed between different measures. Poorer SRH were strongly related to higher public inpatient utilisation, with stronger association observed for ever hospitalisation than bed-days. Poorer SRH measures were also related to higher outpatient uilisation in both sectors during the past month. All the selected chronic conditions were related to increased healthcare use. The associations were less clear for hospital bed-days and the private sector. Future studies should focus on the predictive validity of SRH on future healthcare utilisation.published_or_final_versionPublic HealthMasterMaster of Philosoph