基于风险前馈控制的冠脉介入操作术后并发症风险预测模型构建

Construction of a Postoperative Complication Risk Prediction Model for Coronary Intervention Procedures Based on Risk Feed-Forward Control

摘要: 目的：构建基于风险前馈控制的冠脉介入操作术后并发症风险预测模型。方法：本研究是一项回顾性研究，回顾性队列研究分析2023年3月~2025年2月入住延安市安塞区人民医院心血管内科行冠脉介入操作的病例，对照组：CAG/PCI病例未发生并发症组；观察组：CAG/PCI病例出现并发症组。两组组间比较计量资料采用独立样本t检验或非参数检验，计数资料比较采用c²检验，比较后有统计学意义的纳入二元Logistic回归模型，CAG/PCI术后出现并发症作为因变量，采用二元Logistic回归筛选出现并发症的独立危险因素，根据二元Logistic结果构建风险前馈评分系统(RFS)，以P < 0.05为差异具有统计学意义。结果：本研究共纳入320例于2023年3月2025年2月在我院行冠脉介入操作术后患者，二元Logistic回归分析，结果显示合并高血压(OR = 2.566)、婚姻状况(OR = 3.074)、心功能(LVEF) (OR = 5.623)、病变血管狭窄程度(OR = 4.329)，4个因素进入方程(P < 0.05)，由标准化偏回归系数计算冠脉介入操作术后患者出现并发症风险值(R) = 2 × 高血压 + 3 × 婚姻状况 + 5 × 心功能(LVEF) + 4 × 病变血管狭窄程度。结论：风险前馈控制强调在并发症发生前基于预测进行干预，区别于传统反馈控制，本研究中OR值排序：按效应强度降序呈现(LVEF > 狭窄 > 婚姻 > 高血压)，突出临床优先级。该模型通过整合生理–社会双维变量，涉及对社会支持薄弱人群的精准预警。建议临床：将婚姻状况纳入术前风险评估表，对LVEF < 40%合并社会支持不足者启动多学科干预，同时依托模型开发实时风险预警系统，推动并发症防控从“被动处置”向“主动前馈”转型。

Abstract: Objective: To construct a risk prediction model for postoperative complications of coronary intervention based on risk feed-forward control. Methods: This study was a retrospective study. A retrospective cohort study was conducted on cases of coronary intervention performed in the Department of Cardiovascular Medicine of Ansai District People’s Hospital, Yan’an City from March 2023 to February 2025. The control group consisted of cases of CAG/PCI without complications, and the observation group consisted of cases of CAG/PCI with complications. Independent sample t-test or non-parametric test was used for comparison of measurement data between the two groups, and chi-square test was used for comparison of count data. Variables with statistical significance after comparison were included in the binary Logistic regression model. Postoperative complications of CAG/PCI were taken as the dependent variable, and binary Logistic regression was used to screen independent risk factors for complications. A risk feed-forward scoring system (RFS) was constructed based on the results of binary Logistic regression. A difference was considered statistically significant when P < 0.05. Results: A total of 320 patients who underwent coronary intervention procedures in our hospital from March 2023 to February 2025 were included in this study. Binary Logistic regression analysis was conducted, and the results showed that four factors, namely, hypertension (OR = 2.566), marital status (OR = 3.074), cardiac function (LVEF) (OR = 5.623), and the degree of stenosis of the lesion vessel (OR = 4.329), entered the equation (P < 0.05). The risk value (R) of complications after coronary intervention procedures was calculated by the standardized partial regression coefficient as R = 2 × hypertension + 3 × marital status + 5 × cardiac function (LVEF) + 4 × degree of stenosis of the lesion vessel. Conclusion: Risk feed-forward control emphasizes intervention based on prediction before complications occur, which is different from traditional feedback control. In this study, the OR values were ranked in descending order of effect strength (LVEF > stenosis > marital status > hypertension), highlighting clinical priorities. This model integrates physiological and social dimensions, involving precise early warning for people with weak social support. Clinical recommendation: Marital status should be included in the preoperative risk assessment form. For those with LVEF < 40% and insufficient social support, multidisciplinary intervention should be initiated. At the same time, a real-time risk early warning system should be developed based on the model to promote the transformation of complication prevention and control from “passive response” to “active feed-forward”.