Use propensity score method to analyze data about effectiveness of coronary heart disease treated by kudiezi injection / 中国中药杂志

ABSTRACT

<p><b>OBJECTIVE</b>Real world clinical research to evaluate the effect on coronary heart disease (CHD) of parenterally administered Kudiezi.</p><p><b>METHOD</b>Eighteen nationwide general hospitals' hospital information systems (HIS) were searched for information on in patients treated with parenterally administered Kudiezi. Evaluation used generalized boosted models (GBM), and three types of logistic regression based on adjusted propensity score to balance confounding factors.</p><p><b>RESULT</b>Two groups of coronary heart patients were compared, one group received parenterally administered Kudiezi and one did not. Seventy two confounding factors were identified and balanced by GBM. The three methods of logistic regression demonstrated that patients receiving parenterally administered Kudiezi had improved outcomes compared to those who did not receive the drug and this was statistically significant.</p><p><b>CONCLUSION</b>Parenterally administered Kudiezi is a potential adjuvant therapy in the treatment of coronary heart disease. Logistic regression based on adjusted propensity score can balance confounding factors in retrospective data analysis. However it's limitation is that it is not applicable in prospective studies.</p>