Factors associated with non-use of inpatient hospital care service by elderly people in China.

The objectives of this cross-sectional study were to estimate non-use of inpatient hospital care service by elderly people in China and to examine associations between service non-use during the past 12 months and an array of predisposing, enabling and need factors. Using a fully structured questionnaire, trained health personnel interviewed 4046 Chinese aged 60 and older in Zhejiang province selected by a two-stage stratified cluster sampling scheme between September and December 2007. Based on the Andersen behavioural model, hierarchical logistic regression analysis was used to explore the predictors of non-use of this service. The rate of non-use was 14.2% [95% confidence interval (CI), 13.1-15.3%] for inpatient hospital care service. Logistic regression analyses revealed that enabling factors were more important than either predisposing or need factors in predicting non-use of inpatient hospital care service.Predisposing factors other than education were not significant, and only the need factor of number of diseases was significant for non-use of inpatient hospital care service. The odds of non-use for those with a college or higher degree were 0.36 times (95% CI, 0.21-0.62) the odds for those with primary or lower education. The odds of non-use in the presence of 4-10 diseases and 1-3 diseases were 3.10 times (95% CI,1.96-4.89) and 2.14 times (95% CI, 1.45-3.14), respectively, of those having no disease. Among the four enabling factors, only degree of living satisfaction score was not significantly associated with non-use of inpatient hospital care service. For elderly persons with higher healthcare expenditure, joining new rural cooperative medical insurance or having low social support, the odds of reporting non-use of inpatient hospital care services were raised by a factor of 1.6-8.0. Findings indicated that, in the absence of universal and comprehensive medical insurance coverage,enabling factors are more important than either predisposing or need factors in predicting non-use.