a APGAR: appearance, pulse, grimace, activity, and respiration.
b ADHD: attention-deficit/hyperactivity disorder.
c ASD: autism spectrum disorder.
Clinicians’ diagnostic conclusion regarding the presence or absence of ADHD or ASD was considered as a dichotomous outcome in separate classification models, that are (1) the “ADHD” group comprised participants with a clinical diagnosis of ADHD and possible comorbid conditions; the “non-ADHD” group comprised participants without a clinical diagnosis of ADHD, that is, subjects who received other neuropsychiatric diagnoses or no categorical diagnosis, and (2) the “ASD” group comprised participants with a clinical diagnosis of ASD and possible comorbid conditions; the “non-ASD” group comprised participants without a clinical diagnosis of ASD, that is, subjects who received other neuropsychiatric diagnoses or no categorical diagnosis.
Preliminary data handling.
Data handling and statistical analyses were done through R software (version 4.1.2; R Core Team) [ 21 ]. Missing data were imputed using the 10 nearest neighbors averaging [ 22 ].
Separate classification models were obtained, addressing the clinical and research questions, that are (1) “should a new help-seeking child be diagnosed with ADHD, considering the parent-reported socio-anamnestic information?” and (2) “should a new help-seeking child be diagnosed with ASD, considering the parent-reported socio-anamnestic information?”
RF models were performed through the “randomForest” toolbox [ 23 ], as previously done [ 3 ]. RF is an ensemble learning technique that generates many DTs and aggregates the results. To prevent overfitting, 2 layers of randomness are added in the procedure through bagging: (1) a bootstrap sample of the data set is considered in each tree (the data that are not considered in the bootstrap sample are called out of bag [OOB]); (2) a subset of mtry-independent variables are selected at each tree node. New data categories are predicted by aggregating all predictions performed by the trees, that is, choosing the majority of the voted categories [ 23 ]. In the tuning phase of the model selection, a leave-one-out cross-validation (LOOCV) approach was applied [ 24 ]. Finally, a SHAP analysis was performed to gain insights into the interpretability of the model [ 25 ]. SHAP values are computed by comparing the model’s predictions with and without a particular feature, and this process is repeated iteratively for each feature and sample in the data set. The magnitude of these values reflects the strength of the effect [ 25 ].
After conducting RF analyses, DT models were computed. The DT, characterized by a flowchart-like structure, is constructed by considering the entire data set positioned at the top of a “root” node. At each decision point, observations meeting the specified splitting condition are allocated to the left branch, while those not meeting the condition are directed to the right branch [ 26 ]. Information gain is a node impurity measure for selecting attributes and dividing each node, continuing until the terminal node, referred to as the “leaf,” is reached [ 26 ]. Finally, the algorithm assigns the most frequently observed class in each leaf as the classification prediction [ 26 ].
LR models were used in addition to the DT and RF models. LR is a traditional statistical method widely used for binary classification tasks. It models the probability of a binary outcome (presence or absence of the considered diagnosis) based on one or more predictor variables. In our study, LR was applied using the “glm” function in R.
Fixed training and test set.
To test the classification accuracy of the previously described models, we used 70% of the whole data set as a training set and the remaining 30% as a test set—the 2 subsamples did not present overlapping subjects. The classification performances of the selected models were evaluated considering the following information on the test set:
An additional cross-validation step was performed to test the results’ robustness. The whole data set was randomly split into 5 folds, and the 3 classification models were performed on each independent fold. The classification performances were calculated on each test set, and the mean performance values were estimated.
The study was approved by the Institute’s Ethical Review Board (protocol number 7/23, “Comitato Etico IRCCS E. Medea—Sezione Scientifica Associazione La Nostra Famiglia”). The research was conducted following guidelines and regulations depicted in the Declaration of Helsinki. The study data are deidentified, and no identification of individual participants in any images of the paper is possible. All the participant’s parents or legal guardians were informed of the aim of the study. Each subject was free to participate voluntarily and gave their written informed consent to the minor’s participation. No monetary compensation was provided for participating in the study.
The maximum percentage of missing data per subject was 36% (3 subjects). Table 2 depicts the sample’s demographic characteristics, considering the total sample and stratification by ADHD and ASD diagnosis.
Variable | Total sample (N=1688) | ADHD stratification | ASD stratification | ||||
ADHD (n=269) | Non-ADHD (n=1419) | ASD (n=270) | Non-ASD (n=1418) | ||||
Age (years), median (SD) | 8 (3) | 9 (3) | 8 (3) | 6 (4) | 9 (3) | ||
Male | 1097 (65) | 215 (80) | 894 (63) | 227 (84) | 879 (62) | ||
Female | 591 (35) | 54 (20) | 525 (37) | 43 (16) | 539 (38) |
a ADHD: attention-deficit/hyperactivity disorder.
b ASD: autism spectrum disorders.
Table 3 shows the RF classification models’ performances. Figure 3 shows the SHAP values (ie, the most important independent variables identified by the RF in accurately classifying the diagnoses).
Classification model | Performance on the fixed training and test set | SHAP values, mean (SD) | Average performance on the 5-fold cross-validation sets (SD) |
ADHD vs non-ADHD | : 50% <.001 | ||
ASD vs non-ASD | <.001 |
a SHAP: Shapley additive explanations.
c NIR: no information rate.
d ASD: autism spectrum disorders.
Table 4 shows the DT model results and performances on the test sets.
Classification model | Performance on the fixed training and test set | Attribute importance to the training set, mean (SD) | Average performance on the 5-fold cross-validation sets (SD) |
ADHD vs non-ADHD | : 50% <.001 | ||
ASD vs non-ASD | <.001 |
b NIR: no information rate.
c ASD: autism spectrum disorders.
Table 5 shows the LR model results and performances.
Classification model | Performance on the fixed training and test set | OR coefficients in the training set ( ) | Average performance on the 5-fold cross-validation sets (SD) |
ADHD vs non-ADHD | : 50% <.001 | <.001) <.001) =.003) <.001) =.016) =.676) | |
ASD vs non-ASD | <.001 | <.001) <.001) <.001) <.001) =.134) <.001) |
a OR: odds ratio.
The primary objective of our study was to develop accurate classification models for the diagnosis of ADHD and ASD within a sample referred for clinical evaluation. To this end, we used an ML approach to analyze internet-based parent-reported socio-anamnestic questionnaires.
Our ML models reached overall reasonable classifications in the test sets for both ADHD and ASD. The RF models exhibited classification accuracies of 84% for ADHD and 86% for ASD, respectively, with high sensitivities (93% for ADHD and 95% for ASD). On the other hand, the DT and LR models reached lower accuracy rates, with 74% and 61% accuracy for ADHD and 79% and 63% for ASD, respectively. The DT and LR models also demonstrated lower sensitivities (82% and 62% for ADHD and 88% and 68% for ASD).
In the 5-fold experiment, all models showed a decline in predictive accuracy, as could be expected due to smaller sample sizes. Nevertheless, the RF model continued to exhibit greater accuracy than other models. Concerning the different levels of accuracy reached by our 3 ML models, it is crucial to acknowledge both the advantages and disadvantages of RF, DT, and LR. One of the distinctive features of RF models is that they can effectively capture complex relationships within the data that may elude human interpretation [ 17 ]. For this reason, RF models can occasionally be considered difficult to interpret. This characteristic needs adequate consideration in the clinical context because the primary aim is to provide clinicians with an accurate “first glance” tool that supports them in forming initial diagnostic impressions.
Notwithstanding their eventual interpretability, RF models are remarkably effective in distinguishing different classes, thus representing an asset in psychopathology diagnosis. Conversely, as mentioned above, the DT and LR models are also readily interpretable for clinicians less familiar with ML techniques [ 17 ]. Therefore, the choice of approach depends on the decisional context and the desired degree of interpretability. In this study, we preferred greater levels of classification accuracy over the readiness of the classification process. However, a noteworthy option to mitigate the interpretability concern associated with RF models is provided by SHAP analysis. By assigning an important value to each feature in the classification model, SHAP analysis directly compares RF and other models regarding their interpretability.
Although slightly different in the achieved performance, the 3 models identified sex as the strongest predictor for both ADHD (all 3 models) and ASD (DT and LR models). It is well documented that males are more likely to be diagnosed with both ADHD [ 27 ] and ASD [ 28 ] than females. Interestingly, SHAP analysis indicated a relatively consistent ranking of features for RF models across the 2 clinical diagnoses. After sex, which showed by far the highest discriminative ability among the cases, the presence of pre- and perinatal risk and other developmental concerns featured as influential predictors of both ADHD and ASD classes. Not surprisingly, given the significant heritability of the 2 conditions, having a family member with reported difficulties was also a relevant predictor of the classification.
On the other hand, DT and LR models identified feature rankings that were, except for sex, significantly different for ADHD and ASD classification. This discrepancy could be due to the underlying assumptions of the different ML methodologies. Whereas LR models assume linear relationships between predictors and outcomes, DT and RF models could exploit nonlinear relationships and interactions within the data [ 18 ]. Consequently, some degree of variation in predictor ranking is expected, further highlighting the diverse nature of insights gained from different analytical methodologies. Finally, it should be remembered that it is impossible to conclude the causality and direction of the interrelations among predictors in the ML model.
Our RF model’s accuracy was in line with previous ML classification approaches to questionnaire data [ 10 , 11 , 15 , 16 , 29 ] and other data sources [ 30 - 37 ].
Nevertheless, these classification models outperformed recent work from our group, where we identified children with ADHD with an accuracy of up to 82% using a DT-based supervised ML approach [ 3 ]. Despite some methodological differences, the higher level of accuracy obtained in the current work underscores the potential of RF models in increasing the precision of computer-aided diagnosis. Altogether, this pattern of findings suggests that the RF model outperformed both the DT and LR models in effectively categorizing neurodevelopmental conditions based on parent-reported socio-anamnestic information, as highlighted by previous studies [ 27 , 28 , 38 ].
In the domain of child and adolescent neuropsychiatry, the diagnostic process includes an initial stage where anamnestic, sociodemographic, and behavioral data need to be collected. This data gathering can be remotely performed through internet-based parent reports, as evidenced by previous studies [ 5 , 6 ]. With this regard, the MedicalBIT platform currently represents the first Italian internet-based screening instrument for child and adolescent neuropsychiatric conditions [ 6 ]. As the data are compiled in databases within MedicalBIT, the exploitation of ML models can prompt the classification of the probable diagnostic risk associated with new subjects seeking assistance. The significant predictive value of the models developed in this study might be valuable to support the clinical practice of diagnosing neurodevelopmental conditions.
Despite the encouraging findings, this study is not free of limitations. First, our ML models exclusively rely on parent-reported data. Existing literature [ 37 ] has previously indicated that the reliability of these reports could be negatively influenced by factors such as the possibility of accessing digital tools, intrinsic comprehension difficulties, or general parental educational attainment. Second, our sample exclusively included children and adolescents from a geographically restricted region (Northern Italy). The generalizability of current findings to populations from different areas needs cautious consideration. Third, the relatively low occurrence of ADHD-ASD comorbidity in our cohort prevents us from developing classification models tailored for more nuanced diagnostic presentations, such as either ADHD- or ASD-only versus ADHD-ASD comorbid presentation. Therefore, future extensions of this study should consider including broader cohorts of participants to consider this possibility.
Within the rapidly evolving context of “psycho-informatics,” we believe that the current work represents a noteworthy effort in the realm of computational psychometrics [ 28 ]. Through an exploration of remotely collected parent-reported socio-anamnestic data, the current research has revealed promising avenues for enhancing the diagnostic process of neurodevelopmental and psychopathological conditions. Integrating digital platforms for data collection and ML could offer clinicians a dynamic tool supporting their diagnostic decisions. Within the health care systems, clinical teams confront a scarcity of personnel, with high emotional and cognitive demands for the actual staff [ 38 ]. In this context, this research represents a preliminary effort to mitigate the clinicians’ workload by automating specific tasks (such as data collection and analysis). If proven effective, this approach could leave more time for clinicians to nurture the essential patient-clinician bond, a facet that remains irreplaceable by artificial intelligence technologies.
SG performed statistical analyses and wrote the first draft. SBC, NB, MN, AS, ST, MM, and PC contributed clinical knowledge regarding neurodevelopmental and psychopathological conditions and telemedicine. AC, GC, and PC contributed scientific knowledge on health care technology. SBC, NB, and PC contributed to the development of the telemedicine platform through clinical consultation for variable selection. All authors revised and approved the final manuscript.
The data sets analyzed during this study are available from the corresponding author on reasonable request.
None declared.
attention-deficit/hyperactivity disorder |
autism spectrum disorder |
Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) |
decision tree |
leave-one-out cross-validation |
logistic regression |
Medea Information and Clinical Assessment On-Line |
machine learning |
no information rate |
out of the bag |
random forest |
Shapley additive explanations |
Edited by A Mavragani; submitted 15.11.23; peer-reviewed by P Washington, S Young, A Madevska Bogdanova; comments to author 10.01.24; revised version received 27.03.24; accepted 25.04.24; published 29.07.24.
©Silvia Grazioli, Alessandro Crippa, Noemi Buo, Silvia Busti Ceccarelli, Massimo Molteni, Maria Nobile, Antonio Salandi, Sara Trabattoni, Gabriele Caselli, Paola Colombo. Originally published in JMIR Formative Research (https://formative.jmir.org), 29.07.2024.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Formative Research, is properly cited. The complete bibliographic information, a link to the original publication on https://formative.jmir.org, as well as this copyright and license information must be included.
IMAGES
VIDEO
COMMENTS
Nashville, TN 3720 3 USA. t [email protected]. Abstract. The physical "brick and mortar" classroom is starting to lose its monopoly as the place of. learning. The Internet has made ...
This CoI model emphasizes the importance of social, cognitive, and teaching presence in creating meaningful and effective online learning. ... including higher academic achievement and greater satisfaction with online learning. The study's findings suggest that there is a need to improve students' online self-regulation skills, learning ...
This article reports on a large-scale (n = 987), exploratory factor analysis study incorporating various concepts identified in the literature as critical success factors for online learning from the students' perspective, and then determines their hierarchical significance. Seven factors--Basic Online Modality, Instructional Support, Teaching Presence, Cognitive Presence, Online Social ...
The purpose of this study is to analyze the effect of online education, which has been extensively used on student achievement since the beginning of the pandemic. In line with this purpose, a meta-analysis of the related studies focusing on the effect of online education on students' academic achievement in several countries between the years 2010 and 2021 was carried out. Furthermore, this ...
Distance learning has evolved over many generations into its newest form of what we commonly label as online learning. In this second-order meta-analysis, we analyze 19 first-order meta-analyses to examine the impact of distance learning and the special case of online learning on students' cognitive, affective and behavioral outcomes.
Online learning is one of the educational solutions for students during the COVID-19 pandemic. Worldwide, most universities have shifted much of their learning frameworks to an online learning model to limit physical interaction between people and slow the spread of COVID-19. The effectiveness of online learning depends on many factors, including student and instructor self-efficacy, attitudes ...
Zhang et al. (2022) implemented a bibliometric review to provide a holistic view of research on online learning in higher education during the COVID-19 pandemic period. They concluded that the majority of research focused on identifying the use of strategies and technologies, psychological impacts brought by the pandemic, and student perceptions.
This study also confirms the positive impact of online learning adaptability on satisfaction towards online teaching, which is in line with previous research that adaptability is an important ...
Cross-cultural online learning is gaining importance along with access in global settings. In addition, providing inclusive opportunities for all learners and in ethical ways is being examined. 12: ... Also, since there is no recent study examining online learning research themes in the last decade, this study strives to address that gap by ...
This systematic analysis examines effectiveness research on online and blended learning from schools, particularly relevant during the Covid-19 pandemic, and also educational games, computer-supported cooperative learning (CSCL) and computer-assisted instruction (CAI), largely used in schools but with potential for outside school.
According to the study of Bernard et al. ( 2004 ), this meta-analysis focuses on the activities done in online education lectures. As a result of the research, an overall effect size close to zero was found for online education utilizing more than one generation technology for students at different levels.
Given the dearth of research on the academic impact of online learning, especially in light of Covid-19 in the educational arena, the present study aims to address that gap by assessing the ...
1.1. Related literature. Online learning is a form of distance education which mainly involves internet‐based education where courses are offered synchronously (i.e. live sessions online) and/or asynchronously (i.e. students access course materials online in their own time, which is associated with the more traditional distance education).
This study explores how students at different stages of their K‐12 education reacted to the mandatory full‐time online learning during the COVID‐19 pandemic. For this purpose, we conducted a province‐wide survey study in which the online learning experience of 1,170,769 Chinese students was collected from the Guangdong Province of China.
This study reveals the importance of online learning since, clearly, the performance of students has been better via this method than traditional learning. During the COVID-19 pandemic student commitment to class attendance online has increased, along with participation and interaction, and marks.
A wide variety of terminology is used in varied and nuanced ways in educational literature to describe student learning mediated by technology, including terms such as virtual learning, distance learning, remote learning, e-learning, web-based learning, and online learning (e.g., Moore, Dickson-Deane, & Galyen, 2011; Singh & Thurman, 2019).For example, in a systematic review of the literature ...
online classes could affect the academic performance of students. This paper seeks to study the. impact of online learning on the academic performance of university students and to determine. whether education systems should increase the amount of online learning for traditional in-class. subjects.
In a recent study, researchers found that 18% of parents pointed to greater flexibility in a child's schedule or way of learning as the biggest benefit or positive outcome related to remote learning ( School Psychology, Roy, A., et al., in press).
The CoI model has formed the basis for a good deal of research on online learning. Most of this research. has focused on one of the three pr esences, social presence being the most frequently ...
Research across disciplines has demonstrated that well-designed online learning can lead to students' enhanced motivation, satisfaction, and learning [1,2,3,4,5,6,7].]. A report by the U.S. Department of Education [], based on examinations of comparative studies of online and face-to-face versions of the same course from 1996 to 2008, concluded that online learning could produce learning ...
used to select 18 students from 3 university departments to participate in the study. Research questions focused on how participants perceived their learning experiences in online courses and how they described interactions with instructors and other students. Data collection was multimodal. The interviews were conducted in face-to-face format,
This quantitative study aims to investigate the relationship between e-education and reading comprehension skills acquisition. It also examines if the previous relationship may impact students' results in the exams. It also analyses the relationship between students' knowledge in ICT and their perception and acceptance of online education.
Gaining different perspectives and learning about different cultures also aids children's thinking skills. This highlights the importance of online learning for students because it can open the door to new opportunities and help them develop skills that will serve them well in their future careers. Offers flexible learning hours.
The incorporation of specific push/pull factors in our model is based on previous research relating to online education, as well as online working, which can, to a certain extent, be applied to the educational context. ... Acknowledging the impact of technostress on satisfaction with online learning has considerable importance in the realm of ...
The COVID-19 pandemic accelerated the adoption of online learning, particularly in higher education institutions. This shift underscores the importance of sustainable education practices aligned with the United Nations' Sustainable Development Goals (SDGs). SDG 4 emphasizes inclusive and equitable quality education, highlighting how online learning environments can enhance accessibility and ...
This study explored online assessment and highlighted significant assessment strategies, such as capstone projects, quizzes, and self-assessments. ... the significance of the learning process and the students' frequent efforts were emphasized by this study. There is an emerging perspective that recognizes the ... Her research interests ...
This study aims to contribute to the themes expounded by Deuchar (2022; British Educational Research Journal, 48, 504-518) concerning international students' strengths, contributions, and practices in higher education by focusing on the experiences and practices of Chinese doctoral students across four English higher education institutions.The investigation deploys Bourdieu's tools of ...
Example Of Online Safety Training: An Award-Winning Case Study The Wonderful Company's PACE 360 Program. The Wonderful Company, a global leader in healthy brands and snacks, faced a unique challenge in ensuring the safety of its diverse workforce, which includes both office staff and field employees operating a wide range of vehicles and equipment.
Key findings indicated that the group difference was insignificant, suggesting the importance of metacognitive strategies in supporting adaptability and career resilience for young and mature employees. Finally, limitations to the study were presented, and recommendations were made for future workplace research and practice.
Background: Diagnosis of child and adolescent psychopathologies involves a multifaceted approach, integrating clinical observations, behavioral assessments, medical history, cognitive testing, and familial context information. Digital technologies, especially internet-based platforms for administering caregiver-rated questionnaires, are increasingly used in this field, particularly during the ...