Abstract:

Objective To construct and validate a risk predictive model for treatment delay of patients with lung cancer. Methods A convenience sampling method was used to select 493 lung cancer patients admitted to a tertiary care hospital in Guangxi from September 2021 to June 2022. They were randomly divided into 345 cases in a modeling group and 148 cases in an internal validation group，and 47 lung cancer patients admitted to another tertiary care hospital were selected as the external validation group. The risk prediction model and nomograms model were constructed using logistic regression analysis and R software，and the area under the receiver operating characteristic（ROC） curve was used to test the prediction effect of the model. Results The established model was Y=e^z/（1+e^z），Z=9.04 × low level of disease perception + 2.01 × irregular physical examination-0.08 × perceived social support scale score + 0.23 × score of perceived barriers to healthcare-seeking decision-Chinese -0.15 × score of Health Literacy Scale for Chronic Patients. The ROC of the internal validation group was 0.849；the sensitivity was 82.39%；the specificity was 78.83%；the positive predictive was 90.61%；the negative predictive was 63.57%. The ROC of the external validation group was 0.830；the sensitivity was 97.36%；the specificity was 80.03%；the positive predictive was 94.62%；the negative predictive was 80.05%. Conclusion The low level of disease perception，lack of regular physical examination，the low level of social support，the high impaired perception of medical seeking behavior，and the low level of health literacy are the influencing factors of treatment delay of lung cancer patients，and the prediction model based on the above factors has a good degree of differentiation and calibration，which can help to identify the high-risk treatment delay lung cancer individuals. Nurses can identify high-risk groups based on this model，so as to take measures to improve their medical behavior and reduce the risk of treatment delay.

Key words: Lung Cancer, Treatment Delay, Prediction Model, Nomograms, Nursing Care

摘要：

目的 构建并验证肺癌患者就医延迟风险预测模型。 方法 2021年9月—2022年6月，采用便利抽样法，选取广西壮族自治区某三级甲等医院收治的493例肺癌患者作为调查对象，将其随机分为建模组345例和内部验证组148例，选取另一所三级甲等医院收治的47例肺癌患者作为外部验证组；采用Logistic回归分析筛选肺癌患者就医延迟的危险因素，使用R软件构建风险预测模型和列线图模型，采用受试者操作特征曲线下面积检验模型预测效果。 结果 最终构建的模型为Y=e^z/（1+e^z），其中Z=9.04 × 低疾病感知水平+2.01 × 未定期体检-0.08 × 领悟社会支持量表得分+0.23 × 就医行为感知障碍量表得分-0.15 × 慢性病患者健康素养量表得分。内部验证组受试者操作特征曲线下面积为0.849，灵敏度为82.39%，特异度为78.83%，阳性预测值为90.61%，阴性预测值为63.57%。外部验证组受试者操作特征曲线下面积为0.830，灵敏度为97.36%，特异度为80.03%，阳性预测值为94.62%，阴性预测值为80.05%。 结论 低疾病感知水平、未定期体检、领悟社会支持水平低、就医行为感知障碍高、健康素养水平低是肺癌患者就医延迟的危险因素，该研究构建的预测模型具有良好的区分度和校准度，能够帮助护士筛查肺癌就医延迟高危人群，采取合适的干预措施改善其就医行为，降低就医延迟风险。

关键词: 肺癌, 就医延迟, 预测模型, 列线图, 护理