key: cord-0891672-1wvtynhz
authors: Soares, Carolina; Carneiro, Diogo Reis; Dias, Rafael; Ferreira, Daniel; Sousa, Mário; Oliveira, Ana; Massano, João; Morgadinho, Ana; Rosas, Maria José; Araújo, Rui
title: Behind the Mask: Recognizing Facial Features of Parkinson's Disease During the COVID‐19 Pandemic
date: 2021-04-28
journal: Mov Disord
DOI: 10.1002/mds.28619
sha: 2c2375c34caf2f1fcf4e0b98829beff0a9323f67
doc_id: 891672
cord_uid: 1wvtynhz

nan

Parkinson's disease (PD) remains a clinical diagnosis. 1 Charcot stated that these patients could be "diagnosed from afar" and described the "masked facies," namely, "the immobility of […] facial features." 2, 3 The MDS-UPDRS part III scores facial expression from 0 to 4, according to several features in silence and while talking. 4 Face masks became ubiquitous because of the COVID-19 outbreak, covering the nose and the mouth and thus concealing the lower half of the face. We investigated the influence of surgical masks on the recognition of facial features suggestive of PD (FFPD).

The faces of PD patients and healthy controls were videorecorded in 4 scenarios for a period of 20 seconds each: (1) silent with mask, (2) talking with mask, (3) silent without mask, and (4) talking without mask (details in supplementary file 1). Only PD patients with the "facial expression" UPDRS item ≥ 1 were included. Exclusion criteria were facial palsy, facial dyskinesia/dystonia, visible tremor, and atypical parkinsonism. Controls were assessed to exclude parkinsonian features; those with history of depression or antipsychotic therapy were excluded. The videos were randomly assigned to 6 blinded expert movement disorder neurologists from 3 centers, who classified FFPD for each subject with and without mask and also their level of assessment confidence (from 1 to 10).

We consecutively included 45 PD patients and 32 controls. A total of 450 assessments were performed: 2 evaluations per subject (with/without mask) times the number of raters ( Table 1) . The assessment confidence level increased significantly both in PD and controls after subjects took the masks off (P < 0.001). The eyeblinking rate was lower in PD patients compared with controls (with mask: 0.62 vs 1.09, P = 0.01; without mask: 0.79 vs 1.18, P = 0.029). PD patients had a significantly lower eyeblinking frequency with masks (0.62 vs 0.79, P = 0.03). Neurologists changed their impression 28.1% of the time after masks were removed, more frequently of controls than of PD patients (18.9% vs 9.2%, P < 0.03). A significant number of masked controls correctly identified as not having FFPD were reclassified after face masks was removed (4.6% vs 15.8%, P < 0.01).

Our study provides insights into the effects of face masks on the recognition of FFPD. First, the ability of neurologists to identify or exclude FFPD is largely unaffected by masks, but their use decreases diagnostic confidence (supplementary video, segment 1). Second, we found that PD patients showed reduced eyeblinking frequency with masks on (segment 2). The reason for this is unclear but may have implications in clinical practice because patients may appear more bradykinetic. Third, neurologists overidentified features of PD in controls, particularly after removing the mask (segment 3). The reason for this is also unclear but may refer to the effects of priming and cognitive biases when assessing visual clues. 5,6 Limitations of our study include low to moderate interrater agreement (supplementary file 2), in line with previous studies. 7 Face masks are likely to remain an integral part of daily life for a long time. This study suggests that the influence of masks in clinical practice, especially regarding the recognition of FFPD, should be taken into account and deserves further research. 

Additional Supporting Information may be found in the online version of this article at the publisher's web-site. The eyeblinking rate was calculated by dividing the number of blinks over the time of recording. c When neurologists shifted their evaluation after mask was removed;. d Before and after removing the mask. e Statistically significant. LEDD, levodopa-equivalent dose; SD, standard deviation; w/wo, with and without; vs, versus.

MDS clinical diagnostic criteria for Parkinson's disease

Charcot and his drawings: images from "les lecons du mardi a la Salpetriere 1887-1888