Recruitment and participant baseline characteristics

A total of 55 participants with AS (aged 1–4 years: n = 16, 5–12 years: n = 27, ≥ 18 years: n = 12) and 20 TDC aged 1–12 years were enrolled over a 12-month period. Seventy-five percent of TDC (n = 15) were co-enrolled with a sibling with AS. Fifty-nine of the 75 participants (79%) were enrolled in the first 6 months of the study (Sept 2019–March 2020), prior to the beginning of the COVID-19 pandemic; the last participant was enrolled in September 2020.

All 75 study participants completed Clinic Visit 1 after screening confirmation (there were no screening failures) and 71 participants completed Clinic Visit 2 (51 in clinic; 20 remote). Four participants with AS discontinued the study prematurely and did not complete Clinic Visit 2 (three participants were identified as lost-to-follow-up and one participant had a major protocol deviation [participant enrolled in an investigational drug trial whilst enrolled in FREESIAS]). The mean (± SD) study duration was 362 ± 86 days, ranging from 222–615 days. Nineteen of 55 participants with AS were co-enrolled in the AS-NHS.

Baseline demographics and genetic characteristics are provided in Table 1. The mean (± SD) age at enrollment for participants with AS was 2.9 (± 0.9), 8.3 (± 2.1), and 24.9 (± 5.9) years for the 1–4 year-old, 5–12 year-old, and ≥ 18 year-old groups, respectively, and 6.5 (± 3.4) years for TDC. In total, 40/55 (73%) had deletion AS and 15/55 (27%) had nondeletion AS. Nondeletion genotypes included UPD (n = 5; 9%), UBE3A mutation (n = 5; 9%), and ID (n = 4; 7%); one participant classified as either UPD or ID as a more specific diagnosis was not obtained.

Table 1 Baseline demographics and genetic characteristics

At baseline, gastrointestinal disorders were reported by 87% of those with AS (48/55), the most common condition being gastroesophageal reflux disease in 60% (33/55) of participants with AS followed by constipation (55%; 30/55). Eye disorders were reported in 47% (26/55) of participants with AS, with strabismus being the most commonly reported (40%; 22/55). Psychiatric disorders were reported in 66% (36/55) of participants with AS including insomnia (15/55) and anxiety (7/55; Table S1). As part of the seizure history assessment a total of 75% (41/55) of participants with AS had epilepsy including generalized epilepsy (34%; 14/41); focal epilepsy (24%; 10/41); and combined generalized and focal epilepsy (39%; 16/41); the epilepsy type of one participant was unknown (Table S2).

Even though no formal anchor was provided (see Discussion), ratings were provided by placing the participant in the context of AS i.e., “mildly ill” when compared to the general AS population known to the expert clinician. The most common CGI-S response for participants with AS aged 1–4 years (31%; 5/16) was a score of 3 (Mildly ill), while the most common response for participants with AS aged 5–12 years (37%; 10/27) was a score of 4 (Moderately ill). Among participants with AS aged ≥ 18 years, an equal number responded with a score of 4 (Moderately ill, 42%; 5/12) and 5 (Markedly ill, 42%; 5/12). Participants with deletion AS were most frequently rated with a score of 4 (35%; 14/40), whereas those with nondeletion AS most frequently received a score of 3 (47%; 7/15). Overall, participants with deletion AS were more frequently rated with a score of 5 (Markedly ill, 28%; 11/40) or 6 (Severely ill, 10%; 4/40), compared with participants with nondeletion AS (Markedly ill, 13%; 2/15 or Severely ill, 0%; 0/15).

Feasibility and adherence of COAs, DHTs, and overnight EEGClinical outcome assessments

An overview of the completion rate from all COAs can be found in Table 2. A completion rate of 89–100% was obtained for COAs at Clinic Visit 1 in participants with AS. At Clinic Visit 2, Early Withdrawal, or Remote Visits, the completion rate was 76–91% apart from Bayley-III, which had a lower completion rate of 62%, since it requires in-person administration (which was not possible at some sites in some cases due to COVID-19 pandemic restrictions). TDC demonstrated adherence of 95–100% during both Clinic Visit 1 and 2.

Table 2 Overall completion rates for clinical outcome assessments

Baseline data for all assessments (presented as raw scores, unless stated otherwise) by genotype and age are shown in Table 3. While many assessments showed different levels of mean variability, participants with nondeletion AS tended to have higher scores (i.e., higher performance/less impairment) than participants with deletion AS across assessments. Age-dependent differences were also evident, with those aged 5–12 years generally having higher scores compared with those aged 1–4 years. There were largely no further gains in scores for those aged ≥ 18 years. Cross-sectional comparison of participants with AS aged ≥ 18 years showed similar characteristics to those aged 5–12 years.

Table 3 Baseline COA data by genotype and ageDigital health technologiesSeizure diary

A total of 631 unique seizure events among 18 participants were reported via the seizure diary throughout the study. Some caregivers provided incomplete seizure information. For example, some answers were left blank for recovery time (17%; 105/631 events), type of seizure (14%; 88/631 events), and seizure duration (12%; 74/631 events). A summary of the characteristics of participants and reported seizures is provided in Table S3.

Seizure diary adherence could not be determined as no specific means to track confirmations about the absence of seizures within the app was available. Consequently, it is likely that seizures were underreported. For most participants reporting seizures, there were more seizures reported in the first 90 days of participation than in the last 90 days (see Fig. S1). This suggests that adherence to the seizure diaries declined over time, since there is no obvious reason to assume fewer seizures in the second part of the observation period.

The frequency of seizures reported through the seizure diary was consistent with that reported at baseline in most participants who completed the diary (61%; 11/18). There were 24 individuals out of 41 (59%) with a history of epilepsy reported at baseline who did not report seizures through the seizure diary app nor as an adverse event during the observation period. Three individuals that did not report a history of epilepsy at baseline developed a first seizure during the study (one participant with nondeletion AS age 4.4 years; two participants with deletion AS ages 1.6 and 2.2 years).

Sleep diary

Caregivers were asked to report every morning on the participants’ sleep during the study. Caregiver adherence to the sleep diary was defined as the percentage of days per week within the 52-week observational period in which they completed the diary. The number of participants included in the adherence calculation decreased over the 52 weeks due either to withdrawal or early completion. Mean adherence was 52% (interquartile range [IQR]: 34–72%) for participants with AS and 47% (IQR: 34–70%) for TDC (see Fig. 2A, B). Mean adherence was higher in the first 26 weeks than in weeks 27–52, respectively (participants with AS: 56% [IQR: 42–75%] vs. 48% [IQR: 20–74%]; TDC: 51% [IQR: 43–73%] vs. 42% [IQR: 16–67%]).

Fig. 2

Adherence/completed assessments for sleep diaries, sleep mat, actigraph, and overnight EEG. Adherence to sleep diary completion in participants with AS (A) and TDC (B). Adherence was defined as the percentage of days within the 52-week observational period in which a diary entry was recorded. Adherence to the sleep mat in participants with AS (C) and TDC (D). Shaded area in graphs A–D represents ± 1 SD. Sleep mat usage was defined as the percentage of days per week within the 52-week observational period in which a sleep mat recording was started. The number of participants included in the adherence calculation for both the sleep diary and sleep mat decreased over the 52 weeks due to either withdrawal, rollover to clinical drug trials, or early completion (see “N” above x-axis). E The percentage of nights that the actigraph was used in the Actigraphy Sleep Monitoring Period for those participants for which any data were collected (AS deletion n = 31; AS nondeletion n = 11; TDC n = 13). F Representation of wearing location of the actigraph during the home monitoring period with unknown location (black); right wrist (red); right ankle (orange); left wrist (yellow); left ankle (green); chest (light blue); abdomen (dark blue); and no actigraphy (black). The actigraph was worn for up to 10 days at home preceding and up to 2 days after the first home EEG visit. G Number of participants who completed overnight EEG recordings with usable data of at least 5 h duration at Home Visit 1, 2, and 3. AS Angelman syndrome, EEG Electroencephalogram, SD Standard deviation, TDC Typically developing children

Sleep mat

A sleep mat placed under each participant’s mattress was used to record sleep daily over the course of the study. The actual use period of the sleep mat was defined as the percentage of days per week in which a sleep mat recording was started and was calculated for every week of the 52-week observational period. The number of participants included in the calculation decreased over the 52 weeks due to either withdrawal or early completion (see Fig. 2C, D). The mean (IQR) actual use period was 59% (42–87) for participants with AS and 53% (32–74) for TDC. For participants with AS, mean (IQR) adherence was higher in the first 26 weeks than in weeks 27–52 (63% [43–90] vs. 56% [30–86]); this was not the case for TDC (53% [33–81] vs. 55% [37–76], respectively). Reasons for nonuse were not collected systematically, but the following reasons were identified by the investigational sites: participant interfered with device (n = 5); participant slept in different location (n = 3); caregiver overwhelmed and did not have time (n = 4); sleep mat made bed uncomfortable (n = 3); sleep mat caused participant distress (n = 2); and sleep mat was incompatible with bed (n = 1).

Sleep actigraphy

Sleep actigraphy recordings were planned for each night between 10 days before, up to, and including Home Visit 1, and up to 2 days following Home Visit 1. Due to COVID-19, 20/75 participants did not have a first home visit and therefore were not asked to use the actigraph. Participant adherence to sleep actigraphy was calculated for the remaining 55 participants, defined as the percentage of days within this period on which the actigraph was worn for ≥ 30 min. Mean (IQR) adherence during this period was 56% (0–100) for participants with AS and 44% (2–98) for TDC (see Fig. 2E). Participants were given the option of wearing the actigraph on either the wrist or ankle, or in a pocket on the chest or abdomen and caregivers were asked to enter the wear position in the sleep diary. Some caregivers recorded different positions for different nights (see Fig. 2F). For 16/55 (29%) participants with AS, no actigraph recording was made during Home Visit 1. Reasons for nonadherence were not collected systematically, but the following were noted based on site follow-up with caregivers: not feasible for the caregiver (n = 3); participant did not tolerate actigraph (n = 3); and participant interfered with device (n = 1).

Overnight electroencephalography at home

Ninety out of 225 (40%) of the planned home visits were attempted and led to 85 successful overnight EEG recordings, defined as > 5 h of data per EEG recording; mean (± SD) duration of 14 (± 4.6) hours. The EEG recording period was purposefully ~ 14 h long on an average, as it would start in the afternoon to capture participants’ awake EEG data. Home visits were not carried out for reasons related to the COVID-19 pandemic, e.g., it was legally not possible, the EEG vendor was not able to provide service, or the caregiver did not agree to the home visit due to COVID-19. EEG recordings were obtained from 47/75 participants at Home Visit 1, including 13 TDC and 34 individuals with AS (Fig. 2G; Table S4).

The quality of the scalp EEG data was considered sufficient by EEG experts (neuroscientists and neurologists) in most cases to perform quantitative analyses and to identify waking and sleep background elements, epileptiform discharges (spikes and sharp waves), and epileptic seizures.

FREESIAS data in the context of the AS-NHS data

Comparison of raw scores of Bayley-III baseline data collected in FREESIAS and the AS-NHS [44] was performed only on the 1–12 years age range due to limited data availability on adult individuals with AS in the AS-NHS (Figs. 3A–E; Table S5). For FREESIAS versus AS-NHS data, the mean (± SD) raw score for the Cognitive domain in participants with AS aged 1–4 years was 35.5 (± 7.4) vs. 41.5 (± 10.3) and 49.9 (± 11.1) vs. 51.1 (± 11.4) for those aged 5–12 years. The mean (± SD) raw score for the Receptive Communication domain was 12.9 (± 4.2) vs. 13.8 (± 4.5) in participants with AS aged 1–4 years and 17.4 (± 5.6) vs. 17.6 (± 7.6) for those aged 5–12 years. For the Expressive Communication domain, the mean (± SD) raw score was 9.0 (± 3.7) vs. 10.4 (± 3.8) in participants with AS aged 1–4 years and 11.9 (± 4.3) vs. 12.7 (± 5.2) for those aged 5–12 years. The mean (± SD) raw score for the Fine Motor domain was 24.1 (± 3.8) vs. 26.6 (± 5.4) in participants with AS aged 1–4 years and 29.6 (± 7.3) vs. 32.5 (± 8.8) for those aged 5–12 years, while for the Gross Motor domain, mean (± SD) raw scores were 36.6 (± 10.7) vs. 38.1 (± 9.4) in participants with AS aged 1–4 years and 49.3 (± 6.7) vs. 48.8 (± 6.6) for those aged 5–12 years (Table S5). A time-to-event analysis was performed to evaluate seizure onset age in participants with AS, and the results were compared with the AS-NHS data [26, 44] (see Fig. 3G). Results from the FREESIAS study indicated that participants with deletion AS have an earlier seizure onset and a higher seizure prevalence overall, consistent with previous findings from the AS-NHS study [26, 44].

Fig. 3

Bayley-III profiles obtained using FREESIAS and AS-NHS data. Mean Bayley-III raw scores in participants with AS deletion and nondeletion aged 1–12 years, from FREESIAS and the AS-NHS: A Cognitive domain, B Expressive Communication domain, C Receptive Communication domain, D Fine Motor domain, E Gross Motor domain. F EEG data in participants with AS deletion, nondeletion, and TDC from FREESIAS, AS-NHS, and BCH. Axes are plotted on logarithmic scales. G Kaplan–Meyer analysis of cumulative seizure data, stratified by underlying genotype and data source, in participants with AS deletion and nondeletion from FREESIAS and the AS-NHS. The table indicates the number of participants at risk for each group at 12-month intervals. One FREESIAS participant that presented their first seizure at over 18 years old was excluded from this analysis in order to allow for a direct comparison to the AS-NHS analysis that included pediatric population exclusively. AS Angelman syndrome, AS-NHS Angelman Syndrome Natural History Study, Bayley-III Bayley Scales of Infant and Toddler Development® – Third Edition, BCH Boston Children’s Hospital, EEG Electroencephalogram, TDC Typically developing children

Delta power derived from 10 min of awake EEG recorded in the afternoon of the first day was quantified for baseline recordings. Excess EEG delta power was detected for participants with AS compared with TDC, in line with previous reports (see Fig. 3F). The data also quantitatively matched previous ~ 30-min awake state recordings in participants with AS, thus confirming that usable EEG data were collected in the home setting in the present study [16, 17].

Study expectations and feedback

Insights from caregivers of individuals with AS (N = 54) were obtained from multiple choice feedback questionnaires (Fig. 4; Table S6). Understanding the motivation of caregivers of individuals with AS to participate in a nondrug observational study supports the design and execution of future clinical trials. Based on multiple choice feedback questionnaires collected from these caregivers, key motivators for choosing to participate in the study included: contributing to medical research (93%; 50/54); getting more treatment options (82%; 44/54); gaining a better understanding of AS (76%; 41/54); and raising awareness of AS in the community (41%; 22/54). In caregivers of TDC (N = 20), the main motivating factors to participate in the study included contributing to medical research (95%; 19/20) and getting more treatment options for individuals with AS (60%; 12/20).

Fig. 4

Study expectations and output from feedback questionnaires. EEG electroencephalogram

In total, 74 caregivers completed the questionnaire about the in-clinic visits, including 54 caregivers of individuals with AS and 20 of TDC, although the number of caregivers responding to different parts of the questionnaire differed depending on the question, their previous involvement in clinical research, and willingness to answer. Most of the caregivers of individuals with AS (86%; 37/43) and all of those of TDC (100%; 17/17) expressed that they were satisfied or very satisfied with the experiences of those participants in completing the in-clinic assessments. Most of the caregivers of individuals with AS also deemed the duration of clinic visits for the study to be acceptable (79%; 34/43) and were satisfied or very satisfied (95%; 41/43) with the support given during the clinic visits. Importantly, due to the large number of assessments at the clinical visits, caregivers could choose to perform the assessments over 2 days, which was preferred by a third of all caregivers (35%; 21/60).

In total, 50 caregivers reported on their at-home visit experiences, including 38 caregivers of individuals with AS and 12 of TDC. Home visits were deemed to be a slight (45%; 17/38) or moderate (29%; 11/38) burden.

When reporting on the ease of smartphone use, 28/43 (65%) of the caregivers of individuals with AS were satisfied or very satisfied with it, 33/43 (77%) rated the frequency of smartphone assessments as acceptable, and 42/43 (98%) deemed the time required to complete these assessments to be acceptable. Further answers to the smartphone part of the questionnaire revealed that 59% (22/37) of the respondent caregivers of individuals with AS were very satisfied or satisfied with the experience of the at-home EEG assessment. This may relate to the percentage of caregivers who found the sleep of the child slightly impacted (31%; 11/35) or very/extremely impacted (54%; 19/35). Also, in contrast to the aforementioned slight/moderate burden reported for at-home visits, caregivers of individuals with AS who commented on their preference for at-home vs. in-clinic visits. (n = 40) largely preferred home as the future visit setting (75%; 30/40).

When reporting on the at-home sleep devices, 91% (39/43) of the respondent caregivers of individuals with AS were very satisfied or satisfied with the sleep mat; the score was higher than the actigraph, for which 64% (23/36) of the respondent caregivers of individuals with AS reported that they were satisfied or very satisfied.

In total, 79% (34/43) of caregivers of individuals with AS who commented on their overall experience of the study were satisfied or very satisfied with their experience, with only 9% (4/43) being dissatisfied or very dissatisfied. The main reasons stated for this dissatisfaction included technical issues with some of the DHTs and the request to use digital rather than paper assessments.