Flexibility and Integration in Clinical Contexts

4 Interpretation The computerized reports that are available for the CDI 2 tool kit include standardized interpretations of the Total Score, scale scores, and, if applicable, subscale scores for each form. However, depending on the purpose of the assessment and the context of testing, some users may wish to interpret the results in a broader framework or integrate CDI 2 results with results from other types of assessments. Thus, the level of interpretation depends on the assessor, and the necessary information is presented below. The present chapter starts with a brief overview of the scores that can be derived from the CDI 2 and the factors that must be considered in their interpretation. Then, interpretive guidelines are presented as a series of five steps. Examples are provided to demonstrate how ratings from different respondents can yield a more comprehensive picture of a child's depressive symptoms, or conversely, point to unanticipated (but clinically meaningful) discrepancies among informants. By noting the various contexts that can determine how CDI 2 results are interpreted, this chapter also highlights the flexibility of the CDI 2 tool kit. CDI 2 Scores Raw Scores Raw scores are derived by summing the numeric values of the item responses that were endorsed on a particular form by a respondent. Recall that response options are defined such that higher scores indicate more marked or definite symptoms. Thus, for example, a CDI 2:Self-Report (CDI 2:SR) Total Raw Score of 28 indicates a considerably higher level of depression symptoms than does a Total Raw Score of 4. However, beyond such generic interpretation, one child's raw score on a scale is not readily comparable to those from other children. (as discussed further on) and the various scale scores of one child are not comparable to each other because they contain different numbers of items. To facilitate interpretation of the results, it is therefore recommended that raw scores be converted to standardized scores. T-scores and Percentile Ranks Standardized 7-scores have a mean of 50 and a standard deviation of 10' and thus can readily reflect what is typical or atypical for a child of a given age and sex. A 7-score of 50 for an 11-year-old girl, for example, is exactly equal to the mean score of 7- to 12-year-old girls in the standardization sample (reflecting the general community). High T-scores on the CDI 2 are indicative of high levels of depression symptoms. Thus, a 7-score of 70, which is two standard deviations above the mean, can be considered as a "Very Elevated" score. Such a high score mirrors the endorse-ment of numerous depression symptoms and multiple areas of functional impairment. As a further example, a T-score of 65 is 1.5 standard deviations above the mean. That score therefore is "Elevated" (as compared to the average score) and should trigger definite concern from the assessor. Table 4.1 provides the verbal interpretive labels or clas-sifications that are associated with various levels of CDI 2 7-scores. As shown in Table 4.1, 7-scores between 60 and 64 represent scores in the "High Average" range. Interpretation of the significance of High Average scores requires clinical judgment, as this range occurs about equally for those with and without clinical concerns. 7-scores that fall below 60 generally indicate typical levels of item endorsements for that child's age and sex. CDI 2 scores also can be expressed as percentile ranks. A percentile rank reflects the percentage of individuals in the normative group who scored the same as, or lower than, the respondent. For example, a score in the "90" percentile" indicates that 90% of the children in the standardization sample had the same or lower raw score. The higher the percentile rank, the more symptoms of depression the child is experiencing relative to his/her peer group. Percentiles are computed using the actual frequency counts in the normative sample (i.e., "empirical" percentiles). Like T-scores, percentiles a comparison of a person's scores Note that CDI 2 7-scores are linear. Linear 7-scores do not change the actual distribution of the CDI 2 scores. Therefore, although the mean and standard deviation are fixed at 50 and 10 respectively, CDI 2 scores remain skewed, reflecting the nature of depressive symptoms in the population. compared to similar age and same-sex peers. Use of either metric is acceptable. Empirical percentiles are available as one of the options in the CDI 2 Scoring Software and the MHS Online Assessment Center (see chapter 3 for details). In addition, empirical percentiles for the CDI 2 can be found in appendix D. The classifications in Table 4.1 also can be applied to the various percentile ranks and can be used to summarize results in nontechnical language. Note, however, that the score cutoffs that define the various categories are conventional guidelines and approximations and should not be used as absolute rules. For example, there is no reason to believe that there is a perceptible difference between a T-score of 69 and a T-score of 70. In fact, scores (such as 69 and 70) that fall near the juncture of two categories warrant special consideration. Indeed, it may be more informative to report a T-score of 69, for example, as falling in between Elevated and Very Elevated scores, rather than unequivocally placing it in the Elevated range. In other words, score-classification guidelines should not be applied automatically and the significance of test results should be determined via careful clinical consideration. Table 4.1. The Meaning of T-scores and Percentile Ranks T-score Percentile Rank Classification 70+ 98+ Very Elevated Score/Percentile Rank (Many more concerns than are typically reported) 65-69 93-97 Elevated Score/Percentile Rank (More concerns than are typically reported) 60-64- 84-92 High Average Score/Percentile Rank (Somewhat more concerns than are typically reported) 40-59 16-83 Average Score/Percentile Rank (Typical number of concerns) <40 <16 Low Score/Percentile Rank (Fewer concerns than are typically reported) Confidence Intervals All measurements contain some error (see Naglieri & Chambers, 2009 for a discussion). Chapter 5 (see Standard Error of Measurement) provides information on the measure-ment errors associated with the CDI 2 forms. The effects of measurement error are taken into account in score interpreta-tion by using confidence intervals. Specifically, confidence intervals take measurement error into account by providing, at a specific level of probability, a range of scores within which the true score is expected to fall. Both the 90% and 95% confidence intervals have been computed for all scales (see appendix E for more information on the computation of confidence intervals). For example, if a 7-year-old boy has a CDI 2:SR T-score of 60 on the Total Score scale, we can be highly confident (as reflected by the 90% confidence interval) that his true 7-score falls within the range of 54 to 66. The scoring software program automatically provides the confidence interval for every T-score. The 90% level of confidence is recommended and is the default setting for the scoring software program, although the user can choose either 90% or 95% confidence intervals. (See chapter 3 for details. on computerized scoring options.) For those who are hand-scoring the CDI 2, appendix E displays the 90% and 95% confidence intervals for all T-scores. Demographic Considerations As reviewed in chapter 1, studies have shown that the extent of depression symptoms in youth is partly a function of age and sex. Such age and sex effects on symptom endorse-ments also have been found on the CDI 2 forms. Therefore, T-score conversions for the normative sample are reported by age and separately for boys and girls. The percentiles in appendix D also are therefore also displayed by age and sex. Please see chapter 5 for detailed information about age and sex differences in CDI 2 scores. Demographic factors other than the child's age and sex also should be taken into consideration when interpreting results. For example, in more communally oriented cultures, interpersonal difficulties due to depression may be noticed more readily than atypical internal mood states. Or, in some families, gender-stereotyping may lead parents to minimize presumably atypical behavior in their offspring (e.g., crying in a 16-year-old boy). Such factors sometimes may account for non-convergent reports from a target youth and his parents and teachers and warrant further exploration. Age Effects Studies of the relations of age and depression symptoms among juveniles have consistently found that there is a marked increase in such symptoms around mid-adolescence (see chapter 1 for a review) especially among girls. Although various propositions have been put forth as to the causes of these age effects, none of them have garnered enough support to be widely accepted. It is none-theless clear that depression scores need to be examined in the context of age; the CDI 2 7-scores achieve exactly that goal. Namely, CDI 2 norms are age-based and are available for two age groups: younger (7- to 12-year-olds) and older respondents (13- to 17-year-olds). Sex Effects There is an impressive body of research evidence on the relations between the respondents' sex and the extent of their depression symptoms (also reviewed in chapter 1). Sex differences in depressive symptoms and disorders emerge around adolescence and tend to persist across the life span. While many researchers believe that these differences reflect the impact of sex hormones, supportive data are scant and the actual causal pathways have not been delineated. Statistical analyses of the CDI 2 data did show significant sex effects on many scales (see chapter 5 for more details). Thus, norms for CDI 2 7-scores also are presented separately for boys and girls. Combined-Sex Norms In some circumstances, it may be desirable to examine T-scores for a given sample regardless of respondents' sex. Therefore, combined-sex norms are available for the CDI 2 tool kit and this option can be selected in the scoring software. Additionally, appendix C presents tables of combined-sex norms that may be used when the CDI 2 forms are hand-scored. Step-by-Step Interpretation Guidelines This section provides a five-step strategy for interpreting the results of any of the forms that comprise the CDI 2 tool kit. Before beginning the step-by-step interpretation, the assessor should verify that the forms have been properly completed. If any of the tools required nonstandard administration, the assessor should keep in mind the broader context within which the results are interpreted. Step 1: Interpret the Scale Scores Interpretation can be at the level of the Total Score, scale scores, or subscale scores, depending on the particular CDI 2 form. Remember that the classifications for T-score levels, as well as the names of scales and subscales, are meant to serve as guidelines rather than absolute rules. A meaningful interpretation needs to consider item-level analysis, along with a consideration of the context of the testing and the broader environment of the child. Step 1a: Total Score The Total Score is an overall index of depression symptoms. A Total T-score of 65 or higher indicates that the child is experiencing an elevated number of depression symptoms and/or that the child is being perceived by parents and/or teachers as manifesting definite depression symptoms. An Chapter 4: Interpretation elevated Total Score must be considered in relation to the scales it comprises. A high Total 7-score could result from high scores in all areas of depression symptoms or from a very high score in a particular symptom area. When scale scores vary significantly (e.g., the Emotional Problems T-score is Very Elevated but the Functional Problem T-score is not), the Total Score should be de-emphasized and the interpretive focus should be on the scales/subscales with elevated T-scores. A T-score of 64 or lower is within the Average to High Average range. It is possible, however, to have a Total Score within this range that includes at least one scale score or subscale score that is elevated. Thus, even with an average 7-score, it is advisable to examine the components of the Total Score before deciding about the child's symptomatic status. Step 1b: Scale and Subscale² Scores As just noted, it is advisable to interpret elevated scale and subscale T-scores even if the Total T-score is within the Average to High Average range. A T-score of 65 or higher on either or both the Emotional Problems and Functional Problems scales indicates that the child may be experiencing an elevated number of symptoms in one of those specific areas. If only parent and teacher ratings are available, assessors should continue to examine the overall profile (see step 2) and the item-level responses (see step 3). Recall that the CDI 2:SR contains subscales in addition to the two scale scores (and the Total Score). If CDI 2:SR data are available, assessors should consider an elevated scale score in relation to the subscales it comprises. A high scale score could result from high scores on just one or both of the related subscales. For example, a child with an elevated score on the Emotional Problems scale may have very negative self-esteem but very few symptoms of negative mood or physical symptoms. Similarly, when the Emotional Problems and Functional Problems scale scores are in the not elevated range (i.e., T-score of 64 or lower), it is still important to examine the subscale scores. For example, a child with severe interpersonal problems might have an elevated score on the Interpersonal Problems subscale, but only an average Functional Problems score. Such a pattern can result if the child endorsed few if any issues pertaining to being inef-fective in various tasks. In this case, further investigation of the child's interpersonal difficulties is warranted. Table 4.2 notes some common characteristics of individuals who have T-scores of 65 or higher (the computerized reports offer similar tables). These text descriptions can be helpful in explaining to clients the meaning of elevated T-scores. Available only on the CDI 2:SR