Validation of a Scale to Assess Activities of Daily Living at Home in Children and Adolescents With Autism Spectrum Disorder

Abstract

skills relative to their cognitive skills (Bal et al. 2015;Farley et al. 2009;Marsack-Topolewski et al. 2021).Therefore the assessment of functional impairment, including different areas of a person's life, such as education, family, social life, work life, leisure and free time are crucial, both for the diagnosis, as well as for the therapeutic approach of young people with ASD (Rotger et al. 2014).
In general, levels of functioning, measures of adaptive behavior, and in particular daily living skills, are the variables that most affect repetitive patterns of behavior in people with ASD (Farley et al. 2009;Vahia, 2013).In fact, most people with ASD remain at home with their families until they enter adulthood, so parents provide continuous support to their children to overcome daily living difficulties (ADLs) (Dudley et al. 2019;Marsack-Topolewski et al. 2021).
In addition, in recent months, following the COVID 19 pandemic, children, adolescents and adults with ASD have been locked in their homes for more than a year, so ADLs at home could undergo substantial changes.
Therefore, to assess these daily activities at home it is necessary to have a simple, brief and easy to apply instrument.Therefore, this study aimed to validate a scale to assess activities of daily living (ADLs) within the home of children and adolescents with ASD.

Type and sample
A cross-sectional study was designed in Chilean children and adolescents with ASD.The sample consisted of 32 boys and 5 girls from public educational institutions in the cities of Talca, Rancagua and Santiago (Chile).The age range ranged from 4 to 20 years old.
The sample selection was non-probabilistic by convenience.All the information about the children and adolescents with ASD was collected through their parents, since they were the ones who answered the survey.The average age of the parents was 36.85±7.61years.The sociodemographic variables of both the young people with ASD and their parents are shown in Table 1.

Procedures
Parents were contacted via telephone in May 2021.Once contacted, the objective of the research was explained to them.Parents who agreed to participate in the study signed the informed consent.This stage lasted two weeks.
Children enrolled in state special education schools (preschool, primary and secondary) and those living directly with one or both parents were included.
Children whose parents did not complete the applied scale and those living with relatives (other than their parents, one or both parents) were excluded.The entire procedure was performed in accordance with the Local Ethics Committee and the Helsinki declaration for human beings.

Validity and reliability of the scale
The SAADL was validated by expert judgment (content validity) and construct analysis, while reliability was assessed by measures of internal consistency and stability (retest).
Content validity was assessed by expert judgment, according to the suggestions described by Wiersma (2001).Six health and education professionals with a minimum of 10 years of professional and research experience in their area were invited to participate as experts.This panel of professionals with extensive experience in ASD reviewed appropriate scale items as described by Schultz (2005) The SAADL was sent to each of the experts by e-mail.They then evaluated the degree of representativeness, relevance, diversity, clarity, simplicity and completeness of each of the items of the elaborated instrument in an index card.The alternatives presented a scoring scale from 1 to 5 points.Each expert evaluated the SAADL and resubmitted the form for further analysis.In the end, the SAADL consisted of 8 questions and can be seen in appendix 1.
Construct validation was carried out by measuring exploratory factor analysis (EFA), allowing the underlying structure of the data to be determined (Bollen, 1989).
Reliability was assessed by internal consistency, calculating Cronbach's alpha per question and total scale.On the other hand, stability measures (retest) were also used with a 14-day time interval between both measurements.For the second measurement, nine parents were used, representing 24% of the total sample.

Statistics
The normality of the data was verified by the Shapiro-Wilk test.Descriptive statistics were analyzed for frequencies, percentages, range, mean (X), standard deviation (SD), skewness, and kurtosis.
For content validity, Aiken's V (Bulger & Housner, 2007) was used which evaluates the adequacy of the items to the content validity criteria.Higher values of Aiken's V ≥ 0.75 were accepted (Bulger & Housner, 2007).
For construct validity, the exploratory factor analysis (EFA) fit model was considered by adopting the Kaiser Meyer-Olkin (KMO) sample adequacy criteria, Bartlett's sphericity value was considered to establish the relevance of the factor analysis, Comparative Fit Index (CFI) root mean square error of approximation (RMSEA).The analysis provided the measure of variance explained, factor loadings, communalities and Chi-square approximation.To assess stability measures, the concordance correlation coefficient (CCC) was calculated, using precision (p) and accuracy (A) according to Lawrence and Lin (1989) approach.In addition, the weighted kappa (Cohen, 1968) was calculated to measure the magnitude of agreement between the two scores (test and retest).In all cases, p < 0.05 was adopted.The results were processed and analyzed initially in Excel spreadsheets and subsequently in SPSS 18.0 and Med Calc 11.1.0,as appropriate.

Results
The sociodemographic variables of the parents and children are shown in Table 1.Thirty-seven parents were surveyed, corresponding to 51.4% cohabiting families, 40.5% married and 8.1% divorced.The majority of respondents were female (81.1%) and only 18.9% were male.In addition, most of these families lived in urban areas (86.5%) compared to those in rural areas (13.5%).Regarding information on children with ASD, the majority were males (86.5%) relative to females (13.5%).Of the children, 67.7% lived with both parents (67.7%) and 32.4% with their mothers, while 18.9% were only children and 81.1% had one or more siblings.Values for the V of Aiken test are presented in Table 2.
The values for each question varied from 0.75 to 0.92 while the values for the dimensions were between 0.82 and 0.88.For all of the cases, the values obtained from the judges reflected an agreement of 0.75 to 0.88.The descriptive values of the SAADL are shown in Table 3. Skewness showed values lower than the average (-0.97 to 1.39) while kurtosis ranged from (-1.88 to 2.68).
The coefficient of variation in all cases was less than 33% and the values of Cronbach's alpha per question ranged from 0.81 to 0.90 and in the total scale it was r = 0.87.
The reliability values analyzed by means of retesting can be seen in Table 5.The DRI was calculated, obtaining CCC values from 0.87 to 1.0, the values of precision (0.96) and accuracy (0.99) were very high, and even the weighted Kappa showed high concordance values (0.74 to 1.0), which guarantee equivalence between both measurements.

Discussion
The results of the study have shown that the scale proposed in this study proved to be valid and reliable for assessing ADLs in the homes of children and adolescents with ASD.(Charter, 2003) and even, the interjudge reliability standard is higher than 0.75, which are considered excellent (Cicchetti, 1994).
Secondly, the validation by AFE revealed two dimensions, personal care and mobility at home.In addition, the model was stable and met the criteria for goodness-of-fit indices in CFI, RMSEA, KMO, variance explanation as described in the literature (Dini et al., 2014;Hu & Bentler, 1999;Schermelleh-Engel, Moosbrugger, & Müller, 2003).
On the other hand, the factor loadings obtained in this study were higher than 0.62, while the communalities >0.53, reflecting acceptable values between item and dimension (Fabrigar et al. 1999;Knekta et al. 2019;MacCallum et al. 2001).These values are similar to those reported in recent studies in youth with ASD (Cassidy et al. 2021;Zhou et al. 2017).
Regarding reliability, data were analyzed by internal consistency and retest.For the first case, Cronbach's alpha was used evidencing high reliability values r = 0.87, being similar to other studies conducted in ASD populations (Breidbord & Croudace, 2013;Brugha et al. 2020;Skuse et al. 2005).In general, the scale proposed here reflects internal consistency among its items according in line with what is suggested in the literature (Nunnally, 1994;Streiner, 2003), highlighting a minimum of 0.80.
In the second case, the retest was used as a measure of stability.The scale was applied to the parents on two occasions with an interval of 14 days, in which it is proposed to consider between 10 and 14 days (Terwee et al. 2007).This time interval has allowed reporting concordance between both measures and high levels of precision and accuracy.This shows that the scale presents stability in the scores of both tests, so they remain without substantial changes when measured on different occasions (Michalos, 2014).In addition, the values obtained in this study are consistent with research that has evaluated reliability by retesting (Berthoz & Hill, 2005;Dutil et al. 2017;Pereira et al. 2008).
Re-testing (test and re-test) as quality control criteria for scales in general are crucial, especially if the scales are intended to be applied in treatment interventions.This is because instruments are required to be stable in their results, especially in scales that have to do with ADLs (Dutil et al. 2017).
In general, this study presented some limitations that deserve to be clarified.A relatively small sample was used, and the scale was applied in COVID-19 pandemic time, through google drive.Probably, these factors could have affect the results obtained in the study.However, despite this, the validity and reliability techniques have shown consistent results.It is even one of the first studies that seeks to propose a new scale to be used in the ASD populations that can serve as a baseline for comparing post-pandemic ADLs.It is suggested that future studies expand the study sample, apply other validity and reliability techniques to achieve external validity of the scale.

Conclusion
This study concludes that the SAADL for children and adolescents with ASD is valid and reliable.These psychometric properties warrant its use and regular application to assess daily activities of self-care and mobility in the home in youth with ASD.For the ability to perform ADLs at home can serve to improve safety conditions, greater participation in the home, alleviate family or caregiver overload, and improve their quality of life.

Table 1 .
Sociodemographic characteristics of the sample studied.

Table 2 .
Content validity of the instrument (SAADL) using Aiken's V by question and dimension.

Table 3 .
Descriptive analysis of SAADL in children and adolescents with ASD.
Legend: SD: Standard deviation, CV: Coefficient of variation

Table 4 .
Factor loadings and communalities based on exploratory factor analysis.

Table 5 .
Values that define the concordance between the values of the test and retest of the SAADL.