Objective: To estimate real world treatment effects of parenterally administered Shuxuetong on abnormal changes in blood urea nitrogen (BUN) index, inorder to to explore whether parenterally administered Shuxuetong affects kidney function and to provide clinical guidance for its safe application.
Method: Analyze hospital information system (HIS) data and laboratory information management system (LIS) data from 18 nationawide general hospitals in China. A retrospective cohort study with two cohorts of patients, assessing abnormal changes of BUN index as an outcome, confounding factors are balanced through use of generalized boosted models (GBM) and multiple logistic regression based on adjusted propensity score to estimate treatment effects of parenterally administered Shuxuetong on abnormal changes of BUN index and to explore possible influencing factors.
Result: Logistic regression based on adjusted propensity score results showed that there was a statistically significant difference (P < 0.05), in abnormal BUN in rates between the two groups. However, the GBM propensity score method showed that abnormal BUN rates were not statistically significantly different between the two groups.
Conclusions: Based on large sample clinical observational data from HIS and LIS database, significant effects of parenterally administered Shuxuetong on abnormal changes of BUN have not been found. The generalized boosted models (GBM) method is more suitable for retrospective non-randomized observational data. These results should be confirmed through largescale prospective randomized controlled trials.