JMIR:开发机器学习模型来预测严重的慢性阻塞性肺疾病恶化

2022-03-06 从医路漫漫 MedSci原创

在美国,慢性阻塞性肺病 (COPD) 影响 6.5% 的成年人,是除 COVID-19 之外的第四大死亡原因。

背景:在美国,慢性阻塞性肺病 (COPD) 影响 6.5% 的成年人,是除 COVID-19 之外的第四大死亡原因。每年,COPD 导致 150 万次急诊 (ED) 就诊、70 万次住院和 321 亿美元的总医疗费用。重度 COPD 加重是指需要 ED 就诊或住院,占与 COPD 相关的总医疗费用的 90.3%,并且通常会导致肺功能和健康状况不可逆转的下降。许多严重的 COPD 恶化(例如,47% 的 COPD 住院患者)被认为可以通过适当的门诊护理来预防,因为 COPD 是一种对门诊护理敏感的疾病。减少 COPD 严重恶化的常用方法是将患者置于高风险的预防性护理管理计划中。可以使用预测模型前瞻性地识别高风险患者。一旦患者进入护理管理计划,护理经理将定期联系患者进行健康状况评估并帮助协调健康和相关服务。这种方法被许多健康计划采用,例如12个大都市社区中的 9 个以及许多医疗保健系统。成功的护理管理可以减少高达 27% 的 ED就诊次数和 40% 的 COPD 患者住院时间。

然而,由于资源和服务能力的限制,只有 ≤3% 的患者可以进入护理管理计划。其有效性的上限取决于这些患者的风险水平,这取决于所使用的预测模型的准确性。单独的 COPD 阶段和既往有严重 COPD 恶化都不能很好地预测患者未来严重 COPD 恶化的风险水平。此前,研究人员已经建立了几个模型来预测 COPD 患者的严重 COPD 恶化。这些模型不准确且不适合用于护理管理,因为它们错过了超过 50% 的未来将经历严重 COPD 恶化的患者,错误地预测了许多其他患者将经历严重的 COPD 恶化,使用常规临床实践中没有数据,或者是为与典型 COPD 患者具有不同特征的患者设计的。此外,这些模型中的大多数仅预测 COPD 的住院时间。为了更好地指导护理管理的使用,我们需要预测COPD 的 ED 就诊次数和住院时间,而这些模型中只有2个可以做到这一点。在实践中,一旦为护理管理部署模型,该模型产生的预测误差将导致患者预后下降和不必要的医疗保健成本。由于患有慢性阻塞性肺病的患者数量众多,即使模型准确性的微小改进加上适当的预防性干预措施,也可以帮助改善结果并避免每年因慢性阻塞性肺病而多次就诊和住院。

目的:本研究的目的是开发一个更准确的模型来预测严重的 COPD 恶化。方法:我们检查了 2011 年至 2019 年期间访问华盛顿大学医学设施的所有 COPD 患者,并确定了 278 个候选特征。通过对 2011 年至 2019 年的 43,576 个华盛顿大学医学数据实例进行二次分析,我们创建了一个机器学习模型来预测 COPD 患者明年的严重 COPD 恶化。

结果:最终模型的受试者工作特征曲线下面积为 0.866。当使用预测风险最大的前 9.99% (752/7529) 的患者设置二元分类的截止阈值时,模型获得了 90.33% (6801/7529) 的准确度,56.6% (103/ 182),特异性为 91.17% (6698/7347)。

表 1. 数据实例和不良结果随时间的分布。

表 2 主要分析训练集中数据实例的患者特征

图 1 主分析中最终模型的受试者工作特征曲线。

表 3. 在主要分析中,最终模型在使用不同的截止阈值进行二元分类方面的性能测量。

表 6. 使用预测风险最大的前 9.99% (794/7944) 的患者设置二元分类的截止阈值时,主分析中最终模型的混淆矩阵。

表 7 主分析中最终模型的性能和性能稳定性分析中的模型

结论:与之前发表的模型相比,我们的模型提供了对明年严重 COPD 恶化的更准确的预测。在进一步改进其绩效指标后(例如,通过添加从临床记录中提取的特征),我们的模型可用于决策支持工具,以指导识别 COPD 患者和高风险的护理管理以改善结果。

原文出处: Zeng S,  Arjomandi M,  Tong Y, et al.Developing a Machine Learning Model to Predict Severe Chronic Obstructive Pulmonary Disease Exacerbations: Retrospective Cohort Study.J Med Internet Res 2022 01 06;24(1)

版权声明:
本网站所有内容来源注明为“梅斯医学”或“MedSci原创”的文字、图片和音视频资料,版权均属于梅斯医学所有。非经授权,任何媒体、网站或个人不得转载,授权转载时须注明来源为“梅斯医学”。其它来源的文章系转载文章,或“梅斯号”自媒体发布的文章,仅系出于传递更多信息之目的,本站仅负责审核内容合规,其内容不代表本站立场,本站不负责内容的准确性和版权。如果存在侵权、或不希望被转载的媒体或个人可与我们联系,我们将立即进行删除处理。
在此留言
评论区 (5)
#插入话题
  1. [GetPortalCommentsPageByObjectIdResponse(id=1688363, encodeId=e291168836367, content=<a href='/topic/show?id=fab46121efa' target=_blank style='color:#2F92EE;'>#机器学习模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=34, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61217, encryptionId=fab46121efa, topicName=机器学习模型)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=2bea28558433, createdName=ms2694732865965676, createdTime=Wed Nov 23 22:26:49 CST 2022, time=2022-11-23, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1930859, encodeId=5681193085926, content=<a href='/topic/show?id=22f61180211' target=_blank style='color:#2F92EE;'>#miR#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=11802, encryptionId=22f61180211, topicName=miR)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=0a27288, createdName=smallant2002, createdTime=Thu Jun 30 06:26:49 CST 2022, time=2022-06-30, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1697000, encodeId=2e8b169e00094, content=<a href='/topic/show?id=d2e99e651e4' target=_blank style='color:#2F92EE;'>#阻塞#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=28, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=97651, encryptionId=d2e99e651e4, topicName=阻塞)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=a23129903429, createdName=tomyang96, createdTime=Mon Jan 09 19:26:49 CST 2023, time=2023-01-09, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1199860, encodeId=165c1199860b9, content=机器学习模型……以前都是互联网+,现在流行AI+, beContent=null, objectType=article, channel=null, level=null, likeNumber=63, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20210127/50a003c4097845e182c08afd64d8a509/bd766dfebe9c40e0bca2dae964601bdc.jpg, createdBy=22de5454917, createdName=carrotlyl, createdTime=Sun Mar 06 10:43:51 CST 2022, time=2022-03-06, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1583358, encodeId=6817158335822, content=<a href='/topic/show?id=f6e1611840d' target=_blank style='color:#2F92EE;'>#机器#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61184, encryptionId=f6e1611840d, topicName=机器)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=13dc16973127, createdName=ms6279672939590805, createdTime=Thu Mar 03 08:26:49 CST 2022, time=2022-03-03, status=1, ipAttribution=)]
  2. [GetPortalCommentsPageByObjectIdResponse(id=1688363, encodeId=e291168836367, content=<a href='/topic/show?id=fab46121efa' target=_blank style='color:#2F92EE;'>#机器学习模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=34, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61217, encryptionId=fab46121efa, topicName=机器学习模型)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=2bea28558433, createdName=ms2694732865965676, createdTime=Wed Nov 23 22:26:49 CST 2022, time=2022-11-23, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1930859, encodeId=5681193085926, content=<a href='/topic/show?id=22f61180211' target=_blank style='color:#2F92EE;'>#miR#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=11802, encryptionId=22f61180211, topicName=miR)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=0a27288, createdName=smallant2002, createdTime=Thu Jun 30 06:26:49 CST 2022, time=2022-06-30, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1697000, encodeId=2e8b169e00094, content=<a href='/topic/show?id=d2e99e651e4' target=_blank style='color:#2F92EE;'>#阻塞#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=28, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=97651, encryptionId=d2e99e651e4, topicName=阻塞)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=a23129903429, createdName=tomyang96, createdTime=Mon Jan 09 19:26:49 CST 2023, time=2023-01-09, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1199860, encodeId=165c1199860b9, content=机器学习模型……以前都是互联网+,现在流行AI+, beContent=null, objectType=article, channel=null, level=null, likeNumber=63, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20210127/50a003c4097845e182c08afd64d8a509/bd766dfebe9c40e0bca2dae964601bdc.jpg, createdBy=22de5454917, createdName=carrotlyl, createdTime=Sun Mar 06 10:43:51 CST 2022, time=2022-03-06, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1583358, encodeId=6817158335822, content=<a href='/topic/show?id=f6e1611840d' target=_blank style='color:#2F92EE;'>#机器#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61184, encryptionId=f6e1611840d, topicName=机器)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=13dc16973127, createdName=ms6279672939590805, createdTime=Thu Mar 03 08:26:49 CST 2022, time=2022-03-03, status=1, ipAttribution=)]
    2022-06-30 smallant2002
  3. [GetPortalCommentsPageByObjectIdResponse(id=1688363, encodeId=e291168836367, content=<a href='/topic/show?id=fab46121efa' target=_blank style='color:#2F92EE;'>#机器学习模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=34, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61217, encryptionId=fab46121efa, topicName=机器学习模型)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=2bea28558433, createdName=ms2694732865965676, createdTime=Wed Nov 23 22:26:49 CST 2022, time=2022-11-23, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1930859, encodeId=5681193085926, content=<a href='/topic/show?id=22f61180211' target=_blank style='color:#2F92EE;'>#miR#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=11802, encryptionId=22f61180211, topicName=miR)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=0a27288, createdName=smallant2002, createdTime=Thu Jun 30 06:26:49 CST 2022, time=2022-06-30, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1697000, encodeId=2e8b169e00094, content=<a href='/topic/show?id=d2e99e651e4' target=_blank style='color:#2F92EE;'>#阻塞#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=28, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=97651, encryptionId=d2e99e651e4, topicName=阻塞)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=a23129903429, createdName=tomyang96, createdTime=Mon Jan 09 19:26:49 CST 2023, time=2023-01-09, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1199860, encodeId=165c1199860b9, content=机器学习模型……以前都是互联网+,现在流行AI+, beContent=null, objectType=article, channel=null, level=null, likeNumber=63, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20210127/50a003c4097845e182c08afd64d8a509/bd766dfebe9c40e0bca2dae964601bdc.jpg, createdBy=22de5454917, createdName=carrotlyl, createdTime=Sun Mar 06 10:43:51 CST 2022, time=2022-03-06, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1583358, encodeId=6817158335822, content=<a href='/topic/show?id=f6e1611840d' target=_blank style='color:#2F92EE;'>#机器#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61184, encryptionId=f6e1611840d, topicName=机器)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=13dc16973127, createdName=ms6279672939590805, createdTime=Thu Mar 03 08:26:49 CST 2022, time=2022-03-03, status=1, ipAttribution=)]
    2023-01-09 tomyang96
  4. [GetPortalCommentsPageByObjectIdResponse(id=1688363, encodeId=e291168836367, content=<a href='/topic/show?id=fab46121efa' target=_blank style='color:#2F92EE;'>#机器学习模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=34, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61217, encryptionId=fab46121efa, topicName=机器学习模型)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=2bea28558433, createdName=ms2694732865965676, createdTime=Wed Nov 23 22:26:49 CST 2022, time=2022-11-23, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1930859, encodeId=5681193085926, content=<a href='/topic/show?id=22f61180211' target=_blank style='color:#2F92EE;'>#miR#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=11802, encryptionId=22f61180211, topicName=miR)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=0a27288, createdName=smallant2002, createdTime=Thu Jun 30 06:26:49 CST 2022, time=2022-06-30, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1697000, encodeId=2e8b169e00094, content=<a href='/topic/show?id=d2e99e651e4' target=_blank style='color:#2F92EE;'>#阻塞#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=28, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=97651, encryptionId=d2e99e651e4, topicName=阻塞)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=a23129903429, createdName=tomyang96, createdTime=Mon Jan 09 19:26:49 CST 2023, time=2023-01-09, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1199860, encodeId=165c1199860b9, content=机器学习模型……以前都是互联网+,现在流行AI+, beContent=null, objectType=article, channel=null, level=null, likeNumber=63, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20210127/50a003c4097845e182c08afd64d8a509/bd766dfebe9c40e0bca2dae964601bdc.jpg, createdBy=22de5454917, createdName=carrotlyl, createdTime=Sun Mar 06 10:43:51 CST 2022, time=2022-03-06, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1583358, encodeId=6817158335822, content=<a href='/topic/show?id=f6e1611840d' target=_blank style='color:#2F92EE;'>#机器#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61184, encryptionId=f6e1611840d, topicName=机器)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=13dc16973127, createdName=ms6279672939590805, createdTime=Thu Mar 03 08:26:49 CST 2022, time=2022-03-03, status=1, ipAttribution=)]
    2022-03-06 carrotlyl

    机器学习模型……以前都是互联网+,现在流行AI+

    0

  5. [GetPortalCommentsPageByObjectIdResponse(id=1688363, encodeId=e291168836367, content=<a href='/topic/show?id=fab46121efa' target=_blank style='color:#2F92EE;'>#机器学习模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=34, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61217, encryptionId=fab46121efa, topicName=机器学习模型)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=2bea28558433, createdName=ms2694732865965676, createdTime=Wed Nov 23 22:26:49 CST 2022, time=2022-11-23, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1930859, encodeId=5681193085926, content=<a href='/topic/show?id=22f61180211' target=_blank style='color:#2F92EE;'>#miR#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=11802, encryptionId=22f61180211, topicName=miR)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=0a27288, createdName=smallant2002, createdTime=Thu Jun 30 06:26:49 CST 2022, time=2022-06-30, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1697000, encodeId=2e8b169e00094, content=<a href='/topic/show?id=d2e99e651e4' target=_blank style='color:#2F92EE;'>#阻塞#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=28, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=97651, encryptionId=d2e99e651e4, topicName=阻塞)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=a23129903429, createdName=tomyang96, createdTime=Mon Jan 09 19:26:49 CST 2023, time=2023-01-09, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1199860, encodeId=165c1199860b9, content=机器学习模型……以前都是互联网+,现在流行AI+, beContent=null, objectType=article, channel=null, level=null, likeNumber=63, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20210127/50a003c4097845e182c08afd64d8a509/bd766dfebe9c40e0bca2dae964601bdc.jpg, createdBy=22de5454917, createdName=carrotlyl, createdTime=Sun Mar 06 10:43:51 CST 2022, time=2022-03-06, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1583358, encodeId=6817158335822, content=<a href='/topic/show?id=f6e1611840d' target=_blank style='color:#2F92EE;'>#机器#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=61184, encryptionId=f6e1611840d, topicName=机器)], attachment=null, authenticateStatus=null, createdAvatar=null, createdBy=13dc16973127, createdName=ms6279672939590805, createdTime=Thu Mar 03 08:26:49 CST 2022, time=2022-03-03, status=1, ipAttribution=)]

相关资讯

Eur Respir J:呼吸道疾病人群的血嗜酸性粒细胞计数高于一般人群

嗜酸性粒细胞计数的分布和范围因研究人群而异,且受年龄、吸烟史、合并症等临床因素的影响,无论其严重程度如何,都应在治疗决策中加以考虑。

Oxid Med Cell Longev:1-磷酸鞘氨醇与S100A4蛋白可能参与了COPD的发病和进展

COPD患者肺和血清中S100A4蛋白表达水平增高,血清中S1P蛋白表达增加。

香港研究:只需低剂量服用,即可降低COPD患者肺癌1/4发生与死亡风险!这次是哪个“神药”呢?

PLoS Med.:低剂量阿司匹林与香港慢性阻塞性肺病患者肺癌的发生率:一项队列研究

JAMA子刊:护士长期暴露于消毒剂和手术烟雾增加慢阻肺发生风险

15年以上手术室工作年限的女性护士患COPD的风险增加。

Eur J Epidemiol:体育锻炼能够减弱吸烟导致的COPD发生风险吗?

吸烟状况和CRF均与芬兰中老年男性COPD发生风险独立相关。

Respirology:胸部CT检测能够提高COPD合并症的诊断率

使用系统方法获得胸部CT确定10种合并症的患病率高于临床诊断的患病率。