Evidence-based assessment/Vignettes/ASEBA DSM
Comparing Dimensional Approaches (e.g., ASEBA) and Categorical (e.g., DSM)[edit | edit source]
One major way of approaching assessment has been to use a categorical system, with the Diagnostic and Statistical Manual (DSM) or International Classification of Diseases (ICD) being major examples. These hinge on deciding whether or not a person has a disorder. The DSM and ICD contain instructions and standard criteria for making a diagnosis. Because they are widely used in psychiatry and medicine, these classification systems are sometimes referred to as "the medical model," although classification is also used in other sciences (such as biology and geology) and applied disciplines.
An alternate method is to use rating scales and checklists to measure number and severity of symptoms. Adding the items up (or averaging them) creates a continuous, dimensional score. We can think of these as measuring the degree or severity of a problem.
The two methods offer fundamentally different ways of thinking about mental health and functioning. The categorical methods assume that people either have the condition or they don't, whereas dimensional methods focus on differences of degree. Is anxiety something that everyone experiences, just in different amounts at different times? Or are there two kinds of people -- those with an anxiety disorder, and those without? Categorical diagnoses and dimensional severity scores have different "levels" of measurement, matching to different forms of statistical analysis, as well.
On this page (and in several exercises) we focus on the Achenbach System of Empirically Based Assessment (ASEBA) as a foil for comparison with the DSM. The ASEBA is a family of checklists and rating scales authored by Thomas Achenbach and his team. The suite of measures includes self report, caregiver/parent report, and other collateral perspectives (teacher ratings for preschool and school aged youths, and "familiar other" for young adults and adults -- potentially including roommates and significant others). The ASEBA measures have been among the most widely used measures in clinical research, especially with youths. They have been updated several times over a three decade period, with translations into dozens or scores of languages. Two other features make the ASEBA scales quite different than the DSM.
First, the scales were built based on statistical analyses, not expert opinion or adherence to a particular diagnostic system. A few of the scales loosely approximate diagnostic symptom clusters, but there are some big differences. Anxious and depressed symptoms group onto a single scale, for example. Most ages have eight statistically guided scales, whereas the DSM has more than 360 different diagnoses.
Second, the scales have norms. The ASEBA uses separate reference groups for males and females, and also splits by age. The T-scores and the percentiles compare the individual's score on the scale to hundred of other people's responses of similar age and same sex. This provides an empirical way of telling how the amount of symptoms compares to what would be typical for a person's age and sex. If depressive symptoms are more common in adolescent women than men (and they are), then the same raw score would get assigned a lower T-score for women than men (because it is less high relative to the average for their age group). In contrast, the DSM uses the same number of symptoms to define a disorder for men or women; thus women reporting a higher average number of symptoms would lead to more women being diagnosed with depression. We see the same effect in reverse with diagnoses such as oppositional defiant disorder or ADHD, which are more commonly diagnosed in boys (who have higher average raw scores on the corresponding Aggressive Behavior and Attention Problems scales).
Which approach is better? They both have strengths and limitations. It is interesting to see how things play out when working with cases clinically, though.
ASEBA Scores for Cases[edit | edit source]
Here is a table summarizing the T-scores for several of our case examples. T-scores are standardized scores, with a score of 50 indicating an average amount of the symptoms for the person's age and sex, and a standard deviation (SD) of 10. Thus scores of 70 or higher are two SDs above average (~97.5th percentile -- although the scales with fewer items often have "lumpy" and "sparse" score patterns that do not follow the normal distribution perfectly). Scores of 70+ are often considered "clinically elevated."
Each column is a different case. The caregiver scores (based on the Child Behavior Checklist/CBCL) are presented first, followed by the Youth Self Report (YSR) and the Teacher Report Form (TRF). Note that Deshawn is too young to fill out the YSR; the norms start at age 11, because younger children often do not have a high enough reading level to fill out questionnaires.
DSM Diagnoses[edit | edit source]
Here is a summary of the diagnoses for each case, based on a semi-structured diagnostic interview. For most of these cases, the interviewers did not know the results of the checklists. That often was intentional -- if the case was participating in a project looking at the accuracy of the scales for predicting a diagnosis, then it is important that the diagnosis gets made without knowing the scale score (i.e., the score should be "masked" from the diagnosis). In many clinics, the ASEBA might not get typed into the computer for scoring until after the diagnostic interview finished. The person that did the intake might never see the scores, and the person providing treatment might get them weeks or months into treatment because of delays in processing. Not great design, but that is what often happens in the real world.
- PTSD (100% confidence)
- ADHD combined type (65% confidence)
- ODD (70% confidence)
- r/o reactive attachment disorder (49% confidence)
- r/o mood NOS (33% confidence)
- Cyclothymic d/o (65% confidence)
- Conduct d/o (90% confidence)
- ADHD NOS (65% confidence)
- Enuresis (100% confidence)
- r/o sexual abuse (20% confidence)
- Major Depressive Disorder (85% confidence)
- Dysthymic Disorder (85% confidence)
- ADHD Inattentive (75% confidence)
- ODD (70% confidence)
- Anxiety NOS (65% confidence)
- r/o panic without agoraphobia (45% confidence)
- Bipolar I (85% confidence)
- Specific phobia - animals (90% confidence)
- ADHD combined (70% confidence)
- r/o chronic tic (45% confidence)
- r/o LD NOS (30% confidence)
- Bipolar II (major depression + hypomanic episode)
- ADHD, Inattentive
- Past substance misuse (Xanax)
- Current alcohol and marijuana use
- Past non-suicidal self-injury (NSSI)
- Major Depressive Disorder (Single Episode, Moderate, with anxious distress)
- Provisional: ADHD, inattentive type (but uncertain if symptoms came before depression)
PTSD = Post-Traumatic Stress Disorder
ADHD = Attention-Deficit/Hyperactivity Disorder
ODD = Oppositional Defiant Disorder
r/o = “Rule out” – meaning that the diagnosis is still a possibility
NOS = DSM-IV “Not Otherwise Specified” – a residual diagnosis. In DSM-5, renamed “Other Specified X and Related Disorder”
LD = Learning disorder (cf. Learning Disability)
NSSI = Non Suicidal Self Injury
Discussion Questions[edit | edit source]
Integrating the CBCL and DSM Assessments
The goal of this assignment is to compare and contrast two major approaches to thinking about psychopathology.
One is the categorical approach, where people either meet criteria for a diagnosis, or they don't. This is the model used by the DSM, by psychiatry in general, and by insurance companies and Medicaid. It is often referred to as "the Medical Model."
The alternative is a dimensional approach. Instead of sorting people into groups (e.g., depressed vs. not-depressed), the model is that everyone has different amounts of the state or trait, and there is no sharp boundary between high and low scores. The dimensional model has been more common in psychology instead of psychiatry, and the Achenbach measures are one of the most well-known and widely used examples.
For each vignette, we have both the Achenbach scores and diagnoses based on a full-day semi-structured diagnostic interview with the family. Your job is to decide how best to use the information to help the patient.
You have the information about the scores and diagnoses on the handout for the homework. The scores and diagnoses are also available above.
Here is a link to form to enter your answers.
Here is a link to a downloadable copy of the questions.