[Study on the results of treating tuberculosis inpatient in the general hospitals: a correspondence analysis]

To understand the current status of treatment among inpatients of tuberculosis (TB) in general hospitals, and to related to different attrributes.

A retrospective study was designed for inpatients with TB who were discharged from general hospitals in Nachong region, 2003. Factors associated with the results of treatment were selected, using Chi-square test. Further correspondence analysis (CA) was used to visualize the relationship between attributes of inpatients and results of treatment in general hospitals.

Statistically significant factors associated with treatment results would include gender, age, state of illness at access to hospital, comorbidity and length of stay, while ways of payment, occupation and marital status were not statistically significant. The joint plot of CA showed results as follows: (1) Attributes of inpatients died in general hospitals were clearly different from that of cured or improved inpatients. (2) Result of hospitalization on treatment was more likely to be 'improved' for TB inpatients who were male, aged > or = 15 yrs, and with urgent condition when administered into the hospitals. (3) Result of cure was likely to be seen among inpatients who were female, length of stay >8 days, with no comorbidity, and with average illness state when accessing to hospitals. (4) Bad treatment results were appeared for inpatients younger than 15 yrs, with critical state when administered to hospitals.

(1)CA provided us with a new way on how to extract useful information from miscellaneous data of the patients. (2) The relationships between the results of treatment from the general hospitals and TB inpatients' attributes might provide tips to develop a series of corresponding strategies for treating TB inpatients with special attributes in order to obtain higher cure rate.