Université de Paris, CRESS, INSERM, INRA, Paris, France; Clinical Epidemiology Unit, Hôtel-Dieu Hospital, Assistance Publique-Hôpitaux de Paris, Paris, France.
Université de Paris, CRESS, INSERM, INRA, Paris, France; Clinical Epidemiology Unit, Hôtel-Dieu Hospital, Assistance Publique-Hôpitaux de Paris, Paris, France; Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, NY.
Mayo Clin Proc. 2021 May;96(5):1236-1247. doi: 10.1016/j.mayocp.2020.07.040. Epub 2021 Jan 21.
To assess the relationship between remote digital monitoring (RDM) modalities for diabetes and intrusiveness in patients' lives.
Online vignette-based survey (February 1 through July 1, 2019). Adults with diabetes (type 1, 2, or subtypes such as latent autoimmune diabetes of adulthood) assessed three randomly selected vignettes among 36 that combined different modalities for monitoring tools (three options: glucose- and physical activity [PA]-monitoring only, or glucose- and PA-monitoring with occasional or regular food monitoring), duration/feedback loops (six options: monitoring for a week before all vs before specific consultations with feedback given in consultation, vs monitoring permanently, with real-time feedback by one's physician vs by anoter caregiver, vs monitoring permanently, with real-time, artificial intelligence-generated treatment feedback vs treatment and lifestyle feedback), and data handling (two options: by the public vs private sector). We compared intrusiveness (assessed on a 5-point scale) across vignettes and used linear mixed models to identify intrusiveness determinants. We collected qualitative data to identify aspects that drove participants' perception of intrusiveness.
Overall, 1010 participants from 30 countries provided 2860 vignette-assessments (52% were type 1 diabetes). The monitoring modalities associated with increased intrusiveness were food monitoring compared with glucose- and PA-monitoring alone (β=0.34; 95% CI, 0.26 to 0.42; P<.001) and permanent monitoring with real-time physician-generated feedback compared with monitoring for a week with feedback in consultation (β=0.25; 95% CI, 0.16 to 0.34, P<.001). Public-sector data handling was associated with decreased intrusiveness as compared with private-sector (β=-0.15; 95% CI, -0.22 to -0.09; P<.001). Four drivers of intrusiveness emerged from the qualitative analysis: practical/psychosocial burden (eg, RDM attracting attention in public), control, data safety/misuse, and dehumanization of care.
RDM is intrusive when it includes food monitoring, real-time human feedback, and private-sector data handling.
评估糖尿病远程数字监测（RDM）模式与患者生活侵入性之间的关系。
在线情景调查（2019 年 2 月 1 日至 7 月 1 日）。患有糖尿病（1 型、2 型或成人隐匿性自身免疫性糖尿病等亚型）的成年人评估了 36 个情景中的 3 个随机选择的情景，这些情景结合了不同的监测工具模式（三种选择：仅监测血糖和身体活动（PA）、偶尔或定期监测食物的血糖和 PA 监测、持续/反馈循环（六种选择：监测一周，然后在所有情况下与特定咨询前进行反馈，或在咨询时给予反馈，或持续监测，由医生或另一名护理人员提供实时反馈，或持续监测，使用人工智能生成的治疗反馈或治疗和生活方式反馈的实时反馈）和数据处理（两种选择：公共部门或私营部门）。我们比较了不同情景下的侵入性（在 5 分制上评估），并使用线性混合模型确定了侵入性的决定因素。我们收集了定性数据，以确定参与者感知侵入性的驱动因素。
来自 30 个国家的 1010 名参与者提供了 2860 个情景评估（52%为 1 型糖尿病）。与单独监测血糖和 PA 相比，与食物监测相关的监测模式侵入性增加（β=0.34；95%置信区间，0.26 至 0.42；P<.001），与在咨询时给予反馈的一周监测相比，与实时医生生成反馈的持续监测侵入性增加（β=0.25；95%置信区间，0.16 至 0.34，P<.001）。与私营部门相比，公共部门的数据处理侵入性降低（β=-0.15；95%置信区间，-0.22 至-0.09；P<.001）。定性分析中出现了四个侵入性的驱动因素：实际/心理社会负担（例如，RDM 在公共场合引起注意）、控制、数据安全/滥用以及护理去人性化。
当 RDM 包括食物监测、实时人工反馈和私营部门数据处理时，它具有侵入性。
