Diagnostic utility. The present data found mixed support for the diagnostic utility of the SCARED. In both studies, HCs consistently had scores below Total, GA, and SA clinical cutoffs, regardless of informant. Results were less consistent for anxious youth: among those in the treatment-seeking sample, parent-report scores exceeded clinical cutoffs for Total, GA and SA subscales among GAD and GAD+SAD samples. There is less support for the SCARED’s diagnostic utility in the community sample, where no patient group fell above clinical cutoffs for any of the three scales; however, this could be a result of the relatively small number of patients in each diagnostic group. ROC analyses corroborate this pattern of findings. In the treatment-seeking sample, across informants, scores successfully discriminated the presence/absence of any anxiety disorder, GAD, and SAD on the Total, GA, and SA subscales, respectively. While specificity was uniformly high across informants and scales (> 85%), sensitivity ranged from 57% detection rate for child-report on the GA subscale to 75% detection rate for parent-report on the GA and SA subscales. This pattern of results is consistent with other ROC analyses of the SCARED which identified high specificity (0.89-0.92) at the cost of low sensitivity (0.44) (e.g., Xxxxxxx et al., 2007). Rates dropped substantially in the community sample. While specificity remained relatively high (> 76%), sensitivity was under 50% for child- and parent-report for Total and GA subscales, reaching a maximum of nearly 67% detection rate for parent-report on the SA subscale. One of the key motivations for developing the SCARED was to address the issue of under-diagnosis and under-treatment of anxious youth who may receive less clinical attention due to an absence of behavioral problems (Xxxxxxxx et al., 1999). Thus, a tool used to screen for anxiety disorders with specific clinical cutoffs should have appropriate levels of sensitivity, to ensure that youth with the disorder do not go undiagnosed or untreated, and specificity, to ensure diagnoses and treatment are not provided unnecessarily to non-anxious individuals. Given the high long-term cost of failing to treat an anxious child before they develop long term, difficult to treat symptoms (Xxxxxxxx et al., 2011), a screening tool for pediatric anxiety may choose to sacrifice specificity to increase sensitivity. Thus, more inclusive clinical cutoffs may be beneficial. Indeed, researchers have suggested clinical cutoff...
Diagnostic utility. One-sample t-tests assessed whether child- and parent-report scores on Total, GA, and SA subscales differed from clinical cutoffs for each diagnostic group (HC, GAD, SAD, GAD+SAD). Diagnostic utility will be confirmed if 1) HC scores are below the clinical cutoff on each scale; 2) each patient group (GAD, SAD, GAD+SAD) exceeds the clinical cutoff for the Total score (≥ 25); 2) the GAD and SAD groups exceed the clinical cutoff for GA (≥ 9) and SA (≥ 8) subscales, respectively; and 3) the GAD+SAD group exceeds the clinical cutoff for both the GA and SA subscales. Diagnostic utility was further assessed with receiver operating characteristic (ROC) analyses to quantify the sensitivity (true positive rate) and specificity (true negative rate) of the established clinical cutoff scores in each sample. Sensitivity indicates the probability that a child’s SCARED score will meet or exceed the clinical cutoff when the child is diagnosed on the K-SADS-PL, and the specificity indicates the probability that a child’s score will fall below the clinical cutoff when the child is not diagnosed on the K-SADS-PL. Positive and negative predictive values and overall correct classification were also quantified. Positive Predictive Values (PPV) and Negative Predictive Values (NPV) reflect rates of true positive and negative scores. Specifically, the PPV and NPV are the number of true observations (according to the K-SADS-PL) divided by the total number of positive or negative observations, according to SCARED cutoff scores. In this case, the PPV was calculated as the percent of participants that met or exceeded the SCARED’s clinical cutoff, and had a diagnosis on the K-SADS-PL. Similarly, the NPV was calculated as the percent of participants that fell below the SCARED’s clinical cutoff and did not have a diagnosis on the K-SADS-PL. The Overall Correct Classification (OCC) is the sum of the number of true positives and true negatives, divided by the overall number of participants. This gives the percent of participants correctly classified by the SCARED as either having a true diagnosis or not having a diagnosis.
Diagnostic utility. Each workstation and laptop computer shall be provided with diagnostic utility software. The software proposed shall be the latest release of the package as of the closing date in the RFP.
