Influence of the Intensive Care Unit Environment on the Reliability of the Montreal Cognitive Assessment

Aim: Cognitive assessment is increasingly recommended in patients surviving aneurysmal subarachnoid hemorrhage (aSAH) and is gradually becoming a part of routine clinical practice. Screening of neuropsychological functions to estimate the need for inpatient rehabilitation is often already performed during the surveillance period on the intensive care unit (ICU). For this, the Montreal Cognitive Assessment (MoCA) has been identified as first choice among the short but comprehensive instruments. It remains unknown, however, whether the distraction in the ICU environment might bias the results. This study evaluated the reliability of the MoCA in the ICU environment.

Methods: Within the framework of a nationwide prospective observational study to determine the impact of delayed cerebral ischemia on the neuropsychological outcome after aSAH, reliability measures of the MoCA were studied. Consecutive stable patients with acute brain injury were evaluated twice within 36 hours using official MoCA parallel versions (ΔMoCA). The sequence of assessment was randomized into (1) busy ICU first or (2) quiet office first with subsequent crossover. We determined sequence, period, and location effects, and the intraclass correlation coefficient (ICC) of repeated MoCA.

Results: We studied 50 patients (20 (40%) females) with a mean age of 57 years (standard deviation [SD] 17.2). The sequence of assessment (“ICU first” mean ΔMoCA –1.14 [SD 2.34] vs “Office first” –0.73 [SD 1.52]) did not affect the MoCA (p = 0.47). Patients scored 0.96 points worse on the second period (SD 2.01; p = 0.001), indicating no learning effect but possibly a difference in MoCA parallel versions. The mean ΔMoCA between locations (office minus ICU) was –0.32 (SD 2.21; p = 0.31), indicating that there was no location effect. The ICC for repeated MoCA was 0.87 (95% confidence interval 0.79–0.92).

Conclusion: The study demonstrates that there was no location effect on the MoCA results. In general, the reliability of the MoCA was found to be excellent, implying that the difference between the MoCA results of both locations is very small. Besides, the study finds that there was no learning effect, but that there may be a difference in the difficulty of the two official parallel versions of the MoCA.

Clinical Trial Registration: NCT03032471.