Paulina Vistoso Monreal Anette, Veas Nicolas, Clark Glenn
Herman Ostrow School of Dentistry, University of Southern California, Los Angeles, CA, USA.
McCombs School of Business, The University of Texas, Austin, TX, USA.
Jpn Dent Sci Rev. 2021 Nov;57:242-249. doi: 10.1016/j.jdsr.2021.11.001. Epub 2021 Nov 20.
This review examines how a highly structured data collection system could be used to create data-driven diagnostic classification algorithms. Some preliminary data using this process is provided. The data collection system described is applicable to any clinical domain where the diagnoses being explored are based predominately on clinical history (subjective) and physical examination (objective) information. The system has been piloted and refined using patient encounters collected in a clinic specializing in Orofacial Pain treatment. In summary, whether you believe a branching hybrid check-box based data collection system with built-in algorithms is needed, depends on your individual agenda. If you have no plans for data analysis or publishing about the various phenotypes discovered and you do not need pop-up suggestions for best diagnosis and treatment options, it is easier to use a semi-structured narrative note for your patient encounters. If, however, you want data-driven diagnostic and disease risk algorithms and pop-up best-treatment options, then you need a highly structured data collection system that is compatible with machine learning analysis. Automating the journey from data collection to diagnoses has the potential to improve standards of care by providing faster and reliable predictions.
本综述探讨了如何利用高度结构化的数据收集系统来创建数据驱动的诊断分类算法。文中提供了一些使用该流程的初步数据。所描述的数据收集系统适用于任何临床领域，在这些领域中，所探索的诊断主要基于临床病史（主观）和体格检查（客观）信息。该系统已在一家专门治疗口面部疼痛的诊所收集的患者病例中进行了试点和完善。总之，是否需要一个基于分支混合复选框且带有内置算法的数据收集系统，取决于个人的计划。如果您没有数据分析或发表关于所发现的各种表型的计划，并且不需要关于最佳诊断和治疗方案的弹出式建议，那么在记录患者病例时使用半结构化的叙述性记录会更简便。然而，如果您想要数据驱动的诊断和疾病风险算法以及弹出式最佳治疗方案，那么您就需要一个与机器学习分析兼容的高度结构化的数据收集系统。将从数据收集到诊断的过程自动化，有可能通过提供更快且可靠的预测来提高医疗护理标准。
