Parental Stress and Family Quality of Life in Families of Individuals Living With Angelman Syndrome

Abstract

Background:

Angelman syndrome (AS) is a developmental disorder caused by one of four molecular aetiologies. Affected individuals have intellectual disability (ID), limited speech, seizures and sleep problems. Parents of individuals with AS exhibit elevated stress compared to parents of individuals with other IDs. We examined parental stress and family quality of life (FQOL) over time in families of individuals living with AS.

Methods:

Data were collected in a natural history study of AS. The Parenting Stress Index, Third Edition (PSI) and the Beach Center FQOL Scale assessed parental stress and FQOL. Stress and FQOL were examined across AS molecular subtypes, and predictors were analysed using a generalised linear model. Relationships between parental stress and FQOL were examined using Pearson correlations and a stepwise mixed-linear model approach.

Results:

Our sample consisted of 231 families of individuals living with AS. Parental stress was clinically elevated and was highest in families of individuals with UBE3A pathogenic variants, whereas FQOL did not differ across subtypes in most domains. Increasing age predicted a decrease in parental stress but did not predict FQOL. Elevated parental stress was additionally predicted by maladaptive behaviours and child male sex, whereas lower FQOL was predicted by child male sex, parent marital status and family income. Parental stress had a small negative impact on FQOL.

Conclusions:

Stress is elevated in parents of individuals with AS across subtypes and has a negative impact on FQOL. Interventions to reduce stress have the potential to improve individual and family well-being.

1 ∣. Background

Angelman syndrome (AS) is a rare neurogenetic disorder caused by loss of expression of UBE3A, which may be due to a deletion of the AS critical region on maternally inherited chromosome 15q11-q13 (65%–70%) (deletion subtype), or paternal uniparental disomy (UPD) for chromosome 15q11-q13 (5%–10% of cases), imprinting centre defects (ImpD) (5%–10%), or UBE3A pathogenic variants (10%–15%) (collectively known as the non-deletion subtypes) (Bird 2014). Deletions are further subclassified based on their size as Class I (5.9 Mb, 40%), Class II (5.0 Mb, 53%) or atypical (7%) (Bird 2014). AS is characterised by severe intellectual disability (ID), lack of speech, seizures and sleep difficulties (Williams et al. 2006). At 6 years of age, AS individuals exhibit cognitive developmental age between 15 and 27 months; individuals with a deletion subtype have more severe impairments than individuals with a non-deletion subtype (Sadhwani et al. 2023). Maladaptive behaviours, including hyperactivity, irritability and aggression, are common, especially among individuals with non-deletion subtypes (Clarke and Marston 2000; Sadhwani et al. 2019). Sleep difficulties (increased sleep latency and frequent nocturnal awakenings) and epilepsy are reported in ~90% of individuals with AS (Bird 2014; Pelc et al. 2007). Communication, sleep and behaviour challenges negatively impact quality of life of individuals and their caregivers and increase family burden (Wheeler et al. 2017). Treatment is currently limited to symptomatic interventions, although promising disease-modifying therapies are in development (Keary and McDougle 2023).

Parents of individuals with AS report elevated stress compared to parents of neurotypical individuals and individuals with other IDs and autism (Griffith et al. 2011a; Thomson et al. 2017; van den Borne et al. 1999; Wulffaert et al. 2010; Hagenaar et al. 2024). A previous study comparing parental stress across AS molecular subtypes found that clinically elevated stress was more common among parents of individuals with UPD or ImpD; parents of individuals with UBE3A pathogenic variants were not included (Miodrag and Peters 2015). Other studies support an association between behaviour and sleep difficulties and parental stress, but the impact of impaired cognitive and adaptive functioning, epilepsy and other medical complications is unclear (Sadhwani et al. 2019; Wulffaert et al. 2010; Miodrag and Peters 2015; Goldman et al. 2012). In the general population, demographic and psychosocial factors, including child sex, maternal education, family income, race and ethnicity may impact parental stress (Fang et al. 2024; Cardoso et al. 2010; Walkowiak and Domaradzki 2025a).

Research has demonstrated that parents of children with AS experience elevated stress, anxiety, physical symptoms such as chronic pain and unique caregiving burdens that can negatively impact both parental well-being and overall family functioning (Wheeler et al. 2017; Griffith et al. 2011a; Miodrag and Peters 2015; Walkowiak and Domaradzki 2025a; Ferreira et al. 2022; Griffith et al. 2011b; Walkowiak and Domaradzki 2025b; Pfarrer et al. 2025). Despite this, little is known about the quality of life of families of individuals living with AS. Family quality of life (FQOL) refers to the diverse ways by which having a family member with an ID affects family dynamics and functioning (Summers et al. 2005; Hoffman et al. 2006). Family functioning is a key component of caregiver well-being. For example, limited family support or feelings of guilt associated with balancing multiple caregiving roles are significant sources of stress for parents of children with IDs (Barratt et al. 2025). Fostering family support is widely accepted as an important aspect of effective ID services and interventions (Summers et al. 2005). As such, assessing and addressing FQOL can enhance our understanding of how caregiving burden affects families, help identify risk and protective factors influencing family functioning and guide the development of effective, family-centred interventions and support systems for families of individuals with AS (Summers et al. 2005).

Moreover, there are limited data on how parental stress or FQOL in families of individuals living with AS change over time or across different age groups, and the relationship between parental stress and FQOL has not yet been described. This study sought to examine predictors of, and associations between, parental stress and FQOL across age and molecular subtypes in AS. The aims of this study were to:

1. Describe parental stress and FQOL in families of individuals living with AS across molecular subtypes and over time.

2. Describe predictors of parental stress and FQOL.

3. Analyse how parental stress correlates with FQOL.

2 ∣. Methods

2.1 ∣. Participants

Participants were enrolled in the AS Natural History Study (NHS) (ClinicalTrials.gov identifier: NCT00296764) at one of six study sites: Rady Children's Hospital-San Diego; Texas Children's Hospital; Greenwood Genetics Center; Boston Children's Hospital; Vanderbilt University Medical Center; and Cincinnati Children's Hospital Medical Center between 2006 and 2014. Participants underwent annual evaluations of cognition, adaptive functioning and behaviour. Included participants were genetically confirmed to have AS (deletion class 1, deletion class 2, UPD, ImpD or UBE3A pathogenic variant). Data from participants 22 years of age and over were excluded from this analysis. Age 22 was selected because it is the age at which individuals with disabilities lose eligibility for special education services in the state of the lead study site (603 CMR 28:00 Special Education n.d.). In instances where parent-reported measures were completed by different caregivers, we included measures completed by the caregiver who attended the majority of visits. The term ‘parent’ will be used to refer to a participant's primary caregiver.

2.2 ∣. Ethics Approval

The NHS was approved by institutional review boards of each participating institution. Informed consent was obtained from a legally authorised representative for each participant.

2.3 ∣. Measures

2.3.1 ∣. Use of Standardised Measures Outside Age Range

Some measures to assess parental stress and development are standardised for use in children of specific ages. However, these measures were administered to individuals of all ages in this study because the developmental age of all the participants in this study fell within the normative age range for which these measures were developed (Sadhwani et al. 2023).

2.3.2 ∣. Parental Stress and FQOL

Parental stress was assessed using the Parenting Stress Index, Third Edition (PSI) (Abidin 1995). The PSI is a parent-completed measure with 120 items in three domains: child characteristics, parent characteristics and situational/demographic life stress. The Child Domain includes six subdomains: Distractibility/Hyperactivity, Adaptability, Reinforces Parent, Demandingness, Mood and Acceptability. The Parent Domain includes seven subdomains: Competence, Isolation, Attachment, Health, Role Restriction, Depression and Spouse/Parenting Partner Relationship. The Life Stress Domain does not contribute to the Total Stress score and was excluded from this study. Parents rate items on a 5-point scale (strongly agree, agree, not sure, disagree, strongly disagree), and domain scores and a total score are generated. Higher scores indicate higher parental stress; scores above the 85th percentile indicate clinically significant stress. The PSI has been reported in several AS studies (Wulffaert et al. 2010; Miodrag and Peters 2015; Goldman et al. 2012). It is standardised for use among parents of children aged 1 month to 12 years.

FQOL was assessed using the Beach Center Family Quality of Life Scale (FQOL Scale) (Hoffman et al. 2006). The FQOL Scale is a parent-report measure that assesses the importance of, and satisfaction with, five domains of quality of life: Family Interaction, Parenting, Emotional Well-Being, Physical/Material Well-Being and Disability-Related Support. Since the aim of our study was to assess families' well-being rather than the perceived relevance of factors that may impact their well-being, only the satisfaction scale was examined for all five domains, consistent with prior studies (Boehm et al. 2015; Staunton et al. 2020a; Jenaro et al. 2020). Parents rate 25 items on a 5-point scale (very dissatisfied, satisfied, neither, satisfied, very satisfied), with higher scores indicating higher levels of satisfaction. Scores are calculated as the mean of each subscale (Werner et al. 2009). The FQOL Scale has been used in studies of individuals with developmental disorders (Boehm et al. 2015; Fong et al. 2020; Hsiao et al. 2017) and in one other AS study (Sadhwani et al. 2019).

2.3.3 ∣. Development, Adaptive and Behavioural Functioning

Development and adaptive functioning were assessed with the Bayley Scales of Infant and Toddler Development, Third Edition (BSID-III) (Sadhwani et al. 2023; Bayley 2005) and the Vineland Adaptive Behavior Scales, Second Edition (VABS-II) (Sparrow et al. 2005; Gwaltney et al. 2024), respectively. The BSID-III assesses cognitive, receptive and expressive language and fine and gross motor skills using standardised activities administered by a trained rater. Raw and age equivalent (AE) scores are generated for each domain. This measure is standardised for use in children up to 42 months. The VABS-II is a clinician-administered parent interview that assesses adaptive functioning in three domains: communication, daily living skills and socialisation; for individuals under the age of 7 years, a fourth domain (motor skills) is also assessed. The motor skills domain was included for all individuals in this study. Domain standard and AE scores, as well as an overall Adaptive Behavior Composite (VABS-II composite) score, are generated.

Maladaptive behaviours were assessed using the Aberrant Behavior Checklist-Community version (ABC-C) (Aman et al. 1985). This measure has 58 items rated on a scale of 0–3, with 0 indicating a behaviour is not a problem and 3 indicating a behaviour is a severe problem. The ABC-C has five domains: Irritability, Agitation, Crying; Lethargy and Social Withdrawal; Stereotypic Behavior; Hyperactivity and Non-compliance; and Inappropriate Speech (Aman et al. 1985). The Inappropriate Speech domain was excluded from this study as the majority of participants in the study are non-verbal. Previous studies have shown significant correlation between ABC-C Irritability and Hyperactivity scores and parental stress levels, whereas ABC-C Stereotypy and Lethargy scores have weak associations with parental stress (Sadhwani et al. 2019). Therefore, only the Irritability and Hyperactivity domains were analysed. The ABC-C has been used in several previous studies of individuals with AS (Clarke and Marston 2000; Sadhwani et al. 2019; Summers et al. 2005).

2.3.4 ∣. Demographic and Medical Characteristics

Demographic data were collected at baseline. A structured medical history was collected at each visit to assess seizures and other medical problems. Participants' sleep habits, including average number of night awakenings, were assessed via a structured parent interview by study investigators.

2.4 ∣. Statistical Methods

2.4.1 ∣. Participant Data and Description of PSI and FQOL Scores

Descriptive analyses were performed on demographic variables, baseline medical and developmental variables and baseline PSI and FQOL Scale scores. Baseline visit is defined as the first visit at which the PSI and/or FQOL Scales were completed. The percentage of respondents with PSI Total Stress and Child and Parent Domain stress scores above the 85th percentile was calculated. Characteristics across subtypes were compared using Pearson chi-square, Fisher's exact or one-way analysis of variance (ANOVA) tests; if significant, post hoc Tukey's honestly significant difference (HSD) test was conducted to determine which subtype groups were significantly different.

2.4.2 ∣. Predictors of PSI and FQOL Scores

We investigated predictors of PSI Child Domain stress, Parent Domain stress, Total Stress and FQOL subscales using generalised linear models (GLM), with adjustments for repeated measures in order to account for within-subject correlation. This was achieved by modelling the subject effect as part of the error structure rather than as a fixed effect. For ease of interpretation, age was centred at 6 years. Molecular subtype, participant sex, age at visit, race, ethnicity, annual household income, number of persons in household, number of siblings, maternal education, parent marital status, seizure severity, average number of hospitalisations per year, frequency of nocturnal awakenings, VABS-II Composite and BSID-III Cognitive AE were analysed as predictors. Variables were individually entered into a bivariate model with either PSI or FQOL score as the outcome. Only the significant terms in the bivariate models were then included in the multivariate model to determine if they were still predictors of PSI/FQOL scores when controlling for all other significant terms in the bivariate model. For models with only one significant term, the multivariate model was the same as the bivariate model. To ensure comparability between the bivariate and multivariate models, the bivariate analyses were restricted to data from the subset of observations with complete data on all covariates included in the multivariate model. Post hoc analyses were performed on significant categorical variables to identify differences in least squares means with Bonferroni adjustments for multiple comparisons by subtype. To analyse predictors of FQOL, a log link function was used to account for a non-normal distribution of FQOL scores and predicted coefficients were exponentiated to rescale to the original scale.

2.4.3 ∣. Relationship Between PSI and FQOL Scores

Multiple analyses were conducted to investigate how parental stress contributes to FQOL. PSI Parent Domain, Child Domain and Total Stress scores were included in the GLM when analysing predictors of FQOL, with adjustments for repeated measures, described above. Pearson correlations were calculated to analyse the relationship between baseline PSI Domain scores and FQOL subscale scores. Finally, a stepwise, generalised linear mixed model approach was used to investigate the effects of PSI Child and Parent Domains on FQOL subscales with adjustments for repeated measures. Similar to previous analyses, age was centred at 6 years, a log link function was used, and predicted coefficients were exponentiated to rescale to the original scale. Starting with the unconditional model (a model with no predictors), various covariates, interaction terms and random effects were added and subtracted using the lowest Bayesian information criterion (BIC) to determine the best model for predicting each FQOL subscale. Covariates were molecular subtype, age, ethnicity, parent marital status, seizure severity, VABS-II Composite, BSID-III Cognitive AE, ABC-C Irritability and Hyperactivity scores, molecular subtype by age interaction and molecular subtype by seizure severity interaction. Although age was not statistically significant, age was included in the conditional models as a covariate to explicitly test the effect of time in addition to the effect of molecular subtype.

SAS (Version 7.15; SAS Institute, Cary, NC) was used for all analyses; p < 0.05 was considered statistically significant.

3 ∣. Results

3.1 ∣. Demographics

Data from 231 participants who completed the PSI during study participation were analysed (Table 1). Of those, 214 completed the FQOL Scale. Participants had an average of two to three annual evaluations. The number of completed evaluations ranged from one to nine. The majority of participants were non-Hispanic White, and the majority of parents were married or in marriage-like relationships. No significant differences in demographic parameters were observed across subtypes except for parental marital status: Parents of individuals with ImpD reported highest rates of marriage/marriage-like relationships, whereas parents of individuals with Class I deletions reported lowest rates of marriage/marriage-like relationships (p < 0.001). Medical and developmental covariates are presented in Table 2. Significant differences across subtypes were observed for seizure severity (p < 0.001), VABS-II Composite score (p = 0.004), ABC-C Irritability score (p < 0.001) and BSID-III Cognitive age equivalent score (p < 0.001), consistent with previous studies (Sadhwani et al. 2023; Sadhwani et al. 2019; Thibert et al. 2009; Gentile et al. 2010).

TABLE 1 ∣.

Demographic characteristics at baseline visit (n = 231).

	Class I deletion (n = 62) 26.8%	Class II deletion (n = 92) 39.8%	UBE3A pathogenic variant (n = 28) 12.1%	ImpD (n = 21) 9.1%	UPD (n = 28) 12.1%	p
Sex, n (%)						0.95
Female	34 (54.8%)	47 (51.1%)	14 (50.0%)	10 (47.6%)	16 (57.1%)
Male	28 (45.2%)	45 (48.9%)	14 (50.0%)	11 (52.4%)	12 (42.9%)
Child age at baseline visit
Mean (SD)	4.6 (3.8)	4.9 (3.9)	5.4 (4.1)	5.4 (4.2)	6.8 (4.5)	0.16
Range	1–17	1–20	1–15	2–21	2–21
Child age at last completed visit
Mean (SD)	7.1 (4.8)	7.2 (4.4)	7.4 (4.3)	7.8 (4.2)	8.3 (4.5)	0.75
Range	1–21	1–21	1–19	3–21	2–21
Race (n (%) White)	54 (87.1%)	82 (89.1%)	28 (100.0%)	20 (95.2%)	26 (92.9%)	0.30
Ethnicity (n (%) Hispanic/Latinx)	13 (21.0%)	16 (17.4%)	0 (0%)	2 (9.5%)	5 (17.9%)	0.06
Annual household income, n (%)						0.39
Under $250 000	6 (9.7%)	7 (7.6%)	1 (3.6%)	1 (4.8%)	2 (7.1%)
$25 000–$49 999	13 (21.0%)	14 (15.2%)	4 (14.3%)	1 (4.8%)	1 (3.6%)
$50 000–$99 999	19 (30.7%)	29 (31.5%)	10 (35.7%)	5 (23.8%)	10 (35.7%)
$100 000 or more	18 (29.0%)	33 (35.9%)	8 (28.6%)	12 (57.1%)	13 (46.4%)
Declined/missing	6 (9.7%)	9 (9.8%)	5 (17.9%)	2 (9.5%)	2 (7.1%)
# of persons supported on income, n (%)						0.12
2	2 (3.2%)	3 (3.3%)	1 (3.6%)	0 (0%)	1 (3.6%)
3	23 (37.1%)	17 (18.5%)	6 (21.4%)	6 (28.6%)	4 (14.3%)
4	22 (35.5%)	35 (38.0%)	10 (35.7%)	9 (42.9%)	10 (35.7%)
5	10 (16.1%)	20 (21.7%)	5 (17.9%)	4 (19.1%)	7 (25.0%)
6 or more	4 (6.5%)	10 (10.9%)	1 (3.6%)	1 (4.8%)	4 (14.3%)
Declined/missing	1 (1.6%)	7 (7.6%)	5 (17.8%)	1 (4.6%)	2 (7.1%)
# of persons in household, mean (SD)	4.0 (1.2)	4.4 (1.2)	3.9 (0.9)	4.1 (1.0)	4.2 (1.0)	0.60
# of siblings, mean (SD)	1.4 (1.2)	1.7 (1.2)	1.1 (0.9)	1.8 (1.9)	1.8 (0.1)	1.00
Respondent relation, n (%)						0.35
Mother	47 (75.8%)	74 (80.4%)	23 (82.1%)	16 (76.2%)	22 (78.6%)
Father	7 (11.3%)	7 (7.6%)	1 (3.6%)	4 (19.1%)	5 (17.9%)
Other	1 (1.6%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Maternal Education, n (%)						0.53
Less than college	27 (43.6%)	38 (41.8%)	6 (28.6%)	6 (28.6%)	12 (46.2%)
College or more	35 (56.5%)	53 (58.2%)	15 (71.4%)	15 (71.4%)	14 (53.9%)
Declined/missing	0 (0%)	1 (1.1%)	7 (25.0%)	0 (0%)	2 (7.1%)
Parent marital status, n (%)						< 0.001
Single, never married	5 (8.1%)	5 (5.4%)	0 (0%)	0 (0%)	0 (0%)
Married/living in marriage-like relationship	45 (75.6%)	74 (80.4%)	23 (82.1%)	19 (90.5%)	24 (85.7%)
Separated/divorced	6 (9.7%)	8 (8.7%)	1 (3.6%)		2 (7.1%)
Declined/missing	6 (9.7%)	5 (5.4%)	4 (14.6%)	2 (9.5%)	2 (7.1%)

Abbreviations: ImpD, imprinting centre defect; UPD, uniparental disomy.

TABLE 2 ∣.

Medical and developmental characteristics at baseline visit (unless otherwise specified) (n = 231).

	Class I deletion (n = 62) 26.8%	Class II deletion (n = 92) 39.8%	UBE3A pathogenic variant (n = 28) 12.1%	ImpD (n = 21) 9.1%	UPD (n = 28) 12.1%	p
Seizures, n (%)						< 0.001
None	4 (6.5%)	11 (39.3%)	8 (8.7%)	11 (52.4%)	9 (32.1%)
Mild	30 (48.4%)	12 (42.9%)	43 (46.7%)	4 (19.1%)	8 (28.6%)
Moderate	11 (17.7%)	2 (7.1%)	9 (9.8%)	2 (9.5%)	3 (10.7%)
Severe	1 (1.6%)		7 (7.6%)		2 (7.1%)
Missing/declined	16 (25.8%)	3 (10.7%)	25 (27.2%)	4 (19.0%)	6 (21.4%)
Average number of hospitalisations per year (at last visit), mean (SD)	0.7 (0.6)	0.5 (0.4)	0.6 (0.5)	0.5 (0.7)	0.6 (0.6)	0.80
Waking up greater than two times a week, n (%)						0.42
No	29 (46.8%)	11 (39.3%)	44 (47.8%)	12 (57.1%)	9 (32.1%)
Yes	33 (53.2%)	17 (60.7%)	47 (51.1%)	9 (42.9%)	19 (67.9%)
Declined/missing	0 (0%)	0 (0%)	1 (1.1%)	0 (0%)	0 (0%)
If yes, # awakenings per night						0.11
1	11 (34.4%)	8 (47.1%)	28 (62.2%)	5 (55.6%)	9 (50.0%)
2–4	17 (53.1%)	7 (41.2%)	17 (37.8%)	3 (33.3%)	9 (50.0%)
5–7	2 (6.3%)	2 (11.8%)	0 (0%)	0 (0%)	0 (0%)
> 7	2 (6.3%)	0 (0%)	0 (0%)	1 (11.1%)	0 (0%)
Declined/missing	1 (3.0%)	0 (0%)	2 (4.3%)	0 (0%)	1 (5.3%)
VABS-II Composite score
Mean (SD)	54.8 (12.4)	55.1 (10.8)	62.9 (12.8)	61.4 (8.0)	57.9 (10.8)	0.004
Range	30–82	26–78	42–94	37–81	32–79
Aberrant Behaviour
Checklist—Community (ABC-C)
Irritability
Mean (SD)	3.1 (4.1)	5.1 (6.0)	8.6 (10.0)	5.8 (5.4)	8.6 (6.9)	< 0.001
Range	0–17	0–30	0–33	0–17	1–29
Hyperactivity
Mean (SD)	12.1 (9.1)	15.3 (11.1)	16.9 (14.7)	15.7 (10.2)	19.1 (11.1)	0.06
Range	0–34	0–41	0–48	2–33	3–44
BSID-III cognitive age equivalent (AE) score
Mean (SD) (months)	11.8 (4.2)	19.9 (6.7)	12.7 (3.9)	22.4 (5.8)	20.0 (4.8)	< 0.001
Range	4.0–22.0	4.0–27.0	4.0–28.0	12.0–38.0	12.0–27.0

Abbreviations: BSID-III, Bayley Scales of Infant and Toddler Development, Third Edition; ImpD, imprinting centre defect; UPD, uniparental disomy; VABS-II, Vineland Adaptive Behaviour Scales, Second Edition.

3.2 ∣. Description of PSI and FQOL Across Subtypes and Over Time

Table 3 summarises baseline PSI scores across subtypes. The percentage of respondents with Total Stress scores above the 85th percentile, indicating clinical significance, ranged from 19% in the Class II deletion group to 48% in the UBE3A pathogenic variant group. Forty to sixty percent of respondents reported clinically significant Child Domain stress, and 11%–26% of respondents reported clinically significant Parent Domain stress. Table 4 summarises baseline FQOL scores across molecular subtype. Total mean score ranged from 3.7 to 4.1.

TABLE 3 ∣.

Parenting Stress Index (PSI) scores at baseline visit (n = 231).

	Class I deletion		Class II deletion		UBE3A pathogenic variant		ImpD		UPD		p
	Mean (SD)	%	Mean (SD)	%	Mean (SD)	%	Mean (SD)	%	Mean (SD)	%
Total stress (101–505)	231.3 (35.1)	21	226.3 (34.4)	19	262.0 (47.5)	48	243.1 (32.3)	30	242.6 (33.5)	28	< 0.001
Child Domain
Total Child Domain score (47–235)	115.0 (18.3)	45	114.0 (17.9)	40	128.5 (25.6)	59	123.1 (17.6)	60	123.0 (18.6)	56	0.003
Distractibility/hyperactivity (9–45)	31.6 (4.8)		31.9 (5.4)		32.2 (6.8)		32.8 (6.4)		32.8 (5.0)		0.86
Adaptability (11–55)	26.0 (7.0)		26.3 (5.9)		29.3 (6.8)		30.0 (5.2)		30.0 (7.8)		0.005
Reinforces parent (6–30)	8.7 (3.0)		8.5 (2.8)		9.6 (3.8)		9.1 (2.8)		7.8 (1.8)		0.24
Demandingness (9–45)	22.7 (5.9)		22.2 (5.9)		27.8 (7.9)		24.9 (6.2)		26.2 (6.3)		< 0.001
Mood (5–25)	7.7 (2.7)		7.8 (2.5)		10.5 (3.7)		8.3 (2.5)		8.5 (3.1)		< 0.001
Acceptability (7–35)	18.4 (3.0)		17.3 (3.5)		19.1 (3.9)		18.1 (4.1)		18.8 (4.0)		0.11
Parent Domain
Total Parent Domain score (54–270)	116.4 (22.5)	12	112.7 (22.3)	11	133.4 (25.6)	26	120 (22.1)	20	120.3 (22.2)	15	0.002
Competence (13–65)	24.9 (5.5)		24.9 (6.4)		28.4 (6.0)		24.8 (4.9)		27.2 (5.3)		0.03
Isolation (6–30)	12.6 (4.3)		12.6 (3.5)		14.5 (4.1)		13.6 (4.5)		13.8 (4.3)		0.15
Attachment (7–35)	11.2 (3.2)		11.3 (3.0)		12.1 (3.3)		11.3 (2.6)		11 (2.0)		0.69
Health (5–25)	13.1 (3.2)		12.7 (3.2)		14.3 (4.1)		13.5 (2.9)		13.2 (2.8)		< 0.001
Role restriction (7–35)	19.3 (5.5)		18.1 (5.0)		23.5 (6.6)		20.5 (5.0)		19.9 (6.3)		< 0.001
Depression (9–45)	17.3 (4.7)		16.7 (4.8)		19.4 (5.4)		18.3 (4.5)		18.3 (8.8)		0.19
Spouse/parenting partner relationship (7–35)	17.9 (5.9)		16.4 (4.9)		21.2 (4.9)		18.2 (5.1)		17 (5.2)		0.002

Note: PSI score ranges are included in parentheses.

Abbreviations: %, percentage of respondents who scored above the 85th percentile; ImpD, imprinting centre defect; UPD, uniparental disomy.

TABLE 4 ∣.

Family Quality of Life (FQOL) Scale scores at baseline visit (n = 214).

FQOL score	Class I deletion	Class II deletion	UBE3A pathogenic variant	ImpD	UPD	p
	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
FQOL total score	4.1 (0.6)	4.1 (0.6)	3.7 (0.6)	3.8 (0.6)	4.0 (0.5)	0.10
Subscale
Family interaction	4.2 (0.8)	4.2 (0.8)	3.7 (0.7)	3.9 (0.9)	4.3 (0.6)	0.006
Parenting	4.0 (0.8)	4.1 (0.7)	3.8 (0.7)	4.0 (0.7)	4.0 (0.6)	0.54
Emotional well-being	3.5 (1.0)	3.5 (1.1)	2.8 (1.1)	2.8 (0.9)	3.1 (0.7)	0.005
Physical/material well-being	4.3 (0.7)	4.4 (0.6)	4.3 (0.7)	4.4 (0.7)	4.4 (0.7)	0.71
Disability-related support	4.2 (0.7)	4.2 (0.8)	4.1 (0.9)	3.8 (0.8)	4.0 (0.7)	0.17

Note: FQOL scale is scored on a 5-point scale where 1 = very dissatisfied, 3 = neither satisfied nor dissatisfied and 5 = very satisfied.

Abbreviations: ImpD, imprinting centre defect; UPD, uniparental disomy.

The impact of molecular subtype on PSI and FQOL scores was explored via GLM and post hoc analysis (Tables 5 and 6 and Figures 1-4). Molecular subtype was not a significant predictor of Child Domain stress (Table 5 and Figure 1). UBE3A pathogenic variant subtype predicted higher Parent Domain (p_adj < 0.001) and Total Stress (p_adj = 0.001) compared to the Class II deletion group (Table 5 and Figures 2-3). There were no significant differences in FQOL by molecular subtype except for in the Family Interaction subscale where the UBE3A pathogenic variant group had lower least-square mean scores than the UPD group (Table 6 and Figure 4; p_adj = 0.033). In within-subtype post hoc comparisons, Emotional Well-Being scores were significantly lower than all other FQOL subscales across subtypes (data not shown).

TABLE 5 ∣.

Generalised linear models for predictors of Parenting Stress Index scores.

	Bivariate		Multivariate
	Est. (SE)	p	Est. (SE)	p
PSI Child Domain (n = 222, obs = 463)
Age (years)	0.98 (0.32)	0.002	−0.50 (0.32)	0.12
Diagnosis
Class 1	− 0.61 (2.72)	0.82	2.50 (2.35)	0.29
Class II (ref.)	Ref.		Ref.
ImpD	11.03 (4.91)	0.02	6.81 (4.47)	0.13
UPD	9.36 (3.73)	0.01	4.40 (2.80)	0.12
UBE3A pathogenic variant	9.06 (4.82)	0.06	6.42 (3.69)	0.08
BSID-III Cognitive AE	0.80 (0.15)	< 0.001	0.28 (0.19)	0.15
ABC Irritability	1.56 (0.11)	< 0.001	0.82 (0.13)	< 0.001
ABC Hyperactivity	0.91 (0.06)	< 0.001	0.54 (0.07)	< 0.001
VABS-II Composite	−0.36 (0.09)	< 0.001	−0.21 (0.10)	0.049
PSI Parent Domain (n = 228, obs = 486)
Male sex	7.55 (3.10)	0.01	8.33 (2.79)	0.003
Hispanic	−8.90 (4.28)	0.04	−5.72 (4.14)	0.17
Diagnosis
Class I	3.46 (3.76)	0.36	6.03 (3.73)	0.11
Class II (ref.)	Ref.		Ref.
ImpD	7.01 (5.24)	0.18	4.45 (5.11)	0.38
UPD	9.78 (4.67)	0.04	7.00 (4.20)	0.10
UBE3A pathogenic variant	23.16 (5.08)	< 0.001	18.22 (4.49)	< 0.001
ABC-C Irritability	0.97 (0.17)	< 0.001	0.71 (0.24)	0.004
ABC-C Hyperactivity	0.45 (0.11)	< 0.001	0.19 (0.14)	0.16
Total Stress (n = 222, obs = 462)
Age (at visit in years)	1.13 (0.55)	0.04	−1.81 (0.85)	0.03
Diagnosis
Class I	2.55 (5.83)	0.66	7.55 (5.56)	0.17
Class II (ref.)	Ref.		Ref.
ImpD	20.33 (8.40)	0.02	14.00 (9.26)	0.13
UPD	19.44 (7.39)	0.01	12.34 (6.80)	0.07
UBE3A pathogenic variant	31.22 (9.27)	< 0.001	26.14 (8.14)	0.001
BSID-III Cognitive AE	1.14 (0.39)	0.004	0.47 (0.57)	0.42
ABC-C Irritability	2.40 (0.27)	< 0.001	1.53 (0.36)	< 0.001
ABC-C Hyperactivity	1.25 (0.14)	< 0.001	0.74 (0.17)	< 0.001
VABS-II Composite	−0.53 (0.17)	0.002	−0.40 (0.23)	0.08

Note: Age is centred at 6 years.

Abbreviations: AE, age equivalent score; BSID-III, Bayley Scales of Infant and Toddler Development, Third Edition; ImpD, imprinting centre defect; obs, number of observations; PSI, Parenting Stress Index; UPD, uniparental disomy; VABS-II, Vineland Adaptive Behaviour Scales, Second Edition.

TABLE 6 ∣.

Generalised linear model for Family Quality of Life Scale predictors with log link.

	Bivariate Parameter		Multivariate Parameter
	Exp(β) (SE)	p	Exp(β) (SE)	p
Model 1: FQOL Family Interaction (n = 214, obs = 329)
Male sex (ref. is female)	0.95 (0.02)	0.043	0.97 (0.018)	0.11
Diagnosis
Class I deletion	1.02 (0.029)	0.60	1.02 (0.023)	0.36
Class II deletion (ref.)	Ref		Ref
ImpD	0.93 (0.043)	0.11	0.98 (0.039)	0.59
UPD	1.00 (0.031)	0.88	1.04 (0.032)	0.17
UBE3A pathogenic variant	0.86 (0.038)	< 0.001	0.92 (0.037)	0.04
Child-Domain Stress (in 10 points)	0.97 (0.005)	< 0.001	0.99 (0.007)	0.13
Parent-Domain Stress (in 10 points)	0.97 (0.003)	< 0.001	0.97 (0.004)	< 0.001
ABC-C Irritability	1.00 (0.002)	0.02	1.00 (0.002)	0.84
BSID-III Cognitive AE	1.00 (0.001)	0.01	1.00 (0.001)	0.38
Model 2: FQOL Parenting (n = 224, obs = 351)
Male sex (ref. is female)	0.941 (0.022)	0.009	0.960 (0.019)	0.04
Child-Domain Stress (in 10 points)	0.969 (0.004)	< 0.001	0.985 (0.005)	0.003
Parent-Domain Stress (in 10 points)	0.967 (0.003)	< 0.001	0.975 (0.004)	< 0.001
Model 3: FQOL Emotional Well-being (n = 209, obs = 336)
Male sex (ref. is female)	0.896 (0.037)	0.008	0.900 (0.032)	0.003
Marital status
Single, never married	1.02 (0.084)	0.002	1.134 (0.067)	0.03
Separated/divorced	1.18 (0.068)	0.005	1.136 (0.034)	0.02
Declined	0.91 (0.035)	0.02	0.996 (0.07)	0.95
Married/living in marriage-like relationship (ref.)	Ref		Ref
Diagnosis
Class I deletion	1.026 (0.050)		1.043 (0.044)	0.32
Class II deletion (ref.)	Ref		Ref
ImpD	0.870 (0.064)	0.058	0.932 (0.053)	0.22
UPD	0.861 (0.047)	0.005	0.907 (0.053)	0.09
UBE3A pathogenic variant	0.809 (0.694)	0.007	0.989 (0.009)	0.23
Child-Domain Stress (in 10 points)	0.967 (0.008)	< 0.001	0.989 (0.009)	0.23
Parent-Domain Stress (in 10 points)	0.954 (0.006)	< 0.001	0.966 (0.007)	< 0.001
Model 4: FQOL Physical/Material Well-being (n = 201, obs = 323)
Annual household income
Under $25 000	0.974 (0.045)	0.57	0.955 (0.039)	0.25
$25 000–$49	999 0.969 (0.040)	0.44	0.972 (0.037)	0.45
$50 000–$99 999 (ref)	Ref		Ref
$100 000 or more	1.077 (0.022)	< 0.001	1.060 (0.019)	0.001
Child-Domain Stress (in 10 points)	0.982 (0.004)	< 0.001	0.992 (0.005)	0.09
Parent-Domain Stress (in 10 points)	0.980 (0.003)	< 0.001	0.985 (0.004)	< 0.001
Model 5: FQOL Disability-related Support (n = 219, obs = 340)
Age (in years)	0.991 (0.003)	< 0.001	0.992 (0.004)	0.04
Male sex (ref. is female)	0.902 (0.024)	< 0.001	0.915 (0.021)	0.001
VABS-II Composite (in 5 points)	1.010 (0.005)	0.04	0.992 (0.006)	0.21
ABC-C Irritability	0.993 (0.002)	< 0.001	1.001 (0.002)	0.59
ABC-C Hyperactivity	0.996 (0.001)	< 0.001	0.998 (0.002)	0.13
Child-Domain Stress (in 10 points)	0.970 (0.005)	< 0.001	0.986 (0.007)	0.053
Parent-Domain Stress (in 10 points)	0.978 (0.004)	< 0.001	0.989 (0.005)	0.03

Note: Effect size in parameter estimate is based on a 5-point or 10-point increase in score. Exp(β) = exponentiated coefficient from a Gamma-generalised linear model (on log scale). Age is centred at 6 years.

Abbreviations: ABC-C, Aberrant Behaviour Checklist—Community; AE, age equivalent score; BSID-III, Bayley Scales of Infant and Toddler Development, Third Edition; FQOL, Family Quality of Life; ImpD, imprinting centre defect; obs, number of observations; UPD, uniparental disomy; VABS-II, Vineland Adaptive Behaviour Scales, Second Edition.

FIGURE 1 ∣.

Model-based Parenting Stress Index Child Domain least-square mean scores by molecular subtype. Note: Multivariate model controlled for age, BSID-III Cognitive AE, ABC-C Irritability, ABC-C Hyperactivity and VABS-II Composite with post hoc test of differences corrected for multiple comparisons. Bars indicate significantly different means across groups. ABC-C, Aberrant Behaviour Checklist—Community; AE, age equivalent; BSID-III, Bayley Scales of Infant and Toddler Development, Third Edition; ImpD, Imprinting Center Defect; LS Means, least-square mean scores; PSI, Parenting Stress Index; UPD, uniparental disomy; VABS-II, Vineland Adaptive Behaviour Scales, Third Edition.

FIGURE 4 ∣.

Adjusted least-square mean scores for Family Quality of Life Emotional Well-Being subscale by molecular subtype with 95% confidence limit with post hoc test of differences correction for multiple comparisons. Note: Least-square means and confidence intervals for adjusted model controlled for gender, marital status, PSI Child Domain stress, PSI Parent Domain stress and were exponentiated for ease of interpretation. Post hoc statistical tests were performed with Bonferroni adjustments. FQOL, Family Quality of Life; ImpD, imprinting centre defect; LS Means, least-square mean scores; UPD, uniparental disomy.

FIGURE 2 ∣.

Model-based Parenting Stress Index Parent Domain least square-mean scores by molecular subtype. Note: Multivariate model controlling for gender, Hispanic ethnicity, ABC-C Irritability and ABC-C Hyperactivity with post hoc test of differences corrected for multiple comparisons. Bars indicate significantly different means across groups. ImpD, imprinting centre defect; LS Means, least-square mean scores; PSI, Parenting Stress Index; UPD, uniparental disomy.

FIGURE 3 ∣.

Model-based Parenting Stress Index Total Stress least-square mean scores by molecular subtype. Note: Multivariate model controlled for age, gender, BSID-III Cognitive AE, ABC-C Irritability, ABC-C Hyperactivity and VABS-II Composite with post hoc test of differences correction for multiple comparisons. Bars indicate significantly different means across groups. ABC-C, Aberrant Behaviour Checklist—Community; AE, age equivalent; BSID-III: Bayley Scales of Infant and Toddler Development, Third Edition; ImpD, imprinting centre defect; LS Means, least-square mean scores; PSI, Parenting Stress Index; UPD, uniparental disomy; VABS-II, Vineland Adaptive Behaviour Scales, Third Edition.

Changes in parental stress and FQOL over time were explored by analysing the impact of child age on PSI and FQOL scores. Child age was a significant predictor of Total Stress scores. Increasing age was associated with increased Total Stress in the bivariate model (p = 0.04) but was associated with decreased Total Stress in the multivariate model after controlling for sex, subtype, cognitive ability, adaptive functioning and behaviour (Table 5; p = 0.03). FQOL scores were not significantly affected by child age except in the Disability-Related Support subscale, where satisfaction decreased with increasing child age (Table 6; p = 0.04).

3.3 ∣. Predictors of PSI and FQOL Scores

ABC-C Irritability score was a significant predictor of Child Domain (p < 0.001), Parent Domain (p = 0.004) and Total (p < 0.001) stress, whereas ABC-C Hyperactivity score predicted an increase in Child Domain and Total stress (Table 5; p < 0.001). Higher BSID-III Cognitive AE predicted higher levels of Child Domain and Total stress in the bivariate model, but this effect was not significant when controlling for age, sex, behaviour and adaptive functioning. Higher VABS-II Composite scores predicted lower Child Domain stress in the multivariate model (p = 0.049). Hispanic ethnicity predicted lower Parent Domain stress in the bivariate model, but the interaction was no longer significant in the multivariate model. Other covariates tested were not significant.

Participant sex was a significant predictor of FQOL Parenting, Emotional Well-Being and Disability-Related Support subscale scores (Table 6). Parents of male individuals scored lower than parents of female individuals in these subscales in the multivariate model. ABC-C Irritability and Hyperactivity scores were not significant predictors of FQOL in the multivariable model. Higher VABS-II Composite scores were associated with increased satisfaction in the Disability-Related Support subscale, but this effect was not significant in the multivariate model. Single and separated/divorced marital status was associated with higher Emotional Well-being subscale scores (p = 0.03 and p = 0.02, respectively). Higher family income was associated with increased Physical/Material Well-Being, with families earning $100 000 or more scoring higher in Physical/Material Well-Being than families earning $50 000–$99 000. Post hoc analysis showed that families with annual incomes of $100 000 or more had significantly higher Physical/Material Well-Being scores than those with incomes below $25 000 (p_adj = 0.041) and $50 000–$99 999 (p_adj = 0.008) (Figure 5).

FIGURE 5 ∣.

Adjusted mean scores for FQOL Physical/Mental Well-being by annual household income with 95% confidence limit with post hoc test of differences correction for multiple comparisons. Note: Least-square means and confidence intervals were adjusted for PSI Child Domain stress and Parent Domain stress as well as for repeated measures and exponentiated for ease of interpretation. Post hoc statistical tests were performed with Bonferroni adjustments. FQOL, Family Quality of Life; LS Means, least-square mean scores; PSI, Parenting Stress Index.

3.4 ∣. Impact of Parental Stress on FQOL

PSI Parent Domain, Child Domain and Total stress scores were analysed as predictors of FQOL (Table 6). After controlling for significant covariates, Child Domain stress was associated with a significant decrease in Parenting subscale FQOL satisfaction (p = 0.003). Parent Domain stress was associated with a significant decrease in FQOL satisfaction across all FQOL subscales.

Table 7 outlines associations of the PSI Child and Parent subdomains with FQOL subscales, adjusted for covariates (Table S1). Correlations between PSI and FQOL scores at the baseline visit were first calculated to examine relationships between PSI Domains and FQOL subscales (Table S2). Parent Domain stress was negatively and significantly correlated with FQOL across subtypes, whereas Child Domain stress was negatively and significantly correlated with FQOL in the deletion groups. Only one significant correlation existed between Child Domain stress and Disability-Related Support satisfaction in the non-deletion groups (Table S2).

TABLE 7 ∣.

Model-based associations between Parenting Stress Index domains and Family Quality of Life subscales (n = 159).

	FQOL subscalesa
	Family Interaction		Parenting		Emotional Well-Being		Physical/ Material Well-Being		Disability-Related Support
	Exp(β)	p	Exp(β)	p	Exp(β)	p	Exp(β)	p	Exp(β)	p
PSI Child Domain
Distractibility/hyperactivity	1.003 (0.002)	0.06	1.001 (0.002)	0.55	1.000 (0.003)	0.94	0.999 (0.002)	0.49	0.999 (0.003)	0.88
Adaptability	1.001 (0.002)	0.60	0.999 (0.002)	0.545	0.999 (0.004)	0.76	0.998 (0.002)	0.18	0.997 (0.003)	0.59
Reinforces parent	0.994 (0.005)	0.18	0.997 (0.006)	0.549	1.001 (0.009)	0.96	0.999 (0.005)	0.84	0.999 (0.006)	0.54
Demandingness	0.994 (0.002)	0.02	0.997 (0.002)	0.204	1.001 (0.004)	0.91	1.003 (0.002)	0.23	0.995 (0.002)	0.02
Mood	0.997 (0.005)	0.50	0.998 (0.004)	0.673	0.998 (0.007)	0.81	0.995 (0.004)	0.18	1.000 (0.005)	0.52
Acceptability	0.997 (0.003)	0.37	0.994 (0.003)	0.082	0.983 (0.006)	0.006	0.996 (0.003)	0.20	0.990 (0.005)	0.01
PSI Parent Domain
Competence	1.001 (0.002)	0.82	0.995 (0.003)	0.048	1.005 (0.005)	0.30	0.997 (0.002)	0.09	0.998 (0.003)	0.47
Isolation	0.997 (0.003)	0.32	0.996 (0.004)	0.23	0.992 (0.006)	0.20	0.998 (0.003)	0.34	0.995 (0.003)	0.06
Attachment	0.997 (0.004)	0.50	1.005 (0.005)	0.25	1.000 (0.008)	0.98	1.003 (0.003)	0.40	0.997 (0.005)	0.75
Health	0.997 (0.003)	0.32	0.997 (0.006)	0.62	0.985 (0.006)	0.007	1.001 (0.003)	0.81	1.006 (0.004)	0.13
Role restriction	1.002 (0.003)	0.44	1.002 (0.003)	0.46	1.004 (0.005)	0.36	1.000 (0.002)	0.99	1.007 (0.003)	0.05
Depression	0.998 (0.002)	0.39	0.996 (0.002)	0.09	1.000 (0.005)	0.94	0.995 (0.003)	0.054	0.998 (0.003)	0.25
Spouse/parenting partner relationship	0.991 (0.003)	< 0.001	0.997 (0.003)	0.20	0.999 (0.004)	0.71	0.999 (0.002)	0.52	1.003 (0.003)	0.28

Note: Age is centred at 6 years.

Abbreviations: FQOL, Family Quality of Life; PSI, Parenting Stress Index.

^aThe estimates in the column are adjusted for molecular subtype, age, marital status, ethnicity, seizure, seizure by molecular subtype interaction, and Vineland Adaptive Behaviour Scales-II composite score.

After controlling for covariates, only a few associations between PSI subdomain scores and FQOL subscale scores remained statistically significant (Table 7). Higher Child Demandingness (PSI Child Domain) and Spouse/Parenting Partner Relationship (PSI Parent Domain) scores were associated with decreased Family Interaction FQOL scores; higher Competence (PSI Parent Domain) scores, indicating lower feelings of parental competence, had a negative association with Parenting FQOL scores; increased Acceptability (PSI Child Domain) scores, indicating differences between parent expectations of child abilities and their perceived actual abilities and increased Health (PSI Parent Domain) scores, indicating stress related to parent health, had an inverse relationship with Emotional Well-Being FQOL scores; and increased Demandingness and Acceptability (PSI Child Domain) scores had a negative association with Disability-Related Support FQOL scores (Table 7).

4 ∣. Discussion

We aimed to describe parental stress and FQOL across AS molecular subtypes and the factors that contributed to each. Nineteen to forty-eight percent of participants reported clinically significant parental stress, depending on molecular subtype, indicating the need for interventions to reduce stress. Parental stress was lower in parents of individuals with deletion subtypes compared to those with non-deletion subtypes, as previously demonstrated using a partially overlapping dataset (Miodrag and Peters 2015). FQOL satisfaction reported here is similar to previously reported levels in families with children with IDs (Hoffman et al. 2006; Boehm et al. 2015; Staunton et al. 2020a). Families experience lowest satisfaction with their support systems and ability to pursue independent interests. Parents of children with IDs often face multiple demands in their caregiving role (managing financial stress, coordinating care, emotional stress and social isolation) (Thompson et al. 2014). These factors likely contribute to lower satisfaction in support systems and time for self in AS families. Resources that foster social support and activities outside the parenting role may improve emotional well-being (Miranda et al. 2019).

4.1 ∣. Parental Stress in Parents of Individuals With UBE3A Pathogenic Variants

This is the first study to our knowledge to describe parental stress in parents of individuals with UBE3A pathogenic variants. We found this group experienced the highest rates of clinically significant stress, driven disproportionately by parent factors compared to other subtypes (Abidin 1995). Individuals with UBE3A pathogenic variants tend to exhibit higher levels of maladaptive behaviours, which may contribute to increased parental stress in this group (Sadhwani et al. 2019). Additionally, these individuals tend to have higher cognitive functioning (Sadhwani et al. 2023), which may contribute to increased parental stress due to more complex social–emotional needs and unique expectations regarding development and independence. However, our results show parents of children with UBE3A pathogenic variants have greater parental stress even when controlling for these factors, suggesting there are other unique aspects of parenting a child with a UBE3A pathogenic variant that contribute to parental stress. For example, anxiety may disproportionately impact individuals with AS due to a pathogenic UBE3A variant and can be a source of stress for parents (Hagenaar et al. 2024; Gentile et al. 2010; Grebe et al. 2022; Wheeler et al. 2019). Anxiety manifests as behaviour problems and separation distress and may be paradoxically associated with higher developmental functioning, which can lead to frustration and worsened anxiety (Grebe et al. 2022; Wheeler et al. 2019; Keary et al. 2022). Our findings highlight the need for further research on unique stressors faced by parents of individuals with UBE3A pathogenic variants.

4.2 ∣. Impact of Age

As participants aged, we saw a small decrease in parental stress but also in Disability-Related Support satisfaction. Older age is associated with lower health-related QoL, possibly due to increased health risks, anxiety and divergence from the developmental level of peers amplifying individuals’ perception of deficits (Hagenaar et al. 2024; Xia et al. 2023). These effects may explain the decreased FQOL associated with increased age observed in our study. Reduction in parental stress over time may be due to families developing support systems and coping strategies that improve parent well-being and child–parent relationships over time (Gerstein et al. 2009). We assessed parenting stress in families with children up to 22 years of age; however, caring for an adult with AS comes with unique challenges for parents and families (Thomson et al. 2017). Future research should examine parenting stress and FQOL in families caring for adults with AS to more fully understand the impact of AS on families over time.

4.3 ∣. Child and Family Characteristics That Impact Stress and FQOL

In terms of predictors, maladaptive behaviours correlated with increased parental stress but did not strongly impact FQOL, consistent with previous studies (Sadhwani et al. 2019; Hagenaar et al. 2024; Tomanik et al. 2004). Irritability may strain the parent–child relationship and increase parental stress (Abidin 1995). Irritability also contributed to parent-related stressors such as perceptions of competence, attachment, well-being and relationships. The relationship between parental stress and maladaptive behaviours may be bidirectional, with increased stress leading to erosion of the parent–child relationship and contributing to behavioural problems (Baker et al. 2003; Neece et al. 2012). Additional studies may seek to examine how behaviour affects stress by analysing the effects of behaviour on specific PSI subdomains. In addition, families with male children had higher stress and lower FQOL than families with female children, possibly due to increased rates of aggressive behaviours (e.g., hairpulling, biting) which occur more frequently in male individuals with AS (Sadhwani et al. 2019; Barroso et al. 2018). Families with higher incomes were more satisfied with their material and physical well-being, reflecting the importance of income to meet basic needs such as transportation, healthcare and feeling safe (Walkowiak and Domaradzki 2025a; Hoffman et al. 2006).

Medical complexity of the individual with AS, indicated by seizure severity and hospitalisations per year, as well as sleep difficulties, did not affect parental stress or FQOL. Individuals with AS have high health burden, which negatively impacts individual QoL (Khan et al. 2019a, 2019b, 2023; Xia et al. 2023). Epilepsy severity and hospitalisation can cause increased parental stress (Zdun-Ryżewska et al. 2021; Operto et al. 2019), but our results suggest that these are not sources of long-term stress in AS. Sleep difficulties in the child did not affect stress, inconsistent with previous research (Hagenaar et al. 2024; Goldman et al. 2012). However, we did not utilise a standardised sleep questionnaire or objective sleep assessment, warranting future research on this association. Cognitive functioning did not affect parental stress or FQOL. Improved adaptive functioning had a minimal positive effect on Child Domain stress.

Hagenaar et al. (2024) reported child characteristics associated with parental stress in 73 children with AS recruited from an outpatient clinic in the Netherlands. The consistency between our findings and those of Hagenaar et al. is notable, suggesting that child factors such as sleep disturbances, seizure severity and cognitive function may not significantly influence parenting stress across different cultures. However, unlike our findings, Hagenaar et al. did not identify molecular subtype as a significant predictor of parental stress. This discrepancy may be due to their smaller sample size or their grouping of all non-deletion subtypes in their analysis. Additionally, the average participant age was lower in our study, raising the possibility that differences in stress across subtypes diminish over time as parenting stress decreases.

4.4 ∣. Relationship Between Parental Stress and FQOL

Parental stress and FQOL are closely related, and this was reflected in our results (Jenaro et al. 2020; Hsiao et al. 2017). Higher levels of parental stress predicted a decrease in FQOL. These effects were not previously studied in AS, but prior research in other ID populations has shown similar associations (Jenaro et al. 2020; Staunton et al. 2020b). Stress–FQOL associations were driven by parent-related factors, such as health concerns and feelings of incompetence, as well as child-related factors, such as child demandingness. Lower satisfaction with outside support and with the parenting role can be caused by, and contribute to, lower feelings of parental competence (Jandrić and Kurtović 2021). Parents who experience high demands for attention, high defiance and aggression and discrepancies in expectations versus realities of child behaviours and abilities may be less satisfied with their family cohesion (Abidin 1995).

4.5 ∣. Intervention Strategies for Reducing Parental Stress and Enhancing FQOL

Interventions that reduce parental stress may directly improve parent well-being and support family interactions, as well as indirectly support effective parenting strategies and foster positive outcomes for children with IDs (Miranda et al. 2019; Baker et al. 2003; Hastings and Beck 2004). Research suggests that referral to professional counselling would be recommended for the 19%–48% of families in our sample who reached the clinical stress cut-off (Abidin 1995), but actual rates of counselling in our study population are not known. Other interventions such as respite care, case management involvement, group cognitive behavioural therapy and parent-led support groups may be effective in lowering parental stress in AS families (Miranda et al. 2019; Hastings and Beck 2004). Treatments that target challenging child behaviours may also reduce parental stress (Sadhwani et al. 2019; Goldman et al. 2012). We identified molecular subtype, child sex and family income as predictors of parental stress and FQOL. Although these are not clinically modifiable, clinicians should be mindful of these factors to identify families who may be at increased risk for stress and family dysfunction. Clinicians who serve individuals with AS and their families are in an ideal position to understand the demands placed on parents, identify families at increased risk for elevated stress and help initiate interventions to reduce parental stress and maximise family well-being.

4.6 ∣. Limitations

There are several limitations to this study. Firstly, our sample lacked racial and socioeconomic diversity, limiting the generalisability of our results. Only 7% of participants reported income below $25 000, and only 9% reported non-White race. Additionally, our participants had access to resources that allowed them to participate in a time-intensive research study, which may not reflect access in the general population. The interaction between social factors and parenting stress is complex—families of children with and without IDs from groups who are marginalised, including non-White, Hispanic and low-income families, may be variably vulnerable to parental stress and its effects on FQOL, depending on structural and cultural factors (Cardoso et al. 2010; Nomaguchi and House 2013; Iadarola et al. 2019). There are no studies to our knowledge that focus on parental stress and FQOL in underrepresented families of individuals living with AS, and these questions should be explored further with a larger and more diverse dataset to support social and cultural competence in assessing and responding to these outcomes.

Assessment of parental stress and FQOL had limitations. Elevated stress may not indicate elevated distress. Child Domain stress was less strongly associated with FQOL than Parent Domain stress despite higher Child Domain compared to Parent Domain stress levels. This suggests that child characteristics evaluated on the PSI may not represent stressful characteristics in these families due to resilience and positive coping (Gerstein et al. 2009; Peer and Hillman 2014). The PSI does not capture the positive aspects of the parent–child relationship; however, the FQOL Scale aims to capture positive aspects of raising a child with ID (Summers et al. 2005). Having a child with AS has diverse impacts on families, and identifying positive aspects of parent–child relationships may elucidate how families adapt to challenges. We did not collect information about stress-reduction strategies or interventions pursued by parents over time and recommend further research to better understand parent coping strategies to elucidate mechanisms behind their resilience. The Beach FQOL Scale was designed to assess outcomes of services for people with IDs and may not encompass all aspects of FQOL that families find important.

Only one primary caregiver completed the PSI and FQOL, limiting the perspectives of parental stress and FQOL to one parenting partner. Mothers and fathers respond to parenting a child with health problems differently (Walkowiak and Domaradzki 2025a; Pelchat et al. 2007; Potter et al. 2022), but we lacked sufficient data to analyse stress in mothers and fathers independently. Our understanding of FQOL is incomplete without including siblings of individuals living with AS, as siblings are uniquely affected by their family members' disability and serve a crucial role in caregiving and support during and after childhood (Wheeler et al. 2017; Turnwald et al. 2022). Future studies should collect PSI data from all parent partners, and FQOL data from multiple members of the immediate family including siblings, to learn how different family members perceive stress and family well-being and facilitate the delivery of family-centred care and support.

Although our study controlled for maladaptive behaviour, cognition and adaptive functioning, it is possible the outcome measures utilised in our research do not capture the spectrum of functioning that is relevant to individuals with AS and their families (Sadhwani et al. 2023; Sadhwani et al. 2019; Keute et al. 2020).

Finally, we collected data at two to three time points for each participant on average, limiting our assessment of how parental stress or FQOL changes over the course of time. Longitudinal studies of parental stress and FQOL over many years may be pursued to further this understanding.

4.7 ∣. Conclusions

Parental stress is elevated but decreases over time and is highest in families of individuals living with UBE3A pathogenic variants. Maladaptive behaviours and child male sex are associated with higher parental stress. FQOL satisfaction is generally similar across molecular subtypes and remains stable over time. Stress has a negative impact on FQOL. Interventions that reduce parental stress may improve parent well-being and support family interactions.

Supplementary Material

Additional supporting information can be found online in the Supporting Information section. Table S1: Test of fixed effects of model controls (n = 159). Table S2: Correlations of Family Quality of Life Subscales with Child Domain, Parent Domain, and Total Parenting Stress Index Scores by molecular subtype at baseline visit (n = 231).

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.