Limited Effect of Masking During COVID-19 Pandemic on ADOS-2 Algorithm Scores in Toddlers With and Without Autism.

Limited Effect of Masking during COVID-19 pandemic on ADOS-2 Algorithm Scores in Toddlers with and without autism - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Autism Res . Author manuscript; available in PMC: 2026 Jun 1. Published in final edited form as: Autism Res. 2025 May 12;18(6):1279–1289. doi: 10.1002/aur.70050 Limited Effect of Masking during COVID-19 pandemic on ADOS-2 Algorithm Scores in Toddlers with and without autism Vernetti Angelina 1 , Morgan Chelsea 1 , Powell Kelly 1 , Macari Suzanne 1 , Chawarska Katarzyna 1 Author information Article notes Copyright and License information 1 Child Study Center, Yale University School of Medicine, New Haven, Connecticut, USA ✉ Corresponding author : Vernetti Angelina angelina.vernetti@yale.edu Issue date 2025 Jun. PMC Copyright notice PMCID: PMC12247036  NIHMSID: NIHMS2080341  PMID: 40356197 The publisher's version of this article is available at Autism Res Abstract The Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) constitutes one of the most widely used diagnostic instruments for autism and involves a direct face-to-face interaction between clinician and child. During the COVID-19 pandemic, administration of the ADOS-2 continued in several countries, with the requirement of face mask protection. It has been hypothesized that mask wearing may have disrupted the dynamic of child-clinician interaction, and differentially affected toddlers with autism. We compared ADOS-2 algorithm scores between cohorts of toddlers evaluated before (pre-COVID-19, n=232)) and during the pandemic (COVID-19, n=116). The COVID-19 cohort included 41 toddlers with autism spectrum disorders (AUT, Mage=25.4, SD=3.8), 34 toddlers with other neurodevelopmental conditions (NDC, Mage=22.3, SD=5.0), and 41 typically developing toddlers (TD, Mage=20.4, SD=3.6) recruited between September 2020 and April 2023. The pre-COVID-19 cohort was selected from 409 assessments conducted from January 2013 to March 2020. Propensity matching was used to match the pre- and COVID-19 cohorts on sex, chronological age, and verbal and nonverbal developmental quotient (DQ) scores. Ordered logistic regression analyses were computed for Social Affect (SA) and Restricted and Repetitive Behaviors (RRB) algorithm total and item scores with cohort (pre-COVID-19/ COVID-19) as fixed factor for each diagnostic group. The analyses revealed a limited impact of cohort on the algorithm scores in all three diagnostic groups. Item-level analysis revealed significant cohort effect only on two out of twenty items: shared enjoyment and joint attention, with higher (more atypical) scores found in the COVID-19 than in the pre-COVID-19 cohorts. The resiliency of the algorithm and item-level scores to the effect of masking speaks to the strength of the diagnostic tool and its ability to capture a range of social, communication, and repetitive behaviors under both standard and non-standard conditions. Keywords: ADOS-2, Autism, Toddler, COVID-19, Mask, Pandemic, Algorithm scores, Item level Lay summary During the COVID-19 pandemic, in research and clinical settings the use the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) continued, despite the challenges posed by mandatory mask-wearing in several countries. The current observational study investigated how mask wearing affected ADOS-2 algorithm scores in toddlers with autism, other neurodevelopmental conditions, and typical development. Overall, mask wearing had a limited impact on algorithm scores. The findings speak of the resilience of the ADOS-2 Toddler module to pandemic-related changes in standard administration. Introduction The diagnosis of autism spectrum disorder (henceforth, autism), a neurodevelopmental disorder consisting of persistent difficulties with social communication and social interaction as well as restricted and repetitive patterns of behaviors, activities or interest 1 involves a careful examination of the child’s medical and developmental history as well as a direct assessment of social and cognitive development conducted by qualified clinicians 2 . In the first months of the COVID-19 pandemic 3 – 6 , many children suspected of having neurodevelopmental conditions were unable to access in-person assessments. 7 When in-person clinical evaluations slowly resumed during the pandemic, the use of personal protective equipment (PPE) including surgical face masks became mandatory in many countries, potentially altering the social environment in which evaluations took place. One of the most widely used instruments for evaluating core symptoms of autism is the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) 8 . The Toddler module is designed for children aged 12 to 30 months with language functioning ranging from preverbal/limited speech to the use of simple two-word phrases. This module consists of 10 play probes (see Table 1 ) administered by a clinician who subsequently rates the child’s behaviors on 41 items. The scores of selected items are then summed and form a diagnostic algorithm score. Higher ADOS-2 scores indicate greater severity of impairment at both item and algorithm levels. Table 1. Description of each of the algorithm Items of the Autism Diagnostic Observation Schedule (ADOS-2) Toddler Module. ADOS-2 Toddler Module Algorithm Items Category Item number Description Social Affect A-2 Frequency of Spontaneous Vocalizations Directed to Others A-7 Pointing A-8 Gestures B-1 Unusual Eye Contact B-4 Facial Expressions Directed to Others B-5 Integration Gaze and Other Behaviors B-6 Shared Enjoyment in Interaction B-7 Response to Name B-8 Ignore B-9 Requesting B-12 Showing B-13 Spontaneous Initiation of Attention B-14 Response to Joint Attention B-15 Quality of Social Overtures B-16b Amount of Social Overtures - Parent/Caregiver B-18 Overall Quality of Rapport Repetitive and Restricted Behaviors A-3 Intonation of Vocalizations or Verbalizations D-1 Sensory Interest Play Material/Person D-2 Hand and Finger Movements and Posturing D-5 Repetitive Interests and Stereotyped Behaviors Open in a new tab The validity of the administration of the ADOS-2 during the COVID-19 pandemic has been widely discussed, and though many research and clinical sites continued to employ it during the pandemic, many anticipated that the use of PPE would interfere with administration of the tool and alter its the scores in a systematic manner. 9 – 12 To remedy to this concern, other assessments tools to support autism diagnosis during the pandemic were developed to respond to the needs of clinical evaluations while social distancing. Among these tools were communication technologies that provided health care at a distance using video observations, live video evaluations, and caregiver-administered assessments 9 guided by clinicians. These procedures prove not only to be promising valid tools, but also show potential to increase access to evaluations to more families 13 . However, while these tools have contributed to help with the continuity of evaluations and in the development of a broader range of diagnostic tools, there are several administrative differences when compared to the ADOS-2. They all involve significantly different social contexts compared to the naturalistic interaction with an expert clinician. The ADOS-2 has the advantage of giving clinicians a standard face-to-face setting in which to observe social communicative skills and situate their findings within a framework of established probes. COVID-19 masking influenced various aspects of social interaction, such as altering communication norms, limiting emotional expressiveness, and potentially prompting adaptive strategies to connect, but its impact was particularly evident in hindering facial visibility and the use of oral-motor cues essential for effective communication. During the pandemic, the mandatory masks occluded the nose and the mouth areas. It has been suggested that such an occlusion could influence the ability to recognize emotions in interactive partners as well as impact speech processing. Specifically, facial occlusion may prevent holistic face processing and lead to less effective emotion recognition 14 – 16 . Indeed, several studies have shown that the identification of basic emotions in preschoolers, older children, and adults may be disturbed by face covering. 17 – 20 However, other studies have demonstrated that despite the covering of the lower face, both children and adults are still able to correctly recognize facial expressions 21 . It is important to note that in many of these studies, emotion recognition has been examined using static pictures of masked faces. However, while configuration of lower facial feature is important for emotion recognition, other cues such as eye expressions, movement of eyebrows, prosody, vocalizations, hand gestures, and other body movements also constitute essential sources of information regarding emotional states of others. 16 , 22 , 23 Moreover, even though some studies report reduced emotion recognition performances with partial face occlusion, the performance is typically still above chance level not only in older children but also in toddlers 17 , 19 . Finally, it is plausible that limiting the number of emotional cues may be in fact helpful to individuals with autism, as shown in a report suggesting that frequent exposure to masks during the pandemic improved emotion perception among individuals with autistic traits 24 . If so, with fewer facial cues present, toddlers with autism may in fact benefit from facial occlusion during the ADOS-2 administration, relying more on other facial, motor, and prosodic cues. Masks could also potentially affect speech perception as they occlude oral motor cues. Indeed, while infants look much more at the eye region of talking faces than the mouth during the first three months of life 25 , 26 , by 8 months of age and throughout toddlerhood, infants increase their gaze toward the mouth region of speaking social partners. 27 – 29 Therefore, any possible detrimental mask effect could be most pronounced in young language learners, who might find it more difficult to understand speech without seeing the lip movements corresponding to speech sounds 26 . Still, while gaze begins to shift at times towards the mouth region during infancy, the eye region 28 still remains a key focus of attention during social interactions. Furthermore, as a group, toddlers with autism already show diminished attention to the mouth region compared to developmentally delayed and typically developing toddlers (Shic et al, 2020) and infants with familial history of autism but without the diagnosis of autism do not appear to benefit from audiovisual prosodic and speech cues in the service of language acquisition despite intact attention to these cues. 30 Thus, much remains to be learned about the role of attention to and processing of multimodal audio-visual cues in the dynamics of social interaction in autism. Although most research on face occlusion and its effects on face and speech perception has been carried out in highly controlled experimental settings with unimodal and static facial stimuli, the extent to which occlusion influences social and communicative behaviors in children, both with and without neurodevelopmental conditions, during live, face-to-face behavioral assessments remains uncertain. Additionally, it is extremely important to report on the administration of the ADOS-2 under the use of PPE knowing its wide application in clinical and research settings 31 and its reduced 13 but maintained 32 use in many institutions during the COVID-19 pandemic. Despite the importance of the question, to the best of our knowledge, there have been no empirical studies as to whether mask wearing affects the results of the ADOS-2 assessment in young children suspected of having autism. In the present study we examined the effect of masking on the ADOS-2 Toddler Module total and item algorithm scores in toddlers with autism, other neurodevelopmental conditions, and typical development. Two cohorts of toddlers were compared: toddlers who were evaluated during the pandemic between September 2020 to April 2023 (COVID-19 cohort, n=116) and toddlers evaluated between January 2013 to March 2020 (pre-COVID-19 cohort, n=232) drawn from a poll of 409 of the ADOS-2 Toddler module assessments. The two groups were carefully matched on age, sex, and verbal and nonverbal DQ scores through a propensity matching procedure. We hypothesized that if facial occlusion disrupts the quality and quantity of social overtures, the COVID-19 cohort would have higher scores on the ADOS-2 Toddler module compared to the pre-COVID-19 cohort, regardless of diagnosis. However, it is also plausible that the diagnostic groups are affected by masking differentially. For instance, while the non-autistic toddlers may be able to compensate for impoverished facial cues with other sources of social-communicative cues, toddlers with autism may not, leading to significant increases in the scores due to cohort effect only in the autism group. Alternatively, in toddlers with autism, impoverished facial cues and lesser load on their social perception system may result in beneficial effects, such that the COVID-19 cohort may exhibit lower scores than the pre-COVID-19 cohort. Methods Participants Participants included toddlers referred to a university-based clinic specializing in the differential diagnosis of neurodevelopmental conditions or were recruited through community advertisements. All participants underwent a comprehensive assessment of verbal and nonverbal skills using the Mullen Scales of Early Learning (MSEL) 33 , autism severity using the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) 34 , and adaptive skills using the Vineland Adaptive Behavior Scales, Second Edition (Vineland-II) and Third Edition (Vineland-3) 35 , in conjunction with a review of medical and developmental history. To be included in the autism group, the child had to receive a clinical best estimate diagnosis of autism, including meeting DSM-5 criteria, and having a total calibrated severity score (ADOS-2 Total CSS) of 4 or higher on the ADOS-2. Toddlers in the group with other neurodevelopmental conditions (NDC) included those with global or specific (e.g., language, cognitive, social, or motor) delays or attentional or emotional/behavioral regulations challenges. No specific MSEL thresholds nor ADOS criteria were applied to the group with other NDC. Toddlers were classified as developing typically (TD) if they had MSEL verbal and nonverbal developmental quotients (VDQ and NVDQ, respectively) above 80, a total ADOS-2 Total CSS of 3 or lower, and confirmation of typical development by an expert clinician. Toddlers with history of prematurity or known genetic abnormalities were not included in the study. The COVID-19 cohort (Mean age = 22.7, SD=4.1) included 41 toddlers with autism (AUT, 29 males), 34 NDC toddlers (23 males) and 41 TD toddlers (16 males) (see Table 2 for participant characteristics). Table 2. Propensity matching results between Pre-COVID-19 and COVID-19 cohorts in toddlers with autism (AUT), other neurodevelopmental conditions (NDC), or exhibiting typical development (TD). Propensity matching results AUT NDC TD Standardized mean difference (d) 0.08>d>0.29 0.03>d>0.10 0.03>d>0.26 Variance ratios (vr) 1.26>vr>1.86 0.47>vr>1.71 0.70>vr>0.90 Open in a new tab Propensity matching. To create a comparison group for the COVID-19 cohort, we employed a propensity matching procedure using a sample of 409 ADOS-2 Toddler-Module assessments from 329 unique toddlers, conducted between 2013 and 2020 on toddlers with AUT (n=136), NDC (n=124), and TD (n=149). The matching was conducted separately for each diagnostic group using the R package Matchit 36 , using the nearest neighbor method and ratio of 2. This method facilitates selection of matched groups and has been widely implemented and recommended to be used in neurodevelopmental disorders research 37 . Matching variables comprised chronological age, sex, VDQ and NVDQ. The final pre-COVID-19 dataset included 232 ADOS-2 Toddler-Module assessments (Mean age = 22.3, SD=3.6), 82 toddlers with autism spectrum disorders (61 males), 124 toddlers with NDC (43 males), and 149 typically developing toddlers (31 males). Procedure Assessment of autism symptoms. In both cohorts, the ADOS-2 probes of the Toddler Module were administered by clinical psychologists with 10–15 years of clinical experience working with toddlers with neurodevelopmental conditions, research reliable in ADOS administration. In both cohorts, the toddlers were unmasked, but during the COVID-19 pandemic, both the clinician and the parent were masked. All the modifications to the ADOS administration during COVID-19 pandemic were consistent with standard ADOS protocol. The “snack probe” was administered using two toys from the ADOS kit that the child previously demonstrated interest in, instead of edible snacks. Two probes were adjusted within the parameters specified by the ADOS-2 manual which allows for replacement tasks to minimize the spread of germs: 1) the “response to joint attention” probe consisted of the use of the remote-controlled car (easier to clean) instead of the more commonly used remote-controlled furry bunny; 2) the “anticipation of a social routine” probe involved the use of rockets instead of latex balloons. Only one probe was skipped (the ‘response to smile’) due to masking, but this item did not contribute to the algorithm scores and thus is not considered in the present study. Dependent measures. Preliminary measures consisted of the propensity matching results based on cutoff values proposed in the literature 38 , 39 (a maximum standardized mean difference (d) of approximately .25 and variance ratios (vr) within the range of .5 to 2) as well as comparison of the cohorts on sex, age, and MSEL verbal and nonverbal developmental quotients (i.e., the matching variables). Primary measures consisted of the ADOS-2 total algorithm scores from the Social Affect (SA) and Restricted and Repetitive Behavior categories (RRB) as well as overall (SA + RRB) total scores. Additional dependent measures comprised the item algorithm scores (A2, A7, A8, B1, B4-B9, B12-B15, B16b, B18, A3, D1, D2 and D5, see Table 2 for description). Although the Toddler Module comprises two algorithm columns with different items totaled based on age and language level, for the purpose of this study, all possible twenty items of both algorithms were examined together in the analyses. To increase the range of scores and improve sensitivity in detecting differences between the cohorts, during the examination of individual items, ratings of 3 were not converted to 2, while ratings of 7, 8, and 9 were treated as missing data, accounting for 1%. Additionally, a supplemental analysis was performed using the standard ADOS-2 algorithm conversion where ratings of 3 were adjusted to 2, and ratings of 7, 8, and 9 were converted to 0. Statistical analysis. In the preliminary analysis, to examine the results of propensity matching, we conducted a series of t-tests or chi-square on sex, age, VDQ and NVDQ between the pre- and COVID-19 cohorts in each of the diagnostic groups. In the primary analyses, the algorithm and item-level scores were compared between the cohorts separately in AUT, NDC and TD groups. Ordered logistic regressions were computed for each total or item scores for each group (AUT/NDC/TD) with cohort (pre-/ COVID-19) as fixed factor and subject as random factor. When samples exhibited only one score, (e.g., TD group only exhibited ratings of 0s for item B15, COVID-19 cohort), ordered logistic regression was not conducted. False discovery rate (FDR) correction was applied and both uncorrected and corrected p-values are reported in the results. Results Preliminary Analysis Propensity matching results. The analyses revealed that the d values were close or below the maximum of .25 and vr were within acceptable ranges in all three groups ( Table 2 ). In each diagnostic group, the cohorts did not differ significantly on any of the matching variables including chronological age (all ps > 0.133), VDQ (all ps>0.185), NVDQ (all ps > 0.185) and sex (ps>0.109). Additional analysis of race and ethnicity representation revealed no significant differences between both pre- and COVID-19 cohorts in the overall sample and within each diagnostic group (all ps > 0.118), see Table 3 for sample characteristics). Table 3. Participants characteristics of the Pre-COVID-19 and COVID-19 cohorts in toddlers with autism (AUT), with other neurodevelopmental conditions (NDC) or exhibiting typical development (TD). AUT NDC TD Pre-COVID-19 (N=82) COVID-19 (N=41) Pre-COVID-19 (N=68) COVID-19 (N=34) Pre-COVID-19 (N=82) COVID-19 (N=41) Age (months) Mean (SD) 24.4 (3.10) 25.4 (3.80) 21.9 (3.81) 22.3 (4.97) 20.5 (3.79) 20.4 (3.59) Median [Min, Max] 25.0 [17.6, 29.9] 26.0 [17.9, 30.7] 23.3 [12.0, 28.9] 22.3 [12.1, 30.8] 19.6 [11.9, 30.4] 19.3 [12.5, 28.6] Sex Female 21 (25.6%) 12 (29.3%) 25 (36.8%) 11 (32.4%) 51 (62.2%) 25 (61.0%) Male 61 (74.4%) 29 (70.7%) 43 (63.2%) 23 (67.6%) 31 (37.8%) 16 (39.0%) Verbal DQ Mean (SD) 51.6 (20.8) 55.5 (23.4) 77.1 (26.8) 79.0 (18.4) 110 (16.8) 106 (15.6) Median [Min, Max] 47.0 [19.4, 106] 55.1 [18.5, 112] 70.5 [23.1, 164] 74.8 [40.1, 120] 109 [81.9, 153] 107 [80.3, 139] Nonverbal DQ Mean (SD) 76.0 (12.6) 71.1 (17.2) 91.9 (13.5) 92.2 (11.3) 108 (13.6) 106 (11.3) Median [Min, Max] 76.3 [36.1, 98.8] 71.3 [32.6, 101] 90.5 [60.2, 124] 91.4 [68.0, 116] 106 [87.2, 151] 103 [87.2, 132] ADOS-2 SA CSS Mean (SD) 7.05 (2.22) 7.76 (2.19) 3.15 (1.44) 3.24 (2.09) 1.83 (1.03) 1.98 (0.961) Median [Min, Max] 7.50 [3.00, 10.0] 8.00 [3.00, 10.0] 3.00 [1.00, 8.00] 3.00 [1.00, 8.00] 1.00 [1.00, 4.00] 2.00 [1.00, 4.00] ADOS-2 RRB CSS Mean (SD) 7.13 (1.60) 7.51 (1.87) 3.91 (2.09) 3.53 (2.35) 2.60 (2.02) 1.61 (1.50) Median [Min, Max] 7.00 [1.00, 10.0] 8.00 [1.00, 10.0] 5.00 [1.00, 7.00] 5.00 [1.00, 7.00] 1.00 [1.00, 6.00] 1,00 [1.00, 6.00] ADOS-2 RRB CSS Mean (SD) 7.13 (1.60) 7.51 (1.87) 3.91 (2.09) 3.53 (2.35) 2.60 (2.02) 1.61 (1.50) Median [Min, Max] 7.00 [1.00, 10.0] 8.00 [1.00, 10.0] 5.00 [1.00, 7.00] 5.00 [1.00, 7.00] 1.00 [1.00, 6.00] 1,00 [1.00, 6.00] ADOS-2 Total CSS Mean (SD) 7.29 (2.00) 8.02 (1.93) 2.87 (1.23) 3.03 (1.90) 1.55 (0.669) 1.66 (0.656) Median [Min, Max] 7.00 [4.00, 10.0] 9.00 [4.00, 10.0] 3.00 [1.00, 8.00] 2.00 [1.00, 7.00] 1.00 [1.00, 3.00] 2.00 [1.00, 3.00] Race Asian 11 (13.4%) 4 (9.8%) 3 (4.4%) 4 (11.8%) 7 (8.5%) 5 (12.2%) Black or African American 6 (7.3%) 2 (4.9%) 8 (11.8%) 2 (5.9%) 6 (7.3%) 0 (0%) More Than One Race 11 (13.4%) 4 (9.8%) 9 (13.2%) 6 (17.6%) 8 (9.8%) 6 (14.6%) Unknown or Not Reported 1 (1.2%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) White 47 (57.3%) 27 (65.9%) 48 (70.6%) 20 (58.8%) 60 (73.2%) 28 (68.3%) Caucasian 0 (0%) 1 (2.4%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) Unknown/Not Reported 0 (0%) 1 (2.4%) 0 (0%) 1 (2.9%) 0 (0%) 1 (2.4%) Missing 6 (7.3%) 2 (4.9%) 0 (0%) 1 (2.9%) 1 (1.2%) 1 (2.4%) Ethnicity Hispanic 16 (19.5%) 6 (14.6%) 21 (30.9%) 6 (17.6%) 10 (12.2%) 4 (9.8%) Nonhispanic 53 (64.6%) 31 (75.6%) 45 (66.2%) 26 (76.5%) 71 (86.6%) 34 (82.9%) Not Specified 5 (6.1%) 1 (2.4%) 2 (2.9%) 1 (2.9%) 0 (0%) 2 (4.9%) Missing 8 (9.8%) 3 (7.3%) 0 (0%) 1 (2.9%) 1 (1.2%) 1 (2.4%) Open in a new tab ADOS-2: Autism Diagnostic Observation Schedule-Second Ediion; CSS: Callorated Severity Score; DQ: Developmental Quotient Primary Analysis ADOS-2 algorithm scores. The cohort effect on the Total score was not statistically significant in all three groups: AUT [β=0.44, p=0.061/FDR-p=0.200], NDC [β=−0.11, p=0.704/FDR-p=0.751], and TD [β=0.20, p=0.399/FDR-p=0.612] (see Table 4 ). There were also no significant cohort effects on the total SA scores in the three groups [AUT [β=0.45, p=0.060/FDR-p=0.200], NDC: β=−0.08, p=0.751/FDR-p=0.751, TD: β=0.38, p=0.106/FDR-p=0.271]]. For the total RRB scores, neither AUT nor NDC groups exhibited a cohort effect: AUT: β=0.26, p=0.272/FDR-p=0.417, NDC: β=−0.25, p=0.374/FDR-p=0.577]. The TD group exhibited significantly higher RRB scores during the pre-COVID-19 period compared to the COVID-19 pandemic period, but this effect was no longer significant after false discovery rate correction [TD: β=−0.89, p=0.006/FDR-p=0.069] ( Table 4 ). Table 4. Mean (SD, Range) of the ADOS-2 Total, Social Affect (SA), and Restrictive and Repetitive Behaviors (RRB) Algorithm Scores in Pre-COVID-19 and COVID-19 cohorts as well as uncorrected and corrected (FDR, false discovery rate correction) p-values, log Odd Ratios, Odd Ratios, and confidence intervals from ordinal regression analyses computed separately for toddlers diagnosed with autism (AUT), with other neurodevelopmental conditions (NDC) or exhibiting typical development (TD). Open in a new tab In a secondary analysis, we compared the number of toddlers with autism falling into the moderate versus severe range of concerns based on the ADOS-2 total algorithm scores. Before the pandemic, 73% of toddlers with autism fell into the severe range of concern compared to 85% during the pandemic but these proportions were not significantly different (Χ 2 (1, 123) = 1.67, p=.196). ADOS-2 Social Affect item scores. As can be seen in Table 5 , i n the AUT group, there was no cohort effect on 13 out of the 16 SA items (ps > 0.094, FDR-ps > 0.240). There was, however, a significant increase in scores on the shared enjoyment in interaction [item B6, β=0.93, p<0.001/FDR-p<0.001], requesting [item B9, β=0.66, p=0.008/FDR-p=0.046], and spontaneous initiation of joint attention [item B13, β=1.17, p<0.001/FDR-p<0.001] items. Table 5. Mean (SD, Range) of the ADOS-2 Social Affect Algorithm Item Scores in the Pre-COVID-19 and COVID-19 cohorts as well as uncorrect and corrected (FDR, false discovery rate correction) p-values, log Odd Ratio, Odd Ratio and confidence intervals from ordinal regression analyses computed separately for all toddlers diagnosed with autism (AUT), with other neurodevelopmental conditions (NDC) or exhibiting typical development (TD). Open in a new tab In the NDC group , there was no significant cohort effect for 13 out of 16 items (ps>0.086, FDR-ps>0.396). There was an increase in the shared enjoyment in interaction item [B6, β=1.12, p=0.003/FDR-p=0.034] and spontaneous initiation of joint attention [item B13, β=1.09, p<0.001/FDR-p=0.023] ( Table 5 ). Although an increase in scores was also seen in the frequency of spontaneous vocalization directed to others item, this effect did not survive the FDR correction [item A2, β=−0.70, p=0.019/FDR-p=0.109]. In the TD group , there was no significant cohort effect on 14 out of 16 SA items (ps>0.066, FDR-ps>0.237) ( Table 5 ). There was a significant increase in the spontaneous initiation of joint attention score [item B13, β=2.22, p<0.001/FDR-p<0.001]. There was an increase in scores for the pointing item, however this effect was no longer significant after correction [item A7, β=0.66, p=0.014/FDR-p=0.086]. The analysis was not computed for the item Quality of social overtures as the model was unstable due to the TD group ratings consisting of nearly all 0s in both cohorts. ADOS-2 Restricted and Repetitive Behaviors item scores In the AUT group , the cohort effect was not significant for 2 out of 4 Restricted and Repetitive Behaviors items (ps>0.071, FDR-p>0.204) ( Table 6 ). There was a decrease in scores for the item intonation of vocalizations or verbalizations, [item A3, β=−1.31, p<0.001/FDR-p<0.001]. While scores for the item repetitive interests and stereotyped behaviors appeared to increase, the effect was no longer significant after the FDR correction [item D5, pre-COVID-19: M=1.29, SD=0.78 / COVID-19: M=1.56, SD=0.78; β=0.55, p=0.039/FDR-p=0.179]. Table 6. Mean (SD, Range) of the ADOS-2 Restricted and Repetitive Behaviors Algorithm Item Scores in the Pre-COVID-19 and COVID-19 cohorts as well as uncorrected and corrected (FDR, false discovery rate correction) p-values, log Odd Ratio, Odd Ratio and confidence intervals from ordinal regression analyses computed separately for all toddlers diagnosed with autism (AUT), with other neurodevelopmental conditions (NDC) or exhibiting typical development (TD). Open in a new tab In the NDC group , the cohort effect was not significant for 3 out of 4 Restricted and Repetitive Behaviors items ( Table 6 ). Although there was a decrease in the intonation of vocalizations or verbalizations item scores, the significance dropped after the FDR correction [item A3, β=−1.02, p=0.017/FDR-p=0.109] ( Table 6 ). Finally, in the TD group , the COVID-19 cohort effect was not significant for any of the 4 Restricted and Repetitive Behaviors items (ps>0.085, FDR-ps>0.244). The analysis was not computed for the item intonation vocalization or verbalizations as the model was unstable due to the TD group ratings consisting of nearly all 0s during the pre- and COVID-19 pandemic ( Table 6 ). The results remain consistent when evaluating individual scores between cohorts within diagnostic groups, where ratings of 3 were converted to 2, and ratings of 7, 8, and 9 were converted to 0 following the standard ADOS-2 algorithm conversion (refer to Supplementary Tables S1 and S2 ). Discussion The COVID-19 pandemic created a unique natural experimental opportunity to evaluate the robustness of the behavioral measures for autism symptoms in young children. Capitalizing on this opportunity, the current study examined the impact of face masking on performance of toddlers with autism, other developmental conditions, and typical development, during an in-person assessment for autism symptoms. Their performance was contrasted with that of toddlers matched on sex, age, and verbal and nonverbal skills who were evaluated before the pandemic, i.e., in the absence of the masking requirement. Surprisingly, the results indicate limited impact of masking on performance in all three diagnostic groups when the ADOS-2 total algorithm scores as well as the SA and RRB domain scores were considered. When we examined cohort differences in individual item scores, only spontaneous initiation of joint attention evidenced a cohort effect across all groups, with pandemic scores being more atypical than pre-pandemic scores. In addition, shared enjoyment in interaction and requesting scores were also more atypical but only in the autism and NDC groups. Lastly, fewer atypical intonation, vocalization, or verbalization behaviors were noted in the autism COVID-19 cohort. These differences, however, did not affect the overall patterns of the algorithm scores. We discuss the results and their potential significance in turn below. The ADOS-2 algorithm scores are used, in conjunction with information from other sources (e.g., verbal and nonverbal skills, adaptive functioning, as well as medical and developmental history) to aid in differential diagnosis of autism. 10 Following the implementation of pandemic safety measures, concerns were raised about a potential systematic bias caused by the masking requirement. 10 , 40 The concerns related to the interpretability of the examiner’s facial expressions during the ADOS-2, on which several probes are based, which many argued could not be realized due to the mandatory use of masks. And yet, many programs continued in-person diagnostic assessments during the pandemic due to the ongoing clinical or research needs. In-person assessment is particularly useful in young children where broader developmental and contextual factors matter greatly for interpretation of variability in social affective deficits and presence of repetitive behaviors. 2 The overall results show a negligible impact of masking on the total algorithm scores as well as Social Affect and Restricted and Repetitive Behaviors domain scores and this effect was consistent in toddlers with and without autism. There were also no significant differences in the proportion of toddlers with autism falling into the severe concern range based on the total algorithm scores. These findings suggest that masking had no major effect on performance on children with and without neurodevelopmental conditions and the ADOS-2 scores obtained during pandemic under otherwise standard administration conditions, are likely to reflect the child’s true range of social communicative and repetitive behaviors. A deeper investigation into the individual item performance revealed that a vast majority of item scores were not affected by the masking requirement. Specifically, in the autism group, scores were comparable in the core social nonverbal communication items including the use of eye contact, facial expressions, gestures, pointing, and showing. There were also no significant differences in responsivity to bids for joint attention or response to name. The summary scores capturing frequency and quality of social overtures and engagement were also comparable before and during pandemic. Items capturing sensory interests, motor mannerisms, and stereotyped interests and behaviors also did not register significant cohort effects. However, during the pandemic, toddlers with autism displayed a lower frequency and quality of spontaneous initiation of joint attention bids, sharing enjoyment, and frequency or quality of requesting behavior. Interestingly, most of these differences were not autism-specific: both NDC and typically developing groups showed more atypical scores on the item capturing initiation of joint attention, and toddlers with NDC were less likely to share enjoyment during the ADOS-2. Interestingly, toddlers with autism and other neurodevelopmental conditions displayed vocalizations with fewer atypical intonation markers. Taken together, the study suggests that the ADOS-2 Toddler Module play probes designed to capture behaviors related to autism are very robust to alterations in the administration procedure due to masking. With few exceptions, behavioral profiles relating to core autism symptoms remained unchanged during the pandemic across all the groups. This is likely due to the fact that the social engagement cues at the examiner’s disposal extend far beyond the oral-motor movements and include language and prosodic cues, global facial and gaze gestures, as well as head and body movements, all of which can scaffold and support emotion recognition and communicative bids in toddlers with and without neurodevelopmental conditions. Thus, it is even more interesting that selected few items appeared differentially affected – both in a positive and negative manner – and these effects remained significant after the correction of multiple comparison. All three behaviors involve directing social overtures to others: sharing enjoyment and interests and requesting. It is plausible that masking has dampened a drive to initiate communicative bids. Other factors, however, need to be considered as well. For instance, the toddlers born either right before or during the pandemic were raised under lock-down or highly restricted contact conditions and thus, may have had fewer opportunities to hone some of their social interaction skills with unfamiliar adults, thereby lowering frequency of these spontaneous acts. Examining further the impact of masking on, for instance, joint attention would be of great importance given the pivotal role this skill plays during social interactions and its targeted remediation during interventions. Significance. The current results are promising on many levels. First, it is very encouraging that toddlers have shown a great degree of resilience in the face of pandemic social restrictions, which was apparent in toddlers with and without autism. Second, the results suggest that the ADOS-2 constitutes a valid measure of autism symptoms, even if some of the aspects of the protocol need to be altered due to safety concerns. The latter is important for clinical reasons, but also research reasons, especially involving a longitudinal design and assessment points including the pandemic. It also suggests that data collected during the pandemic under strict research conditions could be pooled with those collected pre- or post-pandemic. Given the widespread use of the ADOS-2 tool in research and clinical settings, a better understanding of the fluctuation (or lack thereof) of the scoring between pre- and COVID-19 pandemic is extremely valuable. Limitations. Amongst the study’s many strengths are careful cohort matching on sex, age, and cognitive abilities, as well as inclusion of children with and without autism. The ADOS-2 protocol was administered by clinic- and research-reliable experts in clinical psychology. It is not clear if a similar pattern of results would be obtained in community-based settings or with less expert administators 10 . Investigation of this question would represent an important follow-up direction. Furthermore, the study controlled many aspects of the administration to ensure adherence to the manualized protocol. It is plausible however, that the examiners implicitly employed additional strategies (e.g., increased the volume of their voice, accentuated their expressions using their eyebrows) to compensate for the non-standard administration procedure with the masking. Lastly, we present the findings from a ‘natural experiment’. Although we have made a significant effort to match the two cohorts (i.e., pre-COVID-19 and COVID-19) on multiple aspects, having a randomized assignment to the assessment conditions would have been more effective. However, this approach would raise ethical concerns. Conclusions. Masking had no major effect on direct assessment results typically used to aid in differential diagnosis of autism in toddlers. It is likely that despite limited access to the oral-motor cues, the children benefited from and relied on the presence of other social communicative cues such as eye contact and gaze direction cues, eyebrow movements, language, prosody, laughter, hand gestures, and other body movements. The study supports the use of the ADOS-2 as a valid component of the autism diagnostic process even when clinician masking is required. Supplementary Material Supinfo NIHMS2080341-supplement-Supinfo.docx (507KB, docx) Key points and relevance. The Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) constitutes one of the most widely used diagnostic instruments for autism. During the COVID-19 pandemic, administration of the ADOS-2 continued in several institutions, however face mask protection worn by the examiner may have potentially disrupted the dynamic of child-clinician interaction. By comparing the ADOS-2 algorithm scores between cohorts of toddlers with and without autism matched on chronological age, sex, and cognitive score, and evaluated before and during the pandemic, we show that the majority of the ADOS-2 algorithm scores were unaffected in all three diagnostic groups. The study provides support for the overall validity of the ADOS-2 scores in toddlers administered live with a masked, research-reliable, trained clinician (and unmasked child) during the pandemic. Acknowledgments This research was supported by NIMH MH124892, NIH P50 HD093078R01, NIMH R01 MH100182, DOD W81XWH-12-ARP-IDA, NIMH R01 MH087554, NIMH P50 MH115716, The Associates of the Child Study Center awarded to Dr. Chawarska. We thank our clinicians Dr. Mariana Torres-Viso, Megan Lyons, Amy Giguere Carney, and Karyn Bailey for their assistance in sample characterization. Additional thanks to Evelyn Pomichter, Amy Margolis, Karen Franchi, and Gitta Selva and dedicated research fellows for their contribution to sample recruitment, and data collection as well as anonymous reviewers for their insightful comments and suggestions. Our deepest gratitude to all the families who participated in our studies. Funding/Support: All phases of this study were supported by the National Institutes of Mental Health R01 MH100182, R01 MH124892 R01 MH087554, R01 MH111652, and P50 MH115716 grants awarded to Katarzyna Chawarska. Footnotes Conflict of Interest Disclosures (includes financial disclosures) The authors have no conflicts of interest to disclose. Ethical considerations. All research activities described in this study adhere to rigorous ethical guidelines. Informed consent was obtained from all participants, and the research was conducted with explicit permission from Yale University’s Institutional Review Board. References 1. American Psychiatric Association. Diagnostic and statistical manual of mental disorders: DSM-5. American Psychiatric Pub; 2013. [ Google Scholar ] 2. Steiner AM, Goldsmith TR, Snow AV, Chawarska K. Practitioner’s guide to assessment of autism spectrum disorders in infants and toddlers. Journal of autism and developmental disorders. 2012;42:1183–1196. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Wu Y-C, Chen C-S, Chan Y-J. The outbreak of COVID-19: An overview. Journal of the Chinese medical association. 2020;83(3):217. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Ciotti M, Angeletti S, Minieri M, et al. COVID-19 outbreak: an overview. Chemotherapy. 2020;64(5–6):215–223. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 5. Chakraborty I, Maity P. COVID-19 outbreak: Migration, effects on society, global environment and prevention. Science of the total environment. 2020;728:138882. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 6. Carranza Carnicero JA, Pérez-López J, Del Carmen González Salinas M, Martínez-Fuentes MT. A longitudinal study of temperament in infancy: Stability and convergence of measures. European Journal of Personality. 2000;14(1):21–37. [ Google Scholar ] 7. Jang J, White SP, Esler AN, et al. Diagnostic evaluations of autism spectrum disorder during the COVID-19 pandemic. Journal of Autism and Developmental Disorders. 2021:1–12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Lord C, Rutter M, DiLavore P, Risi S, Gotham K, Bishop S. Autism diagnostic observation schedule–2nd edition (ADOS-2). Los Angeles, CA: Western Psychological Corporation. 2012; [ Google Scholar ] 9. Dow D, Holbrook A, Toolan C, et al. The brief observation of symptoms of autism (BOSA): development of a new adapted assessment measure for remote telehealth administration through COVID-19 and beyond. Journal of Autism and Developmental Disorders. 2022;52(12):5383–5394. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Bishop SL, Lord C. Commentary: Best practices and processes for assessment of autism spectrum disorder–the intended role of standardized diagnostic instruments. Journal of Child Psychology and Psychiatry. 2023;64(5):834–838. [ DOI ] [ PubMed ] [ Google Scholar ] 11. Brunson McClain M, Roanhorse TT, Harris B, Heyborne M, Zemantic PK, Azad G. School-based autism evaluations in the COVID-19 era. School Psychology. 2021;36(5):377. [ DOI ] [ PubMed ] [ Google Scholar ] 12. Krach SK, Paskiewicz TL, Monk MM. Testing our children when the world shuts down: Analyzing recommendations for adapted tele-assessment during COVID-19. Journal of Psychoeducational Assessment. 2020;38(8):923–941. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Zwaigenbaum L, Bishop S, Stone WL, et al. Rethinking autism spectrum disorder assessment for children during COVID-19 and beyond. Autism Research. 2021;14(11):2251–2259. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Farah MJ, Wilson KD, Drain M, Tanaka JN. What is” special” about face perception? Psychological review. 1998;105(3):482. [ DOI ] [ PubMed ] [ Google Scholar ] 15. Young AW, Hellawell D, Hay DC. Configurational information in face perception. Perception. 1987;16(6):747–759. [ DOI ] [ PubMed ] [ Google Scholar ] 16. Bassili JN. Emotion recognition: the role of facial movement and the relative importance of upper and lower areas of the face. Journal of personality and social psychology. 1979;37(11):2049. [ DOI ] [ PubMed ] [ Google Scholar ] 17. Bourke L, Lingwood J, Gallagher-Mitchell T, López-Pérez B. The effect of face mask wearing on language processing and emotion recognition in young children. Journal of Experimental Child Psychology. 2023;226:105580. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Chester M, Plate RC, Powell T, Rodriguez Y, Wagner NJ, Waller R. The COVID-19 pandemic, mask-wearing, and emotion recognition during late-childhood. Social Development. 2023;32(1):315–328. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Ruba AL, Pollak SD. Children’s emotion inferences from masked faces: Implications for social interactions during COVID-19. Plos one. 2020;15(12):e0243708. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Gori M, Schiatti L, Amadeo MB. Masking emotions: Face masks impair how we read emotions. Frontiers in Psychology. 2021:1541. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. Calbi M, Langiulli N, Ferroni F, et al. The consequences of COVID-19 on social interactions: an online study on face covering. Scientific reports. 2021;11(1):2601. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Witkower Z, Tracy JL. How and why head position changes the perception of facial expressions of emotion. Emotion. 2022;22(5):931. [ DOI ] [ PubMed ] [ Google Scholar ] 23. Sowden S, Schuster BA, Keating CT, Fraser DS, Cook JL. The role of movement kinematics in facial emotion expression production and recognition. Emotion. 2021;21(5):1041. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Ong JH, Liu F. Frequent experience with face coverings for 10 months improves emotion perception among individuals with high autistic traits: A repeated cross-sectional study. Quarterly Journal of Experimental Psychology. 2023;76(7):1599–1608. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Haith MM, Bergman T, Moore MJ. Eye contact and face scanning in early infancy. Science. 1977;198(4319):853–855. [ DOI ] [ PubMed ] [ Google Scholar ] 26. Carnevali L, Gui A, Jones EJ, Farroni T. Face processing in early development: a systematic review of behavioral studies and considerations in times of COVID-19 pandemic. Frontiers in Psychology. 2022;13:778247. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Hunnius S, Geuze RH. Developmental changes in visual scanning of dynamic faces and abstract stimuli in infants: A longitudinal study. Infancy. 2004;6(2):231–255. [ DOI ] [ PubMed ] [ Google Scholar ] 28. Lewkowicz DJ, Hansen-Tift AM. Infants deploy selective attention to the mouth of a talking face when learning speech. Proceedings of the National Academy of Sciences. 2012;109(5):1431–1436. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Tenenbaum EJ, Shah RJ, Sobel DM, Malle BF, Morgan JL. Increased F ocus on the M outh A mong I nfants in the F irst Y ear of L ife: AL ongitudinal E ye-T racking S tudy. Infancy. 2013;18(4):534–553. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 30. Chawarska K, Lewkowicz D, Feiner H, Macari S, Vernetti A. Attention to audiovisual speech does not facilitate language acquisition in infants with familial history of autism. Journal of Child Psychology and Psychiatry. 2022;63(12):1466–1476. [ DOI ] [ PubMed ] [ Google Scholar ] 31. Hurwitz S, Yirmiya N, Patel V. Autism diagnostic observation schedule (ADOS) and its uses in research and practice. Comprehensive guide to autism. 2014:345–353. [ Google Scholar ] 32. Spain D, Stewart GR, Mason D, et al. Autism diagnostic assessments with children, adolescents, and adults prior to and during the COVID-19 pandemic: a cross-sectional survey of professionals. Frontiers in Psychiatry. 2022;13:789449. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 33. Mullen EM. Mullen scales of early learning [measurement instrument]. AGS Circle Pines, MN; 1995. [ Google Scholar ] 34. Lord C, Luyster R, Gotham K, Guthrie W. Autism diagnostic observation schedule, (ADOS-2) manual (Part II): Toddler module. Western Psychological Services, CA. 2012; [ Google Scholar ] 35. Sparrow SS, Balla DA, Cicchetti DV, Harrison PL, Doll EA. Vineland adaptive behavior scales. 1984; [ Google Scholar ] 36. Ho D, Imai K, King G, Stuart E, Whitworth A. Package ‘MatchIt’. Version[Google Scholar]. 2018; [ Google Scholar ] 37. Bang JY, Sharda M, Nadig AS. Towards greater transparency in neurodevelopmental disorders research: use of a proposed workflow and propensity scores to facilitate selection of matched groups. Journal of Neurodevelopmental Disorders. 2020;12(1):1–15. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 38. Stuart EA, Rubin DB. Best practices in quasi-experimental designs. Best practices in quantitative methods. 2008:155–176. [ Google Scholar ] 39. Rubin DB. Using propensity scores to help design observational studies: application to the tobacco litigation. Health Services and Outcomes Research Methodology. 2001;2:169–188. [ Google Scholar ] 40. Posar A, Visconti P, Giuberti V. The impact of the COVID-19 pandemic on the assessment of autism spectrum disorder. Turkish Archives of Pediatrics. 2021;56(3):280. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supinfo NIHMS2080341-supplement-Supinfo.docx (507KB, docx) ACTIONS View on publisher site PDF (490.4 KB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Closed-Loop Oxygen to Verify That Healthcare Workers Interventions Decrease During SARS-CoV-2 Pneumonia (COVID-19)#841456
• Psychological Impact of Autism Screening on Caregivers#1119956
• Assessing Performance of the Testing Done Simple Covid 19 Antigen Test#807040
• Plataforma web para la detección de autismo utilizando ciencia de datos: una herramienta innovadora para la evaluación temprana#236786
• Registry of Seraph®-100 Microbind® Affinity Blood Filter for COVID-19 Under EUA#1047615