Defining a minimal clinically important difference for the Vanderbilt ADHD Parent Rating Scale in children with Attention-Deficit/Hyperactivity Disorder

Abstract

Background

Attention-deficit/hyperactivity disorder (ADHD) is a prevalent neurodevelopmental condition in children and adolescents that poses substantial challenges to health, functioning, and well-being. This study aimed to estimate the minimal clinically important difference (MCID) for the Chinese version of the Vanderbilt ADHD Parent Rating Scale (VAFPRS) and its key score domains using anchor-based and distribution-based methods.

Methods

In this multicenter prospective observational study, 99 children aged 4–17 years who were diagnosed with ADHD were recruited from four tertiary hospitals in China. Caregivers completed the VAFPRS at baseline and follow-up and the Clinical Global Impression–Improvement (CGI-I) scale at follow-up. MCID values were estimated using the anchor-based median change method, the receiver operating characteristic (ROC) curve method, and the distribution-based 0.5 standard deviation (SD) method.

Results

A total of 81 patients completed the follow-up assessment. CGI-I scores were significantly correlated with changes in the total VAFPRS score, core symptoms, inattention, and hyperactivity/impulsivity (r = 0.38–0.43, all p < 0.01). The corresponding anchor-based median MCID estimates were − 8, −5, − 2, and − 3, respectively. ROC curve analysis yielded lower thresholds for these domains (− 3.5, − 2.5, − 1.5, and − 2.5), with acceptable discrimination (AUC ≥ 0.70) for identifying improvement.

Conclusion

This study estimated MCID values for the Chinese VAFPRS in children and adolescents with ADHD. Anchor-based median estimates derived from the minimally improved CGI-I group were − 8 for the total score, − 5 for core symptoms, − 2 for inattention, and − 3 for hyperactivity/impulsivity. ROC-derived thresholds were lower and may provide complementary classification cutoffs for identifying improvement status.

Clinical trial number

Not applicable.

Background

Attention-deficit/hyperactivity disorder (ADHD) is a prevalent neurodevelopmental condition in children and adolescents, characterized by persistent inattention, hyperactivity, and impulsivity, and is associated with impairments in academic achievement, quality of life, and social functioning [1]. According to global estimates, the age-standardized prevalence of ADHD is 1131.9 per 100,000 people (95% CI: 831.7-1494.5), with male individuals affected nearly twice as often as females [2]. Among children, the prevalence of ADHD is significantly elevated, with rates of 7.6% (95% CI: 6.1–9.4%) in those aged 3–12 years and 5.6% (95% CI: 4.8-7.0%) in those aged 12–18 years [3]. Although the specific etiology of ADHD remains uncertain, studies have identified genetic predispositions, prenatal influences, and environmental factors such as nutrition and micronutrient intake in childhood [4]. ADHD is diagnosed clinically through direct observation, caregiver interviews, and validated rating scales, all of which are subject to varying degrees of subjectivity. The Vanderbilt ADHD Parent Rating Scale (VAFPRS) developed by Wolraich and colleagues, provides a broader caregiver-reported assessment by covering core ADHD symptoms as well as oppositional defiant behaviors and academic/social functioning. It is based on the Diagnostic and Statistical Manual of Mental Disorders (DSM) ADHD symptom criteria and has been widely used in clinical and research settings [5]. The American Academy of Pediatrics (AAP) and the National Initiative for Children’s Healthcare Quality (NICHQ) have incorporated this tool into the ADHD diagnosis and treatment toolkit and recommend its clinical use [6]. To address the lack of a validated Chinese version of the VAFPRS, our team previously undertook a systematic translation and cultural adaptation process, supported by expert consensus and pilot testing. Subsequent clinical trials demonstrated that the scale exhibited strong item correlation as well as good reliability and validity, supporting its use in the clinical assessment of ADHD.

Previous clinical trials on ADHD treatments have predominantly used p values and standardized mean differences (SMDs) to evaluate treatment effects [4]. However, statistical significance—affected by sample size and group variance—does not necessarily reflect clinically meaningful changes. In large-sample studies, even minor changes may achieve statistical significance. Thus, determining the minimal clinically important difference (MCID), the smallest change in score that patients perceive as beneficial—has become increasingly important in clinical trials [7]. The MCID, also known as the minimum important difference (MID), is defined as the smallest change that, without considering side effects or cost burdens, is perceived by patients as beneficial and sufficient to warrant a change in treatment management [8]. When the change in a scale score exceeds the MCID value, it is regarded as clinically meaningful. Furthermore, the MCID plays an important role in interpreting clinical trial outcomes and estimating sample sizes [9]. Over the past 30 years, as an increasing number of studies, meta-analyses, and guidelines have incorporated the MCID as a tool to enhance the measurement of patient-reported outcomes, its application has become increasingly widespread [10]. Accordingly, this multicenter prospective observational study aimed to estimate the MCID for the Chinese VAFPRS via both anchor-based and distribution-based methods, thereby supporting clinical trial endpoint selection and sample size calculation.

Methods

Study design

This multicenter prospective observational study included children and adolescents with ADHD receiving standard treatments. The objective of this study was to estimate the MCID of the Chinese VAFPRS and its key score domains by analyzing changes in scores between baseline and follow-up assessments conducted at weeks 3 to 5. The inclusion criteria were as follows: (1) diagnosis of ADHD on the basis of DSM-5 criteria, confirmed by qualified clinicians; (2) age between 4 and 17 years; (3) caregivers had adequate reading comprehension and were able to complete the questionnaires reliably; (4) willingness and ability to comply with study visits; and (5) written informed consent provided by the caregiver, with assent or consent from the participant when aged ≥ 8 years. In this study, caregivers were defined as parents or other primary caregivers who regularly cared for the participant and were able to report the participant’s symptoms. The exclusion criteria were as follows: children with ADHD comorbid with other psychiatric disorders. The study protocol was approved by the Medical Ethics Committee of the First Affiliated Hospital of Tianjin University of Traditional Chinese Medicine (Approval No. TYLL2022[K]028). Written informed consent was obtained from all caregivers, and assent or consent was obtained from participants when applicable. Participants were recruited from four tertiary hospitals: the First Affiliated Hospital of Tianjin University of Traditional Chinese Medicine, Changzhou Children’s Hospital, Hubei Provincial Hospital of Traditional Chinese Medicine, and Baoding First Central Hospital. Clinical trial number: not applicable.

Vanderbilt ADHD Parent Rating Scale

The VAFPRS was completed by caregivers at baseline and again during follow-up visits (weeks 3 to 5). This scale has demonstrated strong psychometric properties for assessing ADHD symptom severity, functional impairments, and treatment-related changes [11–13]. The scale contains 34 items divided into four domains: inattention (items 1–9), hyperactivity/impulsivity (items 10–18), oppositional defiant behaviors (items 19–26), and academic/social functioning (items 27–34). Among them, items 1–18 are referred to as core symptoms. Items 1–26 are rated on a 4-point scale (0 = Never, 1 = Occasionally, 2 = Often, 3 = Very Often), whereas items 27–34 use a 5-point scale (1 = Excellent, 2 = Above Average, 3 = Average, 4 = Somewhat of a Problem, 5 = Problematic).

Clinical global impression–improvement

To anchor perceived clinical change, the same caregivers also completed the Clinical Global Impression–Improvement (CGI-I) scale at weeks 3 to 5. The CGI-I provides a global assessment of improvement based on the caregiver’s perspective relative to the baseline. The CGI-I uses a 7-point rating scale and asks caregivers, “Compared to the first time you completed this scale, how have the participant’s symptoms changed?” The response options included “Very much improved (1 point) “, “Much improved (2 points) “, “Minimally improved (3 points) “, “No change (4 points) “, “Minimally worse (5 points) “, “Much worse (6 points) “, and “Very much worse (7 points) “. The CGI-I has a detailed classification, which helps clinicians more effectively capture subtle yet meaningful changes and is commonly used to support the assessment of clinical improvement in ADHD treatment [14].

Statistical analysis

Anchor-based methods were used to evaluate changes in VAFPRS scores from baseline to follow-up across CGI-I categories. The CGI-I ratings were used as an external anchor to reflect caregiver-perceived global improvement [15]. For normally distributed differences between baseline and follow-up scores, the mean change was used to estimate the MCID. For nonnormally distributed data, the median change was reported [16]. To justify the use of the anchor-based method, correlations between CGI-I scores and changes in VAFPRS scores were tested [17]. Anchor validity was evaluated by examining the correlation coefficients between CGI-I scores and changes in VAFPRS scores. Statistically significant correlations with sufficient magnitude were considered to provide support for the use of CGI-I as an external anchor for MCID estimation. In this approach, the primary anchor-based MCID estimate was derived from participants classified as minimally improved on the CGI-I, because this category represents the smallest caregiver-perceived meaningful improvement.

We also applied the anchor-based receiver operating characteristic (ROC) curve method. CGI-I scores were dichotomized into “improved” (ratings of 1 to 3) and “not improved” (ratings of 4 to 7) [18]. ROC curves were generated, and optimal cutoff points were identified via the maximum Youden index (sensitivity + specificity – 1), minimizing classification errors. The change score at the selected cutoff was considered the ROC-derived threshold. Unlike the anchor-based estimate derived from the minimally improved group, the ROC-derived threshold was interpreted as a classification cutoff for distinguishing participants with improvement from those without improvement. These two approaches were therefore considered complementary: the median change estimate emphasized caregiver-perceived meaningful change, whereas the ROC derived threshold emphasized discriminatory performance. Discriminative accuracy was evaluated via the area under the ROC curve (AUC), with an AUC ≥ 0.7 considered acceptable and an AUC ≥ 0.8 considered excellent [19]. Bootstrap resampling (n = 500 iterations) was used to assess the robustness of the ROC-derived estimates. For each resample, the MCID and AUC values were recalculated, and the medians with 95% confidence intervals were reported [20]. Additionally, distribution-based MCID estimates were computed using 0.2, 0.5, and 0.8 times the baseline standard deviation (SD), representing small, moderate, and large changes, respectively. The 0.5 SD threshold was selected as the primary distribution-based reference [21].

Cases with more than 20% missing items on any scale were excluded from the analysis. For missing data below this threshold, multiple imputation via chained equations (MICE) was conducted. All the statistical analyses were performed via R version 4.4.1. Statistical significance was defined as a two-tailed p value < 0.05.

Results

Patient characteristics

Between September 2022 and September 2023, 99 pediatric patients diagnosed with ADHD were enrolled from four participating centers. Of these, 81 completed the follow-up assessment at week 4 (± 1 week). The sample included 70 boys and 11 girls, with a mean age of 9.4 ± 1.9 years. The baseline demographic and clinical characteristics are summarized in Table 1.

Table 1.

Baseline characteristics of pediatric patients with ADHD

Variable	Category	N	Percentage (%)
Study Site	Changzhou Children’s Hospital	30	37.0%
	Hubei Provincial Hospital of TCM	26	32.1%
	First Teaching Hospital of Tianjin University of TCM	18	22.2%
	Baoding First Hospital of TCM	7	8.6%
Sex	Male	70	86.4%
	Female	11	13.6%
Ethnicity	Han	60	74.1%
	Miao	1	1.2%
	Unknown	20	24.7%
Education Level of Respondent	High school or below	8	9.9%
	Junior college	33	40.7%
	Bachelor’s degree	31	38.3%
	Master’s degree or above	5	6.2%
	NA	4	4.9%
ADHD clinical presentation	Inattentive presentation	46	56.8%
	Hyperactive/impulsive presentation	20	24.7%
	Combined presentation	15	18.5%

Note: ADHD = attention-deficit/hyperactivity disorder; NA = not available; TCM = traditional Chinese medicine

VAFPRS scores and CGI-I

The mean scores (± SD) at baseline and follow-up for the total VAFPRS score, core symptoms, and domain scores were as follows: total score, 56.5 ± 14.8 (baseline) and 44.4 ± 11.6 (follow-up); core symptoms, 25.9 ± 9.1 and 18.8 ± 8.2; oppositional/defiant behavior, 8.1 ± 5.3 and 5.1 ± 2.8; and academic and social performance, 22.5 ± 4.6 and 20.5 ± 3.9.

At follow-up, patients were classified according to their CGI-I scores as follows: very much improved (n = 3), much improved (n = 18), minimally improved (n = 41), no change (n = 18), and minimally worse (n = 1). CGI-I scores were moderately correlated with changes in the total VAFPRS score (r = 0.41, p < 0.001) and core symptom score (r = 0.43, p < 0.001). Significant correlations were also observed between CGI-I scores and changes in the inattention domain (r = 0.38, p < 0.01) and hyperactivity/impulsivity domain (r = 0.39, p < 0.001). The correlation with changes in the oppositional/defiant behavior domain was weaker (r = 0.27, p < 0.05), whereas the correlation with changes in the academic and social performance domain was not statistically significant (r = 0.22). The detailed Pearson correlation coefficients and corresponding significance levels are presented in Table 2.

Table 2.

Pearson correlations (r) of CGI-I scores with baseline and change scores of the VAFPRS

VAFPRS score domain	r with baseline score	r with change score
Total VAFPRS Score	-0.29**	0.41***
Core symptoms (1–18)	-0.40***	0.43***
Inattention (1–9)	-0.22*	0.38**
Hyperactivity/impulsivity (10–18)	-0.39***	0.39***
Oppositional/Defiant Behavior (19–26)	-0.23*	0.27*
Academic and Social Performance (27–34)	0	0.22

Note: Change score was calculated as follow-up score minus baseline score; negative values indicate improvement. CGI-I = Clinical Global Impression–Improvement; VAFPRS = Vanderbilt ADHD Parent Rating Scale

***p < 0.001, **p < 0.01, *p < 0.05

Anchor-based MCID estimation using CGI-I

Owing to nonnormal distributions of change scores, MCID estimates are reported as medians with interquartile ranges (IQRs: 25th–75th percentiles). The primary anchor-based MCID estimates were derived from participants classified as “minimally improved” on the CGI-I, as this category represents the smallest caregiver-perceived meaningful improvement. Among participants rated as minimally improved (n = 41), the median changes (IQRs) in scores were as follows: total score, − 8 (− 15, − 3); core symptoms, − 5 (− 10, − 1); inattention, − 2 (− 6, − 1); hyperactivity/impulsivity, − 3 (− 4, 0); oppositional defiant behavior, − 2 (− 5, 0); and academic and social performance, − 1 (− 3, 0). The detailed results are presented in Table 3.

Table 3.

Meaningful changes in VAFPRS scores across CGI-I groups

CGI-I anchor	n	Meaningful VAFPRS change estimate
Median (25th,75th)		Total VAFPRS score	Core symptoms	Inattention	Hyperactivity/ impulsivity	Oppositional defiant behavior	Academic and social performance
Much improved	18	-10.5(-20.8, -6.3)	-7(-13.5, -4.25)	-4.5(-6, -2)	-3.5(-7.5, -1)	-3(-5.8,0)	-2(-5,0)
Minimally improved	41	-8(-15, -3)	-5(-10, -1)	-2(-6, -1)	-3(-4,0)	-2(-5,0)	-1(-3,0)
No change	18	-3(-11.3, -0.3)	-1(-5,0)	0(-4.3,0)	-0.5(-2,0)	-0.5(-2.8,0)	-0.5(-2.8,0)

Note: Values are medians with interquartile ranges (25th–75th percentiles). Clinically, the minimally improved group provides the primary anchor-based MCID estimates for improvement, corresponding to reductions of approximately 8 points in the total VAFPRS score, 5 points in core symptoms, 2 points in inattention, and 3 points in hyperactivity/impulsivity. CGI-I = Clinical Global Impression–Improvement; VAFPRS = Vanderbilt ADHD Parent Rating Scale

Because the “very much improved” (n = 3) and “minimally worse” (n = 1) CGI-I categories included very few participants, these categories were not included in Table 3 and were not used for group-based estimation of medians and IQRs. For descriptive completeness, the observed median changes among the three participants rated as very much improved were − 44 for the total VAFPRS score, − 27 for core symptoms, − 14 for inattention, − 13 for hyperactivity/impulsivity, − 9 for oppositional defiant behavior, and − 10 for academic and social performance. The single participant rated as minimally worse showed changes of 2, 2, 3, − 1, 0, and 0 in these domains, respectively. These observations were considered descriptive only and were not used to estimate the MCID. Given the very small number of patients reporting worsening, an MCID for deterioration was not estimated.

ROC-based MCID estimation anchored to CGI-I

Using an effective sample size of 81 participants who completed the follow-up assessment and were included in the CGI-I anchored change-score analysis, ROC analysis with 500 bootstrap samples yielded the following ROC-derived thresholds and 95% confidence intervals (CI) for the VAFPRS total and domain scores: total score − 3.5 (− 20.5, − 1.5), core symptoms − 2.5 (− 8.5, − 1.5), inattention − 1.5 (− 7.5, − 0.5), hyperactivity/impulsivity − 2.5 (− 3.5, − 0.5), oppositional defiant behavior − 1.5 (− 5.5, 1.5), and academic and social performance − 1.5 (− 7.03, 2.5). Except for the oppositional defiant and academic and social performance dimensions, the AUC values for all other domains exceeded 0.70, indicating acceptable discriminatory power. The detailed AUC, sensitivity, and specificity metrics (with 95% CI) are presented in Table 4. The corresponding ROC curves and cutoff points are illustrated in Fig. 1. Furthermore, AUC density plots showed concentrated peak distributions for all domains except academic and social performance, supporting the stability of the ROC-based classification model, as shown in Fig. 2.

Table 4.

ROC-derived thresholds and classification performance for VAFPRS score domains using bootstrap method

Items	Total VAFPRS score	Core symptoms	Inattention	Hyperactivity/ impulsivity	Oppositional defiant behavior	Academic and social performance
ROC-derived threshold (95%CI)	-3.5 (-20.5, -1.5)	-2.5 (-8.5, -1.5)	-1.5 (-7.5, -0.5)	-2.5 (-3.5, -0.5)	-1.5 (-5.5, 1.5)	-1.5 (-7.03, 2.5)
AUC (95%CI)	0.72 (0.56, 0.87)	0.74 (0.60, 0.87)	0.72 (0.56, 0.85)	0.70 (0.57, 0.80)	0.65 (0.44, 0.78)	0.58 (0.43, 0.73)
Sensitivity (95%CI)	0.81 (0.35, 0.95)	0.76 (0.57, 0.88)	0.76 (0.31, 0.89)	0.58 (0.39, 0.75)	0.61 (0.11, 0.98)	0.57 (0.10, 0.99)
Specificity (95%CI)	0.65 (0.39, 1.00)	0.74 (0.50, 0.96)	0.68 (0.44, 1.00)	0.85 (0.63, 1.00)	0.72 (0.25, 1.00)	0.67 (0.12, 1.00)

Note: The ROC-derived thresholds indicate the cutoff values used to classify participants with improvement versus those without improvement. Acceptable discrimination was observed for the total VAFPRS score, core symptoms, inattention, and hyperactivity/impulsivity, whereas weaker discrimination was observed for oppositional defiant behavior and academic and social performance. AUC = area under the curve; CI = confidence interval; MCID = minimal clinically important difference; ROC = receiver operating characteristic; VAFPRS = Vanderbilt ADHD Parent Rating Scale

Fig. 1.

ROC curves and cutoff points of variables based on the bootstrap method. Note: diff_VAF1_9 = inattention (items 1–9); diff_VAF10_18 = hyperactivity/impulsivity (items 10–18); diff_VAF19_26 = oppositional defiant behaviors (items 19–26); diff_VAF27_34 = academic and social performance (items 27–34); diff_VAF1_18 = core symptoms (items 1–18); diff_VAF1_34 = total VAFPRS score. ROC = receiver operating characteristic; VAFPRS = Vanderbilt ADHD Parent Rating Scale

Fig. 2.

AUC density curves of VAFPRS scores based on the bootstrap method. Note: diff_VAF1_9 = inattention (items 1–9); diff_VAF10_18 = hyperactivity/impulsivity (items 10–18); diff_VAF19_26 = oppositional defiant behaviors (items 19–26); diff_VAF27_34 = academic and social performance (items 27–34); diff_VAF1_18 = core symptoms (items 1–18); diff_VAF1_34 = total VAFPRS score. AUC = area under the curve; VAFPRS = Vanderbilt ADHD Parent Rating Scale

As a complementary analysis, distribution-based estimates based on 0.5 times the baseline SD were 7.4 for the total score, 4.6 for core symptoms, 2.4 for inattention, 3.1 for hyperactivity/impulsivity, 2.6 for oppositional defiant behavior, and 2.3 for academic and social performance. Given that lower VAFPRS scores indicate improvement, these estimates correspond to score reductions and were broadly consistent in absolute magnitude with the anchor-based median estimates.

Discussion

Patient-reported outcome measures (PROMs) are increasingly used in clinical research, but their clinical interpretation remains challenging because statistically significant differences do not necessarily indicate meaningful change from the caregiver’s perspective. Identifying thresholds for clinically meaningful change is therefore important for evaluating treatment effects and informing clinical decision-making [22]. In pediatric populations, MCID estimates based on caregiver-reported measures can provide quantitative references for interpreting treatment-related changes.

This study collected baseline and follow-up data at weeks 3–5 from participants with ADHD to evaluate changes in VAFPRS and CGI-I scores. Anchor-based and ROC-based methods were used to estimate MCID values across VAFPRS domains. Initial correlation analysis showed that changes in academic and social functioning scores were not significantly associated with CGI-I ratings. This finding may be attributable, at least in part, to contextual factors within the Chinese socio-educational environment. For example, the predominance of single-child households in China may limit children’s opportunities for sibling interaction, thereby affecting the development and observation of social functioning. In addition, academic evaluations in China are often conducted periodically rather than continuously, and caregiver involvement in educational activities may vary. These factors may reduce the sensitivity of academic and social functioning measures to short-term clinical improvement. A similar caution applies to the oppositional defiant behavior dimension. Although its correlation with CGI-I was statistically significant, the coefficient was < 0.3, indicating a weak association. This may partially reflect caregiver response patterns toward behaviorally sensitive items, such as “Is angry or resentful” and “Is spiteful and wants to get even.” Some caregivers may be reluctant to endorse these items, which could introduce reporting bias. Overall, the VAFPRS appears most useful for monitoring changes in the total score, core symptoms, and the inattention and hyperactivity/impulsivity domains in this clinical context. The anchor-based analysis across CGI-I groups yielded MCID estimates of − 8, −5, − 2, and − 3 for the total score, core symptoms, inattention, and hyperactivity/impulsivity, respectively. These estimates were consistent with the scale content and broadly aligned in absolute magnitude with the distribution-based estimate of 0.5 times the baseline SD. ROC analysis using the CGI-I anchor yielded lower ROC-derived thresholds of − 3.5, − 2.5, − 1.5, and − 2.5, with corresponding AUC values ≥ 0.70 for these domains, indicating acceptable discrimination. As specified in the Methods, the ROC-derived threshold was interpreted as a classification cutoff for distinguishing participants with improvement from those without improvement, whereas the anchor-based estimate was derived from the minimally improved group. The median changes in the minimally improved group were larger than the ROC-derived thresholds, suggesting that caregiver-perceived minimal meaningful improvement corresponded to greater score reductions than the classification cutoffs identified by ROC analysis. These findings suggest that ROC-based thresholds may be useful for identifying improvement status, whereas anchor-based MCID estimates may better reflect changes perceived as meaningful by caregivers [23].

This study has several limitations. First, no restrictions were placed on treatment modality or symptom severity during enrollment, which may have introduced heterogeneity in response patterns and influenced MCID estimates. Second, owing to the small number of participants showing clinical deterioration (n = 1), the analysis focused primarily on MCID values indicating improvement. Third, use of the caregiver-rated CGI-I as the sole anchor may have introduced informant-related subjectivity. Because both CGI-I and VAFPRS were caregiver-reported, common-method bias may have inflated the observed association between the anchor and VAFPRS change scores. Nevertheless, caregiver-reported global improvement remains methodologically acceptable for estimating MCID values for caregiver-reported outcome measures in pediatric ADHD, because caregivers are key informants of symptom change in daily settings. Future studies should incorporate additional anchors, such as clinician-rated and teacher-reported global improvement measures, to triangulate MCID estimates and reduce informant-related bias. Stratification by intervention type is also needed to assess treatment-specific MCID values and improve generalizability.

Conclusions

This study estimated MCID values for the Chinese VAFPRS in children and adolescents with ADHD. Anchor-based estimates derived from the minimally improved CGI-I group were − 8 for the total score, − 5 for core symptoms, − 2 for inattention, and − 3 for hyperactivity/impulsivity. ROC-derived thresholds showed acceptable discrimination for these domains and may provide complementary classification cutoffs for identifying improvement status. These findings may support interpretation of short-term treatment-related changes in VAFPRS scores.

Abbreviations

ADHD

Attention-Deficit/Hyperactivity Disorder

AUC

Area Under the Curve

CI

Confidence Intervals

CGI-I

Clinical Global Impression–Improvement

IQRs

Interquartile Ranges

MCID

Minimal Clinically Important Difference

ROC

Receiver Operating Characteristic

SD

Standard Deviation

VAFPRS

Vanderbilt ADHD Parent Rating Scale

Biographies

Shuaishuai Deng

is a postdoctoral researcher at the First Teaching Hospital of Tianjin University of Traditional Chinese Medicine. His main research interest is the evaluation of minimal clinically important difference (MCID) in pediatric scales. Lili Cai, MD, is a physician at the First Teaching Hospital of Tianjin University of Traditional Chinese Medicine. Her research focuses on the diagnosis, treatment, and evaluation of ADHD.

Lili Cai

is a physician at the First Teaching Hospital of Tianjin University of Traditional Chinese Medicine. Her research focuses on the diagnosis, treatment, and evaluation of ADHD.

Qiuhan Cai

is an attending physician at the First Teaching Hospital of Tianjin University of Traditional Chinese Medicine. Her research focuses on the development and evaluation of pediatric scales.

Siyuan Hu

is an attending physician at the First Teaching Hospital of Tianjin University of Traditional Chinese Medicine. Her research focuses on the development and evaluation of pediatric scales.

Jianting Wu

is an attending physician at Changzhou Children’s Hospital.

Yao Zhang

is an attending physician at the First Hospital of Traditional Chinese Medicine in Baoding.

Author contributions

Drs. Q.C. and S.H. were responsible for the clinical trial design and coordination among the study centers, and they are co-corresponding authors. Drs. S.D. and L.C. contributed to manuscript writing and data analysis, and they are co-first authors. Dr. J.W. and Dr. Y.Z. were responsible for data collection at Changzhou Children’s Hospital and Baoding First Hospital of Traditional Chinese Medicine subcenters, respectively.

Funding

This project was funded by the China Tianjin Municipal Education Commission Research Plan Project (2021KJ138).

Data availability

The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author upon reasonable request.

Declarations

Ethical approval and consent to participate

This study was conducted in accordance with the principles of the Declaration of Helsinki. The study protocol was approved by the IRB of The First Teaching Hospital of Tianjin University of TCM (Approval No. TYLL2022[K]028). Written informed consent was obtained from caregivers, and assent or consent was obtained from participants when applicable.

Consent for publication

Not applicable as this study does not include any identifying data.

Competing interests

The authors declare no competing interests.