{"doi":"10.1002/jcv2.12101","title":"Challenges and opportunities for the early identification of ADHD across multiple levels of analysis – A reflection on Tobarra‐Sanchez et al. (2022)","abstract":"The number of publications examining early markers of ADHD is rapidly growing (see recent review and meta-analysis by Shephard et al., 2022), each contributing important scientific information. Although some of the reported findings are also clinically informative, the chasm between these scientific findings and clinical translation remains large. This is, in part, due to approaches that prioritize individual early predictors rather than more comprehensive, multivariable examinations of early markers of ADHD which evaluate contributions of numerous predictors simultaneously. Moreover, many published studies fail to incorporate predictors spanning multiple levels of analysis, from biology to behavior. As the first study to include a combination of genetic risk via ADHD polygenic risk scores (PRS) and clinical, developmental, and behavioral predictors acquired over the first 30 months of life in relation to later childhood ADHD, Tobarra-Sanchez et al. (2022) have taken an important step toward better understanding early predictors of ADHD. Making use of the Avon Longitudinal Study of Parents and Children (ALSPAC), a large and well-characterized UK population-based birth cohort, the authors examined a sequence of univariable models culminating in a multivariable model aimed at predicting later ADHD outcomes. Ultimately, they found that the ADHD PRS, temperament-based activity level at 2 years of age, lower parent-rated fine motor skills at 18 months of age, and lower parent-rated vocabulary at 24 months of age predicted later dimensionally measured ADHD symptoms, while ADHD PRS and temperament-based activity level predicted DSM-IV ADHD diagnosis. The potential implications of this work for future research are numerous. For example, it highlights the pressing need for prospective studies from birth specifically designed to address questions about early markers of ADHD (despite known limitations of such designs). Large birth cohort studies provide invaluable information but, due to their nature, may rely more heavily on parent-report metrics versus direct assessment and deep phenotyping, along with variable batteries across time. When used to address research questions that they were not necessarily intended to answer at the outset, some limitations are therefore imposed. In this case, data were acquired across multiple ages between 15 and 30 months and were not always simultaneously measured, likely resulting in a loss of developmental nuance. For example, hearing/vision concerns were assessed at 15 months, motor skills were assessed at 18 months, vocabulary and temperament were assessed at 24 months, and regulatory issues were evaluated at both 24 (feeding) and 30 (sleeping/crying) months; these data were pooled in multivariable models despite dramatic changes in development during this period. Indeed, development is particularly dynamic at such early ages, making it difficult to extrapolate precise developmental meaning or potential effects of developmental cascades. Prospective studies focused on early markers of ADHD will provide opportunities to build upon Tobarra-Sanchez et al.’s (2022) findings by carefully tracking development of each of these identified domains. They will also more fully address questions related to heterotypic and homotypic continuity and, relatedly, whether there is variation in the composition of predictive multivariable models based on different ages and stages of development. Equally important to what did predict ADHD outcomes, this study raises intriguing questions based upon what did not predict ADHD outcomes in the univariate models, or which did predict ADHD outcomes in univariate models but not once included in the multivariable model. This includes a range of pre- and perinatal variables (e.g., young maternal age at birth, preterm birth, intrauterine growth restriction, low APGAR score), as well as child-level behavioral variables including temperament-based distractibility and regulatory probl","journal":"JCPP Advances","year":2022,"id":295695,"datarank":0.0,"base_score":0.0,"endowment":0.0,"self_citation_contribution":0.0,"citation_network_contribution":0.0,"self_endowment_contribution":0.0,"citer_contribution":0.0,"corpus_percentile":null,"corpus_rank":null,"citation_count":1,"citer_count":0,"citers_with_citation_signal":0,"citers_with_endowment":0,"datacite_reuse_total":0,"is_dataset":false,"is_dataset_confidence":0.9566,"is_data_producer":false,"deposit_databanks":null,"is_oa":true,"file_count":0,"downloads":0,"has_version_chain":false,"published_date":"2022-01-01","fair_score":null,"fair_percentile":null,"algorithm_id":"datarank_citation_only_1hop_v6","ranking_scope":"data_only","authors":[{"id":326470,"name":"Meghan Miller","orcid":"0000-0002-1260-4149","position":0,"is_corresponding":true}],"reference_count":8,"raw_metadata":null,"created_at":"2026-07-19T00:31:08.861377Z","pmid":"37431385","pmcid":null,"fwci":null,"citation_percentile":null,"influential_citations":0,"oa_status":null,"license":null,"views":0,"total_file_size_bytes":0,"version_count":0,"fair_f":null,"fair_a":null,"fair_i":null,"fair_r":null,"fair_zscore":null,"fair_rationale":null,"fair_model":null,"fair_agent_version":null,"fair_fulltext_source":null,"fair_has_llm":null,"fair_computed_at":null,"clinical_trials":[],"software_tools":[],"db_accessions":[],"linked_datasets":[],"topics":[]}