Abstract
Background: To overcome the drawback of individual item-by-item box plots of disclosure for patient views on healthcare service quality, we propose to inspect interrelationships among items that measure a common entity. A visual diagram on the Internet is developed to provide thorough information for hospitals. Methods: We used the Rasch rating scale model to analyze the 2003 English inpatient questionnaire data regarding patient satisfactory perception, which were collected from 169 hospitals, examined model-data fit, and developed a KIDMAP diagram on the Internet depicting the satisfaction level of each hospital and investigating aberrant responses with Z-scores and MNSQ statistics for individual hospitals. Differential item functioning (DIF) analysis was conducted to verify construct equivalence across types of hospitals. Results: 18 of the 45 items fit to the model's expectations, indicating they jointly defined a common construct and an equal-interval logit scale was achieved. The most difficult aspect for hospitals to earn inpatients' satisfaction were item 29 (staff told you about any medication side effects to watch when going home). No DIF in the 18-item questionnaire was found between types of hospitals, indicating the questionnaire measured the same construct across hospitals. Different types of hospitals obtained different levels of satisfaction. The KIDMAP on the Internet provided more interpretable and visualized message than traditional item-by-item box plots of disclosure. Conclusion: After removing misfit items, we find that the 18-item questionnaire measures the same construct across types of hospitals. The KIDMAP on the Internet provides an exemplary comparison in quality of healthcare. Rasch analysis allows intra- and inter-hospital performances to be compared easily and reliably with each other on the Internet. Copyright © 2009 Chien et al; licensee BioMed Central Ltd.
Original language | English |
---|---|
Article number | 135 |
Journal | BMC Health Services Research |
Volume | 9 |
Issue number | 1 |
DOIs | |
Publication status | Published - Jul 2009 |