Department of Psychiatric, The Fourth People's Hospital of Zhangjiagang City, Suzhou, Jiangsu, China.
Soochow University Affiliated Guangji Hospital, Suzhou, China.
BMJ Open. 2024 Nov 14;14(11):e087312. doi: 10.1136/bmjopen-2024-087312.
Treatment compliance among psychiatric patients is related to disease outcomes. How to assess patient compliance remains a concern. Here, we established a predictive model for medication compliance in patients with psychotic disorders to provide a reference for early intervention in treatment non-compliance behaviour.
Clinical information for 451 patients with psychotic disorders was downloaded from the Dryad database. The Least Absolute Shrinkage and Selection Operator regression and logistic regression were used to establish the model. Bootstrap resampling (1000 iterations) was used for internal validation and a nomogram was drawn to predict medication compliance. The consistency index, Brier score, receiver operating characteristic curve and decision curve were used for model evaluation.
35 Italian Community Psychiatric Services.
451 patients prescribed with any long-acting intramuscular (LAI) antipsychotic were consecutively recruited, and assessed after 6 months and 12 months, from December 2015 to May 2017.
432 patients with psychotic disorders were included for model construction; among these, the compliance rate was 61.3%. The Drug Attitude Inventory-10 (DAI-10) and Brief Psychiatric Rating Scale (BPRS) scores, multiple hospitalisations in 1 year and a history of long-acting injectables were found to be independent risk factors for treatment noncompliance (all p<0.01). The concordance statistic of the nomogram was 0.709 (95% CI 0.652 to 0.766), the Brier index was 0.215 and the area under the ROC curve was 0.716 (95% CI 0.669 to 0.763); decision curve analysis showed that applying this model between the threshold probabilities of 44% and 63% improved the net clinical benefit.
A low DAI-10 score, a high BPRS score, multiple hospitalisations in 1 year and the previous use of long-acting injectable drugs were independent risk factors for medication noncompliance in patients with psychotic disorders. Our nomogram for predicting treatment adherence behaviour in psychiatric patients exhibited good sensitivity and specificity.
精神疾病患者的治疗依从性与疾病结局有关。如何评估患者的依从性仍然是一个关注点。在这里，我们建立了一个预测精神障碍患者药物依从性的模型，为治疗不依从行为的早期干预提供参考。
从 Dryad 数据库中下载了 451 例精神障碍患者的临床信息。使用最小绝对收缩和选择算子回归和逻辑回归建立模型。Bootstrap 重采样（1000 次迭代）用于内部验证，并绘制诺模图预测药物依从性。一致性指数、Brier 评分、接收者操作特征曲线和决策曲线用于模型评估。
35 家意大利社区精神病服务机构。
从 2015 年 12 月至 2017 年 5 月，连续招募了 451 例服用任何长效肌内（LAI）抗精神病药物的患者，并在 6 个月和 12 个月时进行评估。
432 例精神障碍患者纳入模型构建；其中，依从率为 61.3%。药物态度量表-10（DAI-10）和简明精神病评定量表（BPRS）评分、1 年内多次住院和长效注射剂使用史被发现是治疗不依从的独立危险因素（均 p<0.01）。诺模图的一致性统计量为 0.709（95%置信区间 0.652 至 0.766），Brier 指数为 0.215，ROC 曲线下面积为 0.716（95%置信区间 0.669 至 0.763）；决策曲线分析表明，在阈值概率为 44%至 63%之间应用该模型可提高净临床获益。
低 DAI-10 评分、高 BPRS 评分、1 年内多次住院和之前使用长效注射药物是精神障碍患者药物依从性的独立危险因素。我们预测精神科患者治疗依从行为的诺模图具有良好的敏感性和特异性。
