A graph-based machine-learning clinical decision support system did not significantly reduce blood pressure or overall economic burden compared to control (P > 0.05).
RCT (n=430)
Cluster-randomized
No
Does a machine-learning based clinical decision support system improve blood pressure management and reduce economic burden in hospitalized hypertension patients?
A machine-learning based clinical decision support system improved antihypertensive prescribing and reduced medical costs for surgical patients with hypertension, though overall blood pressure reduction was not significantly different.
p-value: p=>0.05
Objectives: To assess the effects of clinical decision support system of graph based machine-learning algorithms on the blood pressure management and economic burden of disease. Methods: A cluster-randomized trial was conducted in the 3rd Xiangya Hospital among 43 physicians and 430 of their hypertension patients. Patients were followed during hospitalization. A clinical decision support system (CDSS) based on the graph based machine-learning algorithms was integrated within hospital information system and was activated for physicians in the intervention group. Results: During hospitalization, the intervention of CDSS could improve antihypertensice prescriptions. And there was no difference between CDSS intervention group and controlled group in reducing blood pressure and improving the economic burden of disease (P > 0.05); The benefit-cost ratio of CDSS was 1.15, and the net present value of benefit cost was 5,792 yuan. The results of subgroup analysis showed that, the intervention of CDSS could improve antihypertensice prescriptions and reduce the direct medical costs (Intervention group = 43,467 ± 39,716 yuan vs. Controlled group = 61,205 ± 66,576 yuan, P = 0.048) and the economic burden of disease (Intervention group = 46,006 ± 40,831 yuan vs. Controlled group = 64,192 ± 67,968 yuan, P = 0.048) among hypertension patients who were treated in surgical department; The benefit-cost ratio of CDSS was 1.44, and the net present value of benefit cost was 18,186 yuan Conclusion: A clinical decision support system based on the graph based machine-learning algorithms changed the antihypertensive prescriptions and reduced the medical expense among hypertension patients.
Xing Liu (Mon,) conducted a rct in Hypertension (n=430). Clinical decision support system (CDSS) based on graph-based machine-learning algorithms vs. Controlled group (usual care) was evaluated on Reducing blood pressure and improving the economic burden of disease (p=>0.05). A graph-based machine-learning clinical decision support system did not significantly reduce blood pressure or overall economic burden compared to control (P > 0.05).