Morrow Emily L, Spieker Andrew J, Greevy Robert A, Roddy McKenzie K, Mayberry Lindsay S
Division of General Internal Medicine & Public Health, Department of Medicine, Vanderbilt University Medical Center, Nashville, USA.
Center for Health Behavior & Health Education, Vanderbilt University Medical Center, Nashville, TN, USA.
J Gen Intern Med. 2025 May;40(6):1333-1339. doi: 10.1007/s11606-024-09101-1. Epub 2024 Oct 25.
Continuous glucose monitor (CGM) use is increasing rapidly among people with type 2 diabetes, although little is known about predictors of CGM use beyond clinical and demographic information available in electronic medical records. Behavioral and psychosocial characteristics may also predict CGM use.
We examined clinical, psychosocial, and behavioral characteristics that may predict CGM use in adults with type 2 diabetes.
This longitudinal observational study comprised a secondary analysis of data collected in a larger trial. Enrollment included HbA1c tests and surveys assessing demographic, clinical, psychosocial, and behavioral characteristics. We queried participants regarding their CGM use during the study on their final self-report surveys, 15 months post-enrollment.
Participants were 245 community-dwelling adults with type 2 diabetes recruited from primary care.
We used logistic regression to predict CGM use during the 15-month trial period from baseline characteristics.
Around one-third of participants (37.1%; 91/245) started CGM. Predictors of starting CGM in bivariate models included younger age, higher socioeconomic status, insulin use, higher HbA1c, and more diabetes distress. When including all potential predictors in a single multivariable model, only younger age (aOR = 0.95, p = 0.001), insulin use (aOR = 2.33, p = 0.006), and higher socioeconomic status (aOR = 0.44, p = 0.037) were significant predictors. Despite the association between higher HbA1c and CGM use, neither diabetes self-care behaviors nor diabetes self-efficacy significantly predicted CGM use. Of participants who tried a CGM, 14.3% (13/91) had stopped, with cost being the most-cited reason.
Even when including behavioral and psychological characteristics, younger age, using insulin, and higher socioeconomic status remain key predictors of CGM use. These findings emphasize the importance of access and affordability for people who may benefit from CGM. Providers should not bias their introduction of CGM towards those with (perceived or actual) optimal or sub-optimal self-care behaviors.
连续血糖监测（CGM）在2型糖尿病患者中的使用正在迅速增加，尽管除了电子病历中可用的临床和人口统计学信息外，对于CGM使用的预测因素知之甚少。行为和心理社会特征也可能预测CGM的使用。
我们研究了可能预测2型糖尿病成人使用CGM的临床、心理社会和行为特征。
这项纵向观察性研究包括对一项更大规模试验中收集的数据进行二次分析。入组包括糖化血红蛋白（HbA1c）检测以及评估人口统计学、临床、心理社会和行为特征的调查。在入组15个月后的最终自我报告调查中，我们询问了参与者在研究期间的CGM使用情况。
参与者是从初级保健机构招募的245名社区居住的2型糖尿病成年人。
我们使用逻辑回归从基线特征预测15个月试验期内的CGM使用情况。
约三分之一的参与者（37.1%；91/245）开始使用CGM。二元模型中开始使用CGM的预测因素包括年龄较小、社会经济地位较高、使用胰岛素、HbA1c较高以及糖尿病困扰较多。当在单一多变量模型中纳入所有潜在预测因素时，只有年龄较小（调整后比值比[aOR]=0.95，p=0.001）、使用胰岛素（aOR=2.33，p=0.006）和社会经济地位较高（aOR=0.44，p=0.037）是显著的预测因素。尽管HbA1c较高与CGM使用之间存在关联，但糖尿病自我护理行为和糖尿病自我效能均未显著预测CGM的使用。在尝试使用CGM的参与者中，14.3%（13/91）已停止使用，费用是最常提及的原因。
即使纳入行为和心理特征，年龄较小、使用胰岛素和社会经济地位较高仍然是CGM使用的关键预测因素。这些发现强调了对于可能从CGM中受益的人群而言，可及性和可承受性的重要性。医疗服务提供者在介绍CGM时不应偏向于那些（被认为或实际）自我护理行为最佳或欠佳的患者。
