Department of Psychiatry, Faculty of Medicine, McGill University, Montreal, Quebec, Canada.
Chronic Viral Illness Service, McGill University Health Center, Montreal, Quebec, Canada.
PLoS One. 2019 Mar 25;14(3):e0213908. doi: 10.1371/journal.pone.0213908. eCollection 2019.
Mild cognitive impairment is common in chronic HIV infection and there is concern that it may worsen with age. Distinguishing static impairment from on-going decline is clinically important, but the field lacks well-validated cognitive measures sensitive to decline and feasible for routine clinical use. Measures capable of detecting improvement are also needed to assess interventions. The objective of this study is to estimate the extent of change on repeat administration of three different forms of a brief computerized cognitive assessment battery (B-CAM) developed for assessing cognitive ability in the mildly-impaired to normal range in people living with HIV. We hypothesized no change over a six-month period in people on effective antiretroviral therapy.
102 HIV+ individuals completed a set of computerized cognitive tasks on three occasions over a six-month period. Rasch analysis was used to determine if change over time (i.e. improvement due to practice) was uniform across tasks and to refine scoring in order to produce three forms of the B-CAM of equivalent level of difficulty. Group-based trajectory analysis (GBTA) was then applied to determine if performance at baseline influenced the magnitude of practice-related improvement on the battery as a whole over the course of follow-up.
Two cognitive tasks (fluency and word recall) had different levels of difficulty across test sessions, related to the different forms of the tasks. These two items were split by testing session. For all other items, the level of difficulty remained constant across all three time points. GBTA showed that the sample was composed of three distinct groups of people with unique trajectories, defined mainly by level of cognitive ability at baseline. Only the highest group showed an apparent improvement over time, but this change fell within measurement error.
Rasch analysis provides mathematical confirmation that these three forms of the B-CAM are of equivalent difficulty. GBTA demonstrates that no adjustment of the total score is required to correct for practice effects. Application of these modern statistical methods paves the way towards rapid and robust quantification of change in cognition.
轻度认知障碍在慢性 HIV 感染中很常见，人们担心它会随着年龄的增长而恶化。区分静态损伤和持续下降在临床上很重要，但该领域缺乏针对下降且适合常规临床使用的经过良好验证的认知测量方法。还需要能够检测到改善的测量方法来评估干预措施。本研究的目的是估计在重复使用三种不同形式的用于评估 HIV 感染者轻度至正常认知能力的简短计算机认知评估电池 (B-CAM) 时，变化的程度。我们假设在接受有效抗逆转录病毒治疗的人群中，在六个月内没有变化。
102 名 HIV+个体在六个月的时间内分三次完成了一组计算机认知任务。Rasch 分析用于确定随着时间的推移（即由于练习而导致的改善）是否在任务之间均匀，以及改进评分以产生三种难度水平相当的 B-CAM 形式。然后应用基于群组的轨迹分析 (GBTA) 来确定基线时的表现是否会影响整个电池在随访过程中与练习相关的改善幅度。
两个认知任务（流畅性和单词回忆）在测试会话之间具有不同的难度级别，这与任务的不同形式有关。这两个项目按测试会话分开。对于所有其他项目，难度级别在所有三个时间点都保持不变。GBTA 显示，该样本由具有独特轨迹的三个不同人群组成，主要由基线时的认知能力水平定义。只有最高组显示出随着时间的推移有明显的改善，但这种变化在测量误差范围内。
Rasch 分析提供了数学上的确认，这三种形式的 B-CAM 具有相同的难度。GBTA 表明，无需调整总分即可纠正练习效应。这些现代统计方法的应用为快速和稳健地量化认知变化铺平了道路。
