Cluster Randomized Trial of a Personalized Clinical Decision Support Intervention to Improve Statin Prescribing in Patients With Atherosclerotic Cardiovascular Disease.
Cluster Randomized Trial of a Personalized Clinical Decision Support Intervention to Improve Statin Prescribing in Patients With Atherosclerotic Cardiovascular Disease.
复制标题
个性化临床决策支持干预改善动脉粥样硬化性心血管疾病患者他汀类药物处方的整群随机试验。
DOI：
10.1161/circulationaha.123.064226
复制
发表时间：
2023
期刊：
影响因子：
37.8
通讯作者：
Matheny,MichaelE
中科院分区：
文献类型：
--
作者：
Virani,SalimS;Ramsey,DavidJ;Westerman,Dax;Kuebeler,MarkK;Chen,Liang;Akeroyd,JuliaM;Gobbel,GlennT;Ballantyne,ChristieM;Petersen,LauraA;Turchin,Alexander;Matheny,MichaelE
他汀类药物和高强度他汀类药物（HIS）在动脉粥样硬化性心血管疾病（ASCVD）患者中的使用率仍然很低。他汀类药物相关副作用（SASE）和治疗惰性都起作用。1我们评估了个性化提醒是否能改善ASCVD患者HIS的使用。在这项在退伍军人事务部进行的随机分组对照试验中，我们测试了一项4年多来开发的干预措施，该干预措施通过自然语言处理4使用结构化3和非结构化数据构建算法，以识别SASE，并通过进行定性访谈，了解患者对SASE的看法和临床医生的信息需求。5利用这一点，开发了一种干预措施，其中包括由研究团队在一个位置处理的提醒，这些提醒针对每位患者进行个性化处理，并在下次就诊前2至7天（同步提醒）或在初级保健就诊外（异步提醒）发送给各自的初级保健临床医生。提醒信息包括ASCVD诊断的日期和类型（缺血性心脏病、外周动脉疾病或缺血性卒中）、他汀类药物和剂量、最后一次填写日期、SASE的日期和类型以及HIS定义和SASE管理的指南资源。为了防止警报疲劳，我们的算法确保临床医生在收到更多提醒之前没有来自该提醒的> 3个未签名警报。干预站点的临床医生可以选择不接收提醒。家庭护理包括临床医生访问显示他汀类药物治疗依从性的患者仪表板。每月更新队列，以纳入更新的他汀类药物剂量、新的排除（转移性癌症、临终关怀、姑息治疗或死亡）和新的SASE。获得了机构伦理审查委员会的伦理批准。研究数据可根据合理要求从相应作者处获得。在指南教育之后，我们随机分配了27个初级保健诊所（36641例患者）：14个进行干预（117名临床医生和18427例患者），13个进行常规护理（128名临床医生和18214例患者）。结局包括干预和常规护理中心之间HIS（主要）和他汀类药物（次要）使用变化前后的结局。由于我们预期提醒将使临床医生能够引起患者对他汀类药物的担忧，因此我们评估了两组之间他汀类药物依从性（使用覆盖天数比例≥ 0.8）前后的差异。审判于2021年8月开始，于2022年11月结束。平均年龄为71.1岁，主要的ASCVD表型包括缺血性心脏病（77.5%的患者）。在干预组的患者中，41.6%的患者在结构化数据或自然语言处理中有与SASE相关的信号。
Statin and high-intensity statin (HIS) use remains low in patients with atherosclerotic cardiovascular disease (ASCVD). 1, 2 Both statin-associated side effects (SASEs) and therapeutic inertia play a role. 1 We evaluated whether personalized reminders improve HIS use in patients with ASCVD. In this cluster-randomized controlled trial performed in the Department of Veterans Affairs, we tested an intervention developed over 4 years by constructing algorithms using structured3 and unstructured data through natural language processing4 to identify SASEs and, by performing qualitative interviews, to understand patient perspectives on SASEs and clinician information needs. 5 Leveraging this, an intervention was developed that included reminders processed by the research team at one location individualized to each patient and sent to their respective primary care clinicians 2 to 7 days before their next visit (synchronous reminders) or outside of the primary care visit (asynchronous reminders). Information on reminders included date and type of ASCVD diagnosis (ischemic heart disease, peripheral artery disease, or ischemic stroke), statin and dose, date of last fill, date and type of SASE, and guideline resources on HIS definition and SASE management. To prevent alert fatigue, our algorithms ensured that clinicians did not have> 3 unsigned alerts from this reminder before receiving more reminders. Clinicians at the intervention sites could opt out from receiving reminders. Usual care included clinician access to a patient dashboard displaying compliance with statin therapy. The cohort was updated monthly to incorporate updated statin dose, new exclusions (metastatic cancer, hospice, palliative care, or death), and new SASEs. Ethical approval was obtained from the institutional ethics review committee. Study data are available from the corresponding author on reasonable request. Following guideline education, we randomly assigned 27 primary care clinics (36 641 patients): 14 to the intervention (117 clinicians and 18 427 patients) and 13 to usual care (128 clinicians and 18 214 patients). Outcomes included before and after changes in HIS (primary) and statin (secondary) use between intervention and usual care sites. Because we expected that the reminder would allow clinicians to elicit patients’ concerns about statins, we evaluated before and after statin adherence (using proportion of day covered≥ 0.8) difference between 2 groups. The trial began in August 2021 and ended in November 2022. Mean age was 71.1 years, and the predominant ASCVD phenotype included ischemic heart disease (77.5% patients). Of the patients in the intervention arm, 41.6% had a signal related to SASEs on either structured data or natural language processing.