Saint Marys University The Messiness of LD Identification Analysis Paper The actual writing: Please write a 4-5 page paper (see format details below) that

Saint Marys University The Messiness of LD Identification Analysis Paper The actual writing: Please write a 4-5 page paper (see format details below) that describes the main ideas of the article. Be sure you include a section that describes what you learned that you did not know, what surprised you, and how this article connects or relates to some aspect of your life or that of someone you know. You can use hypothetical situations (e.g., if I had no toes…).Format for assignment paper must be typed neatly and contain no misspelled words, typos, or grammatical errors. Assignments must be double-spaced (no extra spaces/lines between paragraphs) using 1″ margins, use New Times Roman font (12 cpi). Number the pages on the top right please. Start on Page 1 – no need for a title page. Be sure your name, date, and the title of the assignment are on your work. If your assignment is shorter than 4 entire pages, it is too short, and points will be deducted). I post the articles in pdf. Canadian Psychology / Psychologie canadienne
2017, Vol. 58, No. 3, 218 –227
© 2017 Canadian Psychological Association
0708-5591/17/$12.00 http://dx.doi.org/10.1037/cap0000115
The Messiness of LD Identification: Contributions of Diagnostic Criteria
and Clinical Judgment
Meadow Schroeder and Michelle A. Drefs
Damien C. Cormier
University of Calgary
University of Alberta
To make a diagnosis, clinicians rely on both diagnostic criteria and clinical reasoning. In the case of
learning disabilities (LDs), reliance on the latter (clinical reasoning) is likely to be heightened, given the
current absence of a consistent LD diagnostic approach or definition. This study investigated the
agreement between diagnostic decisions of reading LD (RLD) made by clinicians and those produced
from strict adherence to 3 commonly employed diagnostic models (i.e., low achievement, abilityachievement discrepancy, processing strengths and weakness). Using a clinical sample of 313
elementary-aged children, the level of agreement between clinician and models was determined using
kappa. The Wechsler Intelligence Scale for Children (4th ed.: Canadian; WISC-IVCDN; Wechsler, 2003)
and the Wechsler Individual Achievement Test (3rd ed.: Canadian; WIAT-IIICDN; Wechsler, 2010) were
used in the comparisons. Overall, the results indicate varying levels of agreement between: (a) clinician
and model identification of reading disabilities, and (b) model-model agreement. The strongest agreement was between the clinician and low achievement model (LAM)—the model with the broadest LD
definition. Clinicians should carefully consider the rationale for their diagnostic approach and be mindful
of additional factors influencing their clinical judgment.
Keywords: clinical reasoning, clinical decision-making, diagnostic agreement, learning disabilities
Further complicating this issue is that clinical decision-making
relies on more than just a strict adherence to a particular diagnostic
model with clinical judgment also playing an important role in the
process. Although clinical judgment generally remains a poorly
understood and largely neglected skill (Lilienfeld, Ammirati, &
David, 2012), we do know that it plays a more prominent role in
instances of diagnostic uncertainty (Redelmeier, Ferris, Tu, Hux,
& Schull, 2001). Thus, it is our contention that clinical reasoning
likely plays a heightened role in LD diagnosis. That is, psychologists are more likely to rely on clinical judgment to make diagnostic decisions in the absence of a consistent, reliable, and valid
identification approach—as describes the current state of LD identification. As an example, a psychologist who primarily adheres to
the discrepancy approach may be more lenient in the required
spread of scores knowing that other identification models exist that
do not require a discrepancy. Leniency might become attractive if
it assists in the obtainment of what is the psychologist deems as
“appropriate” levels of support for a particular student.
The use of clinical judgment is not without certain cautions. We
know largely from the considerable study of clinical reasoning
within the medical field that practitioners are prone to several
thinking errors and issues (Wilcox & Schroeder, 2015) and evidence within psychology has found practitioners inconsistently
apply diagnostic criteria (Maki, Burns, & Sullivan, 2017). However, it is unclear as to whether and under what conditions judgment leads to increased or decreased diagnostic accuracy.
Given the extant multiple approaches and the possible heightened role of clinical judgment in LD identification, the study of LD
identification practices renders a potentially useful avenue by
which to gain insights into the use of clinical judgment in assessment activities. In the present study, we directly compared the rate
At present, diagnosing an LD is a challenging task given the
lack of consensus regarding the best and more comprehensive
method for the diagnosis of LDs (Berninger, Richards, & Abbott,
2015). Traditionally, a formal diagnosis of an LD has been necessary for students to receive additional support and funding in the
academic classroom to address the areas in which they struggle
(Philpott & Cahill, 2008). Typically, psychologists or other individuals with formal training in cognitive and academic assessment
make this diagnosis by referencing established criteria for identifying LDs (Fagan & Wise, 2007). Of issue at present is that a
number of diagnostic models exist. For example, learning disorders were first included in the Diagnostic and Statistical Manual of
Mental Disorders (3rd ed.; American Psychiatric Association
[APA], 1980) and the model adopted in the description of learning
disorders is essentially that of the discrepancy model (DM). In
contrast, the concordance-discordance model (C-DM), based on a
pattern of strengths and weaknesses (PSW) approach (Hale &
Fiorello, 2004), has recently emerged as a viable alternative. This
availability of multiple approaches to LD identification has been
identified as highly problematic (Harrison & Holmes, 2012), with
research supporting LD identification as method contingent (i.e.,
who receives a diagnosis depends on the identification method
employed; e.g., Maki, Floyd, & Roberson, 2015).
Meadow Schroeder and Michelle A. Drefs, Educational Psychology,
University of Calgary; Damien C. Cormier, Department of Educational
Psychology, University of Alberta.
The authors would like to thank Amanda I. Radil.
Correspondence concerning this article should be addressed to Meadow
Schroeder, Educational Psychology, University of Calgary, 2500 University Drive NW, Calgary, AB T2N 1N4. E-mail: schroedm@ucalgary.ca
218
MESSINESS OF LD IDENTIFICATION
of LD diagnoses made by clinicians with those obtained using
three commonly employed psychometric approaches to LD identification—the DM, LAM, and C-DM. To further explore diagnostic decision-making, when there was disagreement between the
clinicians and psychometric approaches, we sought to determine
the source of the disagreement as an initial step to better understanding the types of information clinicians use to inform their
decision-making.
219
LD is reflected in the current diagnostic criteria for specific LD
described in the Diagnostic and Statistical Manual of Mental
Disorders (5th ed.; DSM–5; APA, 2013), which does not require a
discrepancy in cognitive and academic profiles. However, by
removing the emphasis on processing deficits or brain-based difficulties, the LAM significantly alters the definition of an LD that
is familiar to many professionals.
C-DM
Models of LD Diagnosis
Historically, LDs have been characterised by one feature: unexpected underachievement (Swanson, 2011). Unexpected underachievement represents the notion that LDs are fundamentally
about individuals of average cognitive abilities having significant
difficulty acquiring content in a specific academic area (i.e., language, reading, writing, mathematics). There is considerable controversy, however, as to how best to operationalize this definition.
Among the various models that have been proposed and utilized to
varying extents within the field are the DM, LAM, and C-DM.
DM
The DM for LD identification has seen the most widespread use
over other models for more than 30 years (Kavale, 2001). Essentially, the DM requires two pieces of information from the assessment process: an ability score (i.e., a standardized measure of
general cognitive functioning) and an achievement score (i.e., a
standardized measure of academic achievement in reading, writing, and/or math). When there is a significant difference between
the IQ and the achievement scores, defined as a standard score
difference of one to two standard deviations depending on the
criteria or sources used (Fletcher, Francis, Morris, & Lyon, 2005;
Shaywitz, Fletcher, Holahan, & Shaywitz, 1992), a LD is identified.
There is a substantial body of evidence suggesting that the DM
is not the most reliable or valid model of diagnosing LDs (see
Harrison & Holmes, 2012, for a review); nevertheless, it remains
the most widely used framework to govern how individuals receive additional support for learning difficulties (Fletcher, Francis,
et al., 2005). Because of the numerous studies that have highlighted problems with the DM, there has been a long history of
attempting to develop alternatives to the DM to identify LDs.
LAM
The LAM is a relatively straightforward method for identifying
LD; individuals who score below a given cut-off point are identified as having an LD (Swanson, 2011). This model was developed
from empirical evidence suggesting that individuals with low
achievement and individuals diagnosed with an LD based on the
DM tend to have similar cognitive or academic achievement
profiles (Shaywitz et al., 1992). Moreover, meta-analyses have
shown that low achievers and individuals diagnosed with LDs are
more similar in processing and cognitive profiles than they are
different (Stuebing et al., 2002). Swanson (2011) claimed that
there is “growing consensus” among professionals that it is more
appropriate to use “an absolute definition of learning disabilities,”
as opposed to a DM definition (p. 347). This shift in views toward
The C-DM of LD identification was originally proposed by Hale
and Fiorello (2004) to provide practitioners with an empirical
approach to examining patterns of performance on cognitive and
academic measures, in addition to information provided “as part of
a comprehensive evaluation that includes non-response to intervention and other data sources” (Hale, Wycoff, & Fiorello, 2010,
p. 190). The purpose of the framework is to identify a processing
deficit as related to a corresponding achievement deficit, while
both are not related to cognitive strengths (see Hale & Fiorello,
2004 or Hale et al., 2010 for a detailed description). Proponents of
the model assert the method can be used to plan targeted interventions based on the profile of cognitive strengths and weaknesses.
However, a number of criticisms have been levied against this
model including a lack of empirical evidence in support the validity of LD identifications, a low LD identification rate, and
limited published information on how to apply the model (e.g.,
Maki et al., 2015; Miciak, Taylor, Denton, & Fletcher, 2015).
Issues With the Multiple Approaches
to LD Identification
A body of literature has found that rates of LD identification are
dependent on the method used and who meets criteria outlined by
the specific method (Maki et al., 2015). Of issue is both the
variability across methods in prevalence rates and composition of
the subgroup (who has LD; e.g., Proctor & Prevatt, 2003). Similar
diagnostic consistency issues have been found with the use of a
single approach. Miciak and colleagues (2015) applied the PSW
approach to data from two comparable psychoeducational assessment batteries. The agreement in LD identification decisions between the two batteries was poor, with positive agreement just over
chance levels at 62%. Such consistency issues in LD identification
are not limited to methods that adhere to strict decision-making
rules for the interpretation of standardized test data. Barth et al.
(2008) found poor agreement among methods used in the response
to intervention (RtI) approach to LD identification. Different students were identified as inadequate responders depending on the
cut score, method (e.g., dual discrepancy), or measure used.
Clinical Decision-Making
Clinical reasoning is a skill that enables practitioners to “bridge
the gap between [their] experience and knowledge base and the
strategies used to solve real-life problems” (Andrews & Syeda,
2016, p. 2). Of the few empirical studies of clinical reasoning
specific to psychology, a recent study by Maki et al. (2017)
highlighted how clinicians may more widely and selectively consider additional information and use clinical judgment in formulating diagnostic decisions. They enlisted 376 school psychologists
SCHROEDER, DREFS, AND CORMIER
220
to review one of nine deidentified student evaluation reports and
make an LD identification (LD or not LD). Practitioners were
randomly assigned to use one of three diagnostic methods (i.e.,
RtI, PSW, or DM). In comparing LD identifications made by the
participants with those made by the researchers, the two groups
agreed on 73.7% of the cases. This agreement rate was further
reduced to 51.2% when the comparison was made using inconclusive case data— operationalized as student scores closely approximating established identification cut points. The results support
that practitioners, even when instructed to do so, do not always
strictly adhere to identification criteria, particularly so when they
encounter inconclusive data. Moreover, the extent to which clinical judgment was used did not appear to be impacted by practitioners’ years of experience, degree level, or training in a particular
identification approach.
In terms of what is known about influences on clinical judgment, there is some evidence to suggest that practitioners may
disregard more stringent diagnostic criteria in favour of loosely
defined criteria to provide services to a wider range of children in
need (MacMillan, Gresham, & Bocian, 1998). A 198% increase in
LD identification between 1976 and 1977 and 1992–1993 was
attributed, among other factors, to “a subjective decision-making
process involved in identifying and planning services for LD
students at the school level” (MacMillan & Siperstein, 2001, p. 3).
An additional factor that may make LD identification less clear
cut and require increased clinical judgment is the presence of
another disorder. Much comorbidity occurs with LDs, including
attention-deficit/hyperactivity disorder (ADHD; DuPaul, Gormley,
& Laracy, 2013), requiring differential diagnosis. There is a
chance that a child who exhibits symptoms of a reading disability
also exhibits symptoms of ADHD. The clinician relies on clinical
judgment to decide if the symptoms of ADHD are severe enough
to meet criteria for a diagnosis, and if the reading disability or the
ADHD is the primary concern.
In other situations, clinicians must decide if a disorder other than
LD better explains low achievement scores. Some individuals have
intelligence quotients that fall in the below-average range but are
not considered intellectually disabled (a range of 70 – 85). Clinicians are left to use clinical judgment to determine if the assessment data for an individual falling in the borderline intellectual
functioning (BIF) range suggests the presence of a true LD or if the
individual is a “slower learner.” In sum, deviation from a strict
adherence to diagnostic criteria may be appropriate based on
external or environmental variables (e.g., co-occurring disorders)
or based on experiential knowledge of the wider range of students
who benefit from additional intervention and supports. However,
such deviation may also result in LD misidentification (i.e., both
false positives and false negatives).
The Present Study
In order to begin to better understand how clinicians respond to the
absence of a single LD identification model, we compared the level of
agreement between diagnostic decisions for an RLD made by clinicians with those obtained by the research team using strict adherence
to three LD diagnostic models: (a) the DM, (b) the low achievement
model, and (c) the C-DM. Similar to the work of Maki et al. (2017)
which used deidentified reports of actual clients, the present study
used an archived data set. Among the benefits of using actual cases is
the ability to observe the outcomes of clinical decision-making when
practitioners are faced with “the polymorphisms of actual clinical
materials and the misleading test results encountered in everyday
practice” (Kassirer, 2010, p. 1118). To further explore diagnostic
decision-making, when there was disagreement between the clinician
and psychometric approaches, we sought to determine the source of
the disagreement as an initial step to better understanding the types of
information clinicians use to inform their decision-making. Two research questions were explored:
1.
What is the agreement between the three diagnostic models and clinician in the identification of a reading learning
disability?
2.
What factors account for the observed model-clinician
agreement and disagreement?
Method
Sample
Assessment data for children who had received psycho-educational
assessments between 2011 and 2015 from one of two clinics
(university-based, community-based) were selected through a homogenous sampling of the clinic databases. Cases were selected if: (a) the
client was age 6 –12 at the time of assessment, (b) the client had
completed the WISC-IVCDN and a WIAT-IIICDN (Reading Composite), and (c) a diagnostic decision was clearly stated. A total of 313
cases met our inclusion criteria. Slightly over half of the cases were
male (57% males, 43% females), and a learning difficulty in reading,
writing, and/or math was listed as either the primary (31%) or secondary (46%) referral concern. Other referral concerns included behavioural, social-emotional, giftedness, and autism spectrum disorder
(ASD). Table 1 provides additional demographic information about
the sample. Assessments were completed by graduate-level school
psychology students under the supervision of registered psychologists
(university-based clinic) or by provisional, or registered psychologists
(community-based clinic).
Measures
The test battery was selected to incorporate measures that assess
student performance in reading and general intelligence using
age-based norms. The Wechsler tests are two measures commonly
used by school psychologists in Canada (Oakland, Douglas, &
Kane, 2016). The WISC-IVCDN (Wechsler, 2003) is a normreferenced measure of intellectual ability. The subtests provide
four index scores and a Full Scale IQ (FSIQ). The internal consistency and test–retest reliability across the four index and FSIQ
scores for the WISC-IVCDN are generally high with values exceeding r ⫽ .90 (Wechsler, 2003).
The WIAT-IIICDN (Wechsler, 2010) is an individualadministered measure of language, reading, written language, and
mathematics. The Total Reading Composite (TRC) is derived from
tests of word recognition, pseudoword decoding, reading comprehension, and oral reading fluency. The composite reliability estimates for the WIAT-IIICDN exceed .90.
MESSINESS OF LD IDENTIFICATION
221
Table 1
Demographics of Sample
Clinic 1
(n ⫽ 68)
Category
Referral source
Agency
Parent/guardian
Medical
School
Other
Gender
Male
Female
Referral reasona
LD
Behaviour
Social-emotional
Giftedness
Autism spectrum disorder
Grade
1
2
3
4
5
6
7
Clinic 2
(n ⫽ 245)
n
%
n
%
0
46
6
16
0
0
68
9
24
0
9
169
57
9
1
4
69
23
4
.5
36
32
53
47
143
102
58
42
55
36
7
5
0
81
53
10
7
0
186
121
70
29
1
76
49
29
12
.1
2
9
19
14
11
9
4
3
13
28
21
16
13
6
12
56
53
59
27
32
6
Clinic 1
(n ⫽ 18)
Comorbid diagnoses with RLD
Math LD1
Writing LD
Behaviour-ADHD
Other behaviour
Social-emotional
Giftedness
Borderline intellectual
functioning/intellectual
disability
Autism spectrum disorder
Other
5
23
22
24
11
13
2
Clinic 2
(n ⫽ 64)
n
%
n
%
8
18
5
0
0
0
0
44
100
28
0
0
0
0
20
27
15
3
4
0
0
31
42
23
5
6
0
0
0
0
0
0
1
6
2
9
Note. N ⫽ 313. LD ⫽ learning disability; RLD ⫽ reading learning disability; ADHD ⫽ attention-deficit/
hyperactivity disorder.
a
Categories add up to greater than 100% because of multiple referral concerns/diagnoses for some cases.
Diagnostic Models
DM. Using the simple difference method (Saklofske, Prifitera,
Weiss, Rolfhus, & Zhu, 2005), cases met criteria for an RLD when
there was a 15 point (one standard deviation) or greater difference
between the WISC-IVCDN FSIQ and the WIAT-IIICDN TRC and
the FSIQ fell within the …
Purchase answer to see full
attachment