Keywords
communicative participation - patient-reported outcomes - quality of life
Learning Outcomes: As a result of this activity, the reader will be able to (1) describe the importance
of including stakeholder perspectives such as through cognitive interviews in the
development of patient-reported outcomes; (2) describe the recommended methods for
scoring and interpreting the CPIB regarding use of T-scores; (3) list at least three
populations of people with communication disorders for whom the CPIB has been validated;
(4) discuss why measurement of communicative participation is important in clinical
assessment and outcomes measurement in addition to conducting assessment of communication
impairment/activity; and (5) describe the content and format of the CPIB.
Patient-reported outcomes (PROs) are becoming increasingly recognized as an essential
component of assessment and outcomes measurement in speech-language pathology (SLP)
clinical practice.[1]
[2] PROs ensure that the viewpoints of clients are represented in the identification
of treatment needs, the selection and prioritization of treatment goals, and the determination
of whether meaningful outcomes have been achieved. Person-centered care in SLP—care
that is defined by and focused on the needs, values, priorities, and cultures of an
individual client—compels us to capture each client's perspective on the communication
disorder, and to document that our interventions have led to changes that they define
as worthwhile.[3]
[4]
[5] PROs can bring measurement rigor to the client values and perspectives component
of evidence-based practice.[1] A variety of PROs relevant to various communication disorders in adults are available
including those pertaining to neurologic communication disorders,[6]
[7]
[8] voice disorders,[9]
[10]
[11] and fluency.[12]
[13]
[14] The purpose of this article is to summarize the Communicative Participation Item
Bank (CPIB)[15] as one example of a PRO that was designed to be applicable for adults across a range
of communication disorders. An instrument that can be used across diagnoses could
facilitate comparisons across different disorder groups,[16] as well as promote aggregating data across multiple groups for purposes such as
obtaining larger samples and performing meta-analyses, identifying common participation
restrictions for advocacy purposes, and providing a common metric for clinics to report
outcomes across their clinical units. This article will focus on the extent to which
the CPIB has achieved the objective of being relevant, valid, and reliable across
disorders, as well as looking at the next steps that are needed to fully realize that
goal.
Setting the Stage for the CPIB
Setting the Stage for the CPIB
Development of the CPIB began in 2004 at a time when the National Institutes of Health
(NIH) made a significant investment in improving the science of PROs through an initiative
referred to as the Patient Reported Outcomes Measurement Information System (PROMIS;
https://www.healthmeasures.net/explore-measurement-systems/promis).[17]
[18] Researchers working under the umbrella of PROMIS harnessed several methods to advance
the scientific rigor of PROs including increased reliance on systematic guidance from
stakeholders through cognitive interviews,[19]
[20]
[21] and use of the modern measurement methods of Item Response Theory (IRT).[22]
[23] While the CPIB is not part of the PROMIS system, its development was patterned after
PROMIS methods. Before exploring the use of the CPIB in different groups of individuals
with communication disorders, the following sections will provide an overview of the
CPIB including defining the concept of communicative participation, discussing the
role of stakeholder input, and introducing basic concepts of IRT most relevant to
clinical use of the CPIB and other IRT-based instruments.
What is the Construct of Communicative Participation?
The construct of communicative participation was adapted from the concept of life
participation in the World Health Organization's (WHO) International Classification
of Functioning, Disability, and Health (ICF).[24] The ICF defines participation as involvement in life situations. Communicative participation
refers specifically to the communication aspects of life situations, and encompasses
the exchange of information and ideas among people in the settings in which people
live, work, socialize, and engage in their communities.[25] To assess and support communicative participation, it is important to think about
the full experience of engagement in communication in a situation, and to appreciate
how that involvement is shaped by a constellation of influences. As SLPs, we have
often focused largely on communication proficiency such as speech intelligibility,
accuracy of reading or auditory comprehension, speech fluency, voice quality, and
other measures of communication skill or ability.[26]
[27]
[28]
[29] Communicative participation, however, encompasses a more holistic view of the communication
experience than accuracy of communication skills alone.[30] Successful or satisfactory communicative participation requires a level of engagement
or involvement that meets each individual's needs or goals, and includes connectedness
to others, integration into the communication situation, or other comprehensive views
of the communication encounter.[31]
[32]
This distinction between communicative participation and communication impairment
is clarified by examining the role of these constructs in the context of the ICF framework.[24] The ICF, as a biopsychosocial model of health and disability, reflects the many
interactions among the physical structure and function of the body (impairment), activity
performance (activity limitations), environmental and personal contextual factors,
and participation in life situations (participation restrictions). Given the multiple,
multidirectional interactions among these elements, communicative participation can
be shaped by any combination of the other factors in the ICF framework. While SLPs
are highly familiar with the ways in which physiologic damage to or atypical development
of the physiologic mechanisms for speech, language, cognition, and hearing can impact
an individual's ability to communicate, we also recognize the profound barriers that
the social and physical environments around us can create for people with communication
disorders. Furthermore, we recognize that each client brings their own individual
background and perspectives to the experience of living with a communication disorder,
which can certainly influence priorities for treatment goals, intervention methods,
and outcomes. Thus, the CPIB was designed to provide an overall patient-reported view
of participation in conversational situations, with the intent that other measures
of communication impairment and skill performance, environmental factors, and personal
factors can then be compared to ascertain how each contributes to CPIB outcomes.
Why is Stakeholder input important in PRO development?
PROs are designed to assess a construct from the viewpoint of the person living with
that experience, and as such, the process of developing PROs should include these
stakeholders. Just because a PRO asks patients to respond to the questions does not
guarantee that those questions ask about experiences or issues that are the most meaningful
or relevant to the people living with the diagnosis or situation.[33] PROs developed solely by clinicians or researchers without the input of people who
live with the condition risk being irrelevant, or even offensive to clients if they
do not address meaningful experiences using language and terminology that is regarded
as acceptable by that population.
Stakeholder input in the development of PROs is often sought through cognitive interviews.[19]
[21] In cognitive interviews, the candidate items are presented to people who live with
the diagnosis (or other relevant experience) who are asked to “think out loud” as
they complete the questionnaires. These participants provide detailed feedback regarding
how they answer the questions and why, wording that they find confusing or offensive,
the relevance of the content in the items, and all other aspects of a questionnaire
including the instructions, response options, format, and any other elements. Instrument
developers work collaboratively with these participants to revise the questionnaire.
This rigorous process helps ensure that the final version of the questionnaire is
acceptable, accessible, and meaningful to the stakeholders.
The input of people living with various communication disorders was incorporated into
the development of the CPIB. Cognitive interviews were conducted early in the development
of the initial item set, and included people with communication disorders associated
with laryngeal dystonia (LD; spasmodic dysphonia), Parkinson's disease (PD), laryngectomy,
multiple sclerosis (MS), amyotrophic lateral sclerosis (ALS), stroke, and stuttering.[31]
[34] Cognitive interviews have been repeated for the CPIB as part of calibrating the
instrument for new populations that may not be sufficiently represented by the speech,
language, or voice disorders included in the original samples. For example, cognitive
interviews conducted with people with hearing loss,[35] and with people who are transgender (manuscript in preparation), identified issues
that warrant modifications in the CPIB which will be discussed later in this article.
What is item response theory?
After candidate items for a PRO have been vetted through cognitive interviews, the
next step in instrument development is psychometric assessment and further modifications
of the item set to optimize validity and reliability. There are two main branches
of psychometric methods used in instrument development—Classical Test Theory (CTT)
and IRT.[36]
[37]
[38] Both CTT and IRT grapple with the issue that many of the constructs that are assessed
in tests and questionnaires such as reading comprehension, quality of life (QOL),
or communicative participation are latent traits. Latent traits are constructs that
we cannot observe and measure directly such as height or weight. They are more obscure
or underlying traits that we can only estimate based on how individuals respond to
test or questionnaire items. For example, we infer what an individual's QOL might
be from opinions they express on a questionnaire. CTT is likely the set of psychometric
methods that most readers are familiar with even if they have not heard that term
before. CTT is based on assumptions that any observed score on a questionnaire reflects
a combination of the respondent's “true” score—or actual level of the latent trait,
as well as some degree of error in measurement that leads to the observed score differing,
to some unknown extent, from the true score. If measures of validity and reliability
are sufficiently favorable, we assume that the observed score sufficiently represents
the true score.
IRT, sometimes referred to as modern measurement theory, provides estimates of the
latent trait by using mathematical modeling of the relationships among the latent
trait, the parameters of items on the instrument, and how people respond to the items.[18]
[36]
[39]
[40]
[41] While a detailed discussion of IRT is beyond the scope of this article, one analysis,
that of differential item function (DIF), is particularly relevant to understanding
CPIB development. DIF is essentially a measure of bias in an instrument. Absence of
DIF suggests an absence of bias, meaning that the measurement properties of the items
(item difficulty and/or item discrimination) do not differ across subgroups such as
age, sex, or other participant subgroups. Absence of DIF does not mean that these
different subgroups experience the same level of the latent trait (e.g., the same
extent of communicative participation restrictions); it means that if subgroups are
statistically equated for their trait levels, they will have the same predicted responses
to the items. Thus, the same items and scoring parameters can be used across subgroups
without measurement confounds stemming from the instrument. Examples of DIF analyses
for the CPIB will be discussed later in this article demonstrating the extent to which
the CPIB can be used with people with different communication disorders. Two other
IRT concepts that SLPs should be particularly aware of when using any IRT-based instrument
are adaptive assessment and standard scale scoring which will be summarized in the
next paragraphs. Readers interested in more detail about applications of IRT to SLP
practice are referred elsewhere.[1]
[40]
IRT instruments are often referred to as item banks. An IRT item bank is a reservoir
of items from which subsets of items can be selected for an assessment purpose. The
items in an IRT item bank usually represent different levels of item difficulty, or
levels of the trait, and thus some might be better suited for some measurement purposes
than others. For example, in a measure of communicative participation, some items
might be more sensitive for capturing the higher level of difficulty of participating
in more demanding or complex communication situations, while other items might be
more sensitive to capturing the lower levels of difficulty of less-demanding interactions.
In contrast to CTT-based instruments (many traditional questionnaires in SLP practice),
which require that all items in a questionnaire be administered to obtain valid scores
(regardless of how sensitive each item is for assessing that particular individual),
IRT allows for adaptive assessment in which only the items in the item bank that are
most sensitive for estimating a particular individual's trait level need to be administered.
In our example of the CPIB and communicative participation, items that represent the
less difficult end of the range of the underlying trait depict situations that are
typically easier to engage in such as talking with people you know, or greeting familiar
people at a social gathering. Items that represent the more difficult end of the trait
range depict situations such as getting your turn in a fast-moving conversation, or
communicating in large groups. On the CPIB, someone whose underlying “trait” would
be best represented by a score in the low-medium range of the scale should not have
to answer items on the more difficult end of the scale, such as those about speaking
in large groups, because those items will not be sensitive to estimating their trait.
Adaptive assessment is most efficiently conducted via computerized adaptive testing
(CAT) in which a computer algorithm selects items to administer to an individual based
on how they have responded to earlier items.[39]
[42]
[43] CAT administrations can obtain scores that are as precise, or more precise than
CTT scores using far fewer items—sometimes as few as four or five items.[43] When CAT applications are not available in clinical and research settings, IRT-based
instruments may offer static, paper-and-pencil short forms that consist of a small
set of items extracted from the full item bank and calibrated with a scoring guide.
Multiple different short forms can be developed from one item bank to target different
levels of the latent trait or other assessment purposes, and each is calibrated with
its own scoring guide. Thus, with IRT item banks, the “full-item set” may never need
to be administered. There may only be CAT applications and short forms, which have
been demonstrated to represent the measurement qualities available in the full item
bank. Adaptive assessment has the potential to help PROs to be more acceptable for
respondents by not asking them to respond to items that may be highly irrelevant to
them because the situations asked about are much too difficult, or too easy. In addition,
the reduction in response burden facilitated by adaptive assessment might significantly
improve feasibility and accessibility of PROs.
This discussion of adaptive testing and multiple short forms raises another key issue
in using IRT-based instruments: how to calculate and compare scores generated using
different item sets. With IRT instruments, raw scores are generally not appropriate
for reporting and interpreting scores. Imagine having two short forms generated from
the same item bank—one consisting of items asking about easier communication situations
(e.g., short form A) and another with items about more difficult situations (e.g.,
short form B). Let's pretend that we administer one short form (e.g., short form A)
to one person, and the other short form (e.g., short form B) to a different person,
and they both get a raw score of 20 on their respective short forms. Because the items
in these two short forms are different, and represent different underlying levels
of the trait, we cannot say that these two people have the same level of the trait
even though they have the same raw scores on their short forms. The only way to interpret
and compare these scores is to convert them to a standard scale to which all items
in the item bank have been calibrated in a way that accounts for the different difficulty
levels, item discrimination abilities, and other measurement parameters of the items.
Many IRT-based instruments use the T-scale as the standard scale. The mean of the
T-scale is 50 and represents the mean of the calibration sample (standard deviation = 10).
The T-scores obtained from different CAT administrations or short forms from the same
instrument can be directly compared with each other.
The Format of CPIB (Version 1.0)
Currently, the CPIB is a 46-item bank with the option for CAT and one static, 10-item
general short form.[15] All items start with the stem, “Does your condition interfere with…” followed by
a conversational situation such as “making small talk,” or “getting your turn in a
fast-moving conversation.” All of the items address verbal conversational situations.
Items related to communication via reading, writing, or other modalities are not included
for two reasons. First, for measurement purposes, the focus on one modality of communication
helps create a unidimensional instrument in which all items measure the same construct
(i.e., interference in communicative participation).[44] This avoids potential measurement confounds from mixing different communication
modalities into one instrument. In addition, focusing on conversational situations
keeps the CPIB relevant to most communication disorders in adults because while most
communication disorders impact conversation in some manner, not all communication
disorders impact reading and writing.
The CPIB item stem refers to “condition” as opposed to a more specific term such as
“communication disorder.”[31]
[34] This choice was made based on guidance from many cognitive interview participants
who had either multiple communication disorders (e.g., a speech disorder and hearing
impairment) or a communication disorder occurring in the context of a more complex
medical condition such as PD or MS. These participants reported that they could not
reliably parse out how much of their communicative participation restrictions might
be due to one aspect of communication versus another (the speech disorder vs. the
hearing impairment). Participants with complex diagnoses also reported that many aspects
of their health conditions such as fatigue, tremor, vision loss, or other symptoms
contributed to restricted communicative participation. They indicated that it was
an artificial and invalid exercise to try to report how much their communicative participation
was restricted by a single symptom when their lived experiences were more complex.
Another term in the CPIB stem that warrants mention is the term “interfere.” This
also reflects the consistent feedback provided by cognitive interview participants
who wanted wording that captured the problems, challenges, and heartache that had
been associated with their communication disorders.[31]
[34] While most people reported restrictions in communicative participation, these restrictions
took different forms. Some people avoided or felt excluded from situations and simply
did not participate. Other people still participated, but did so with considerable
effort, discomfort, or frustration. The term “interfere” captured, for most participants,
the various ways that their participation might be restricted.
Currently, the CPIB can be administered via either a static paper-and-pencil 10-item
general short form that represents the range of the item difficulty levels available
in the item bank, or via a CAT application (
https://sites.google.com/uw.edu/cpib
).[15] Thus, there is no “full length” form for the full 46-item set. If using the static
short form, scoring and interpretation involves a two-step process. First, the user
must calculate the raw or summary score using the point values at the top of each
response column corresponding to the extent of interference reported for each item
(not at all = 3 points; a little = 2, quite a bit = 1; very much = 0; this system
allows for higher scores to indicate more favorable participation). Then, the user
should refer to the scoring table on the second page of the short form to convert
the raw scores into T-scores. A T-score of 50 on the CPIB is the mean of the calibration
sample which was a large sample representing different communication disorders including
those associated with both neurologic etiologies and head and neck cancer (HNCA).[15] Thus, a T-score of 50 means that the respondent is reporting about the same level
of restrictions in communicative participation as people with other communication
disorders report on average. Higher scores are more favorable (i.e., they represent
better communicative participation, or less interference). Perhaps a more meaningful
way to interpret the scores, particularly in a clinical situation, is to compare a
client's score to the maximum possible score which is what a client would achieve
if they reported no interference on every item they were administered—an ideal treatment
outcome. On the general short form, the maximal possible score is T = 71.0, and on
the CAT application the maximal possible score is T = 74.3. Data regarding what constitutes
clinically meaningful differences are forthcoming, but one general guide that has
been found to be consistent among PRO assessments in other healthcare fields is that
a change of half a standard deviation (5 points on the T-scale) is a common indicator
of clinically significant change.[45]
Using the CPIB with various populations
Using the CPIB with various populations
The following sections describe the different populations for which the CPIB has been
calibrated, and how the CPIB has been used in research or clinical practice with these
groups.
Neurologic Speech and Language Disorders
Currently, the CPIB has been calibrated and used with adults representing a range
of neurologic motor speech and language impairments. Further work remains to be done,
particularly in the areas of cognitive-communication disorders and for more severe
impairments requiring use of augmentative and alternative communication (AAC).
Parkinson's Disease
People with PD were included in the cognitive interviews and the initial across-disorder
calibration of the CPIB.[15]
[31] For people with PD, the CPIB has been moderately correlated with measures of health-related
QOL (both general and PD-specific questionnaires), suggesting that it captures a construct
that has some overlap with, but is distinct from QOL.[46] An analysis of predictors of CPIB scores in people with PD in both the United States
and New Zealand found that better communicative participation was associated with
less severe self-reported symptoms in several areas including speech symptoms, cognitive
difficulties, emotional symptoms, fatigue, and dysphagia. More favorable participation
was also associated with higher levels of speech usage, but with an interaction with
age.[46] Relationships between CPIB scores and clinician-administered measures of PD severity
may be more nuanced. For example, Spencer et al[47] found that clinician-administered measures of motor severity and a cognitive screening
did not significantly predict CPIB scores,[47] while Barnish et al[48] did find that scores on a cognitive screening did predict CPIB scores.[48] Both of these studies demonstrated that the ability of speech intelligibility measures
to predict CPIB scores depended on the task. Intelligibility measures based on isolated
read sentence production did not significantly predict CPIB scores, whereas intelligibility
measures based on more naturalistic, contextually produced sentences or monologue
samples did.[47]
[48] These findings suggest that communication proficiency measures that replicate naturalistic
communication may be stronger predictors of communicative participation.
As an outcome measure, the CPIB captured statistically significant changes in communicative
participation after Lee Silverman Voice Training (LSVT) for people with PD with large
effect sizes.[49] However, after LSVT, participants still reported scores approximately 1.5 standard
deviations below the maximum possible score, suggesting that while improved, they
experienced ongoing participation restrictions. These findings support qualitative
and quantitative reports from people with PD regarding the complexity of communication
experiences, the success of which hinges on more than the extent or nature of dysarthria
symptoms.[32]
[50]
[51]
[52] SLPs may need to take a broader approach to helping people with PD with their communicative
participation in ways that include, but also extend beyond some of the current impairment-focused
speech training programs.
Multiple Sclerosis
People with MS were also included in the initial CPIB cognitive interviews and across-disorder
calibration.[15]
[31] Similar to findings in people with PD, more favorable participation has been associated
with multiple self-reported factors including less severe speech symptoms, fatigue,
cognitive symptoms, limitations in physical abilities, and depression. Better participation
has also been associated with being employed, as well as with higher levels of perceived
social support, education, and speech usage.[53]
[54]
[55] The CPIB demonstrated moderate-high correlations with a PRO specifically developed
to address communication symptoms in people with MS.[56] There is evidence of a significant trend in worsening CPIB scores, as clinician-judged
cognitive-communication profiles became more complex in that people without cognitive
or dysarthria symptoms reported most favorable communicative participation, followed
by people with either cognitive or dysarthria symptoms alone, and then people with
combined cognitive and dysarthria symptoms who reported the most restricted participation.[57] In contrast, CPIB scores have not been significantly associated with clinical measures
of sentence intelligibility or other clinician-judged speech severity measures.[57] As with PD, these findings suggest that increasing complexity and/or severity of
communication symptoms likely contribute to participation restrictions. However, the
results also suggest that many of our clinical measures that assess isolated communication
skills and tasks may not sufficiently capture those relationships.[57] These findings also support qualitative data that people with MS may be struggling
with communicative participation long before any speech symptoms are noticeable to
the casual observer.[31]
[58]
[59] Symptoms such as mild cognitive or word-finding challenges, as well as fatigue and
unpredictability of symptoms, may lead people with MS to withdraw from or avoid interactions
with other people even though they may be able to mask their communication struggles
sufficiently that other people do not notice the symptoms. People with MS may benefit
from SLP services to help them maintain, or bolster, their communicative participation
much earlier in the course of their disease than they may currently be referred for
frank evidence of overt dysarthria or cognitive impairments.
Aphasia (with/without Apraxia of Speech)
After the initial CPIB calibration, a later study demonstrated no clinically significant
DIF on the CPIB between people with aphasia and the original calibration sample. This
means that the original item set, General Short Form, CAT, and scoring parameters
can be used with people with aphasia.[60] Comparison of levels of support needed from SLPs to complete the CPIB to scores
from the Western Aphasia Battery-Revised (WAB-R)[61] suggested that people with WAB-R aphasia quotient scores above 80 required minimal
assistance. Participants with aphasia quotient scores ranging from 50 to 80 were generally
able to complete the CPIB, although with increasing reliance on support as aphasia
severity worsened. Participants with WAB-R scores lower than 50 were generally unable
to complete the CPIB or did so with considerable difficulty.[60] The CPIB was moderately correlated with the ASHA Quality of Communication Life (ASHA-QOL)
scale, again showing that the CPIB represents a related but not identical construct
to QOL.[60] Correlation between self-report and family proxy report on the CPIB was low with
proxy scores being over 5 points lower on the T-scale (over half a standard deviation)
than the self-report scores, a difference that was statistically significant.[60]
In a study with people with primary progressive apraxia of speech (AOS) with and without
co-occurring aphasia, CPIB scores revealed restricted communicative participation
in all subgroups, but more so in people with aphasia co-occurring with AOS than with
AOS alone (or AOS with dysarthria).[62] Findings supported other trends in the research that CPIB scores do not always correlate
strongly with measures of speech or language abilities.[62] Early studies on changes in CPIB scores after treatment have been mixed with one
study showing significant gains after treatment for aphasia,[63] but another showing mixed results after treatment for AOS.[64]
IRT instruments with adaptive testing such as CAT may be of particular benefit for
people with aphasia or other cognitive or language impairments, for whom the demands
of reading and responding to questions may be particularly fatiguing. CAT applications
that can get highly precise and reliable measurement with just a small number of items
may facilitate greater involvement of clients and research participants in PRO measurement.
This approach may reduce the historical problem of people with cognitive or linguistic
impairments being excluded from research because of difficulties with self-report.[65]
[66]
[67] We should also keep in mind that completing PROs is not an assessment of reading comprehension. Clients can be assisted in completing PROs,
including having items read to them, as long as they understand the intent of the
items and can provide their own responses in some reliable manner.[68]
Degenerative Diseases
People with ALS were included in the cognitive interviews and initial calibration
of the CPIB.[15]
[31] Another study with people with ALS showed a wide range of CPIB scores, with communicative
participation steadily worsening with increasing dysarthria severity.[69] CPIB scores correlated strongly with other bulbar symptoms but only moderately with
speech intelligibility scores.[69] The CPIB has also undergone assessment for use with people with Huntington's disease,[70] and has been used with individuals with progressive ataxic dysarthria.[71]
One notable limitation of the early work with people with ALS is that all participants
were still using natural speech for at least some of their communication.[15]
[69] The CPIB has yet to be calibrated for people who communicate primarily or solely
with AAC. Validating the CPIB for AAC communicators needs to begin with cognitive
interviews to assess the acceptability of many of the items which refer to “speaking.”
This qualitative work should be followed by the psychometric analyses needed to either
demonstrate no DIF with prior calibration samples, thus allowing use of the current
item set and scoring parameters, or advising on the modification of the CPIB for AAC
communicators.
Head and Neck Cancer
People with HNCA, including people with laryngectomy, were included in the cognitive
interviews and initial CPIB calibration.[15]
[31] Many of the trends with HNCA are similar to those with other disorder groups. For
example, correlations between the CPIB and both general health-related and diagnosis-specific
QOL instruments in HNCA vary from low to moderately high.[72]
[73] Those relationships strengthen to moderate-strong correlations with instruments
that focus on voice-specific QOL such as the Voice Handicap Index (VHI), particularly
in studies that have included a large representation of individuals treated for laryngeal
cancer.[72]
[74]
As with other diagnoses, clinical measures of speech severity and speech intelligibility
are only weakly associated with communicative participation in HNCA.[75]
[76] However, also consistent with other populations, the severity of self-reported speech symptoms is the variable most strongly associated with CPIB scores in HNCA
(i.e., increased self-rated severity of speech is associated with worse communicative
participation). Other self-reported variables including cognitive symptoms, depression,
perceived social support, and resilience are also significantly associated with communicative
participation.[73]
[77]
[78] These findings support the premise that communicative participation is affected
by multiple non–speech impairments as well as environmental and personal factors in
HNCA survivors, as well as in other populations.
One finding specifically relevant to people with HNCA is that CPIB scores may vary
with different cancer sites. For example, scores have been shown to be worse for people
with oral cancer compared with those with oropharyngeal cancer.[73]
[78] When combined with other studies, it appears that regardless of cancer site, many
patients with different types of HNCA may be at high risk for not returning to their
baseline function in communicative participation, even many years after their HNCA
treatment. Future studies using longitudinal designs are forthcoming.
Voice Disorders
People with LD (also known as spasmodic dysphonia) have been represented at many stages
in the development of the CPIB, including the earliest cognitive interviews,[34] the first preliminary psychometric IRT analyses,[79] and a later additional item calibration finding no DIF with other populations included
in CPIB development.[80] Correlations between the CPIB and VHI for people with LD have been shown to be moderate-high.[79]
[80]
A comparison of CPIB scores with qualitative patient accounts revealed that even for
people who have been established in treatment with botulinum toxin (BoNT) injections,
the most common treatment for LD, there is a wide range of persistent communicative
participation restrictions. Some people receiving long-term BoNT injections experience
negligible restrictions in communicative participation, while others still experience
severe restrictions. Mixed results have been found regarding changes in CPIB scores
after BoNT treatment, with both statistically significant[80] and nonsignificant[81] findings reported. Two factors in the latter study may have contributed to lack
of statistical significance.[81] First, most participants appeared to be experienced with BoNT injections, and it
is possible that people who know from experience how to manage the ups and downs of
the BoNT treatment cycle may not experience wide swings in PRO perspectives in any
single treatment cycle. Second, the post-injection data collection point appeared
to be within 2 weeks after the injection, and some participants may have not yet reached
their optional voice quality at that time. These studies illustrate the need to consider
issues unique to cyclical treatment such as BoNT when assessing communicative participation
or other patient-reported constructs in people with LD.
One of the issues that has become apparent in exploring communicative participation
in people with LD, similar to PD and other populations described earlier, is that
there is often a gap left between the communicative participation gains that are made
with traditional interventions and optimal communicative participation.[34] In LD, residual dysphonia, speaking effort, fluctuations in symptoms over the treatment
cycle, unpredictability of treatment response, and unaccommodating communication environments
are just some factors that continue to exert restrictions on communicative participation,
even when people have been receiving BoNT treatment for years.[82] SLPs should play a critical role in helping people with LD maximize their communicative
participation even when they are established in BoNT treatment.
Use of the CPIB with other voice populations is gradually growing. Significant improvements
in CPIB scores have been observed after intervention for unilateral vocal fold immobility.
The CPIB has also been shown to be sensitive to change in people with presbyphonia
undergoing SLP intervention.[83] A cross-sectional study of a community-based sample representing diverse voice disorder
diagnoses supported trends found in other populations that the CPIB demonstrates moderate
correlations with other PROs, such as the VHI, indicating that the CPIB addresses
a construct related to, but sufficiently distinct from other PROs to warrant its use
in clinical assessment and outcomes measurement.[84]
Fluency
People who stutter were included in the qualitative cognitive interviews as part of
developing the CPIB.[31] In terms of quantitative explorations of communicative participation restrictions
and contributing factors, the patterns seen in the populations described thus far
appear to continue with people who stutter. Strong correlations have been observed
between the CPIB and self-reported speech severity. Moderate correlations were observed
with self-esteem, self-efficacy, and social support; and significant but small correlations
were observed with age, education, speech usage, and stuttering self-help/support
group history.[85]
Additional Populations
Use of the CPIB has begun to spread to additional populations which may provide insight
into the impact of a wide range of health conditions on communicative participation,
even if speech, language, cognitive, or hearing impairments may not be the most salient
aspects of those diagnoses. Restrictions in communicative participation were captured
by the CPIB for people with spinal cord injury (SCI),[86] facioscapulohumeral muscular dystrophy (FSHD),[87] and facial paralysis.[88]
[89] These studies revealed a range of communicative participation restrictions with,
in some cases, the extent of the restrictions being commensurate with those resulting
from more traditionally identified speech or language impairments. One finding of
interest is that across these groups, conversational situations that require some
element of speed or endurance such as having long conversations, getting your turn
in fast-moving conversations, and having to say something quickly are situations that
these participants found particularly challenging.[86]
[87]
Cross-Cultural Investigations and Translations
Two studies have investigated the psychometric properties of the CPIB in other English-speaking
countries including New Zealand[90] and the United Kingdom[48] with favorable results. At the time of this writing, the CPIB is being used in Swedish.[69] Efforts are underway for translations of the CPIB into other languages as well,
with results pending.
Future Directions
The evidence summarized in this article suggests that the CPIB captures the construct
of communicative participation in a manner that is meaningful, relevant, valid, and
reliable for people across a range of communication disorders and other health conditions.
Combined, the evidence to date suggests that the CPIB measures a construct that is
not uniquely captured by other PROs that focus either more specifically on speech
or language symptoms, or more broadly on overall QOL. Furthermore, CPIB scores are
not fully accounted for by clinical measures of speech or language skill or ability.
Despite these positive indicators that the CPIB provides a useful tool in the measurement
toolbox, the relevance and acceptability of the original CPIB version are not universal.
Recent and ongoing research has suggested needed modification of the CPIB for some
populations.
One population for whom the CPIB may warrant modification is people with hearing loss.
Cognitive interviews suggested that 12 of the 46 CPIB items were not optimally relevant
and acceptable to people with hearing loss.[35] The primary reason that these items were unacceptable is that they were written
in a way that focused almost entirely on expressive communication (e.g., giving someone
detailed information) instead of on the back-and-forth exchange of communication.
Items that referred more holistically to conversations that encompassed both receptive
and expressive communication were regarded as highly relevant. Given these findings,
one option is to create another short form for the CPIB focused on items with the
highest relevance to that population for a hearing-loss–specific short form.
One group for whom the original CPIB version is unacceptable is people who are transgender.
Based on cognitive interview findings, to be reported in a later article, SLPs and
healthcare providers are strongly advised to not use the original CPIB with this population because wording of the item stem is not
acceptable to many people who are transgender. At the time of this writing, work is
underway calibrating a new version of the CPIB that uses more inclusive wording, and
readers are encouraged to contact the corresponding author for more details if further
information is desired before results are published. The new CPIB version that is
undergoing psychometric analyses may also be more inclusive for other populations
who would not be considered as having communication disorders such as people seeking
accent modification, or even for typically communicating adults who may experience
other factors influencing communicative participation. The addition of a CPIB “version
2.0” will not render the original version invalid. Clinicians and researchers who find the original
version suitable for their needs and want a version with established validity and
reliability data behind the CPIB for the populations for which these data exist may
choose to continue to use the original version.
Other areas need to be addressed to achieve truly widespread applicability of the
CPIB. As mentioned earlier, studies of validity and reliability have not been conducted
for some groups including those with cognitive-communication disorders resulting from
various etiologies, and people who communicate primarily or solely using AAC. Finally,
further exploration into translations and applicability in different languages and
cultures is needed.
In conclusion, in this article we have attempted to summarize the development of the
CPIB and its use in clinical and research applications up to the time of writing this
article. While we have attempted to be as inclusive as possible in capturing the existing
peer-reviewed literature relevant to the CPIB, we acknowledge the possibility that
some articles may be missed in this review, particularly if they reach publication
status between completion of writing and publication of this article. We hope that
researchers and clinicians will find the CPIB useful in promoting holistic, patient-centered
communication care, and we appreciate the efforts of clinicians and researchers, as
well as those of people with communication disorders and their families and colleagues,
to promote communication access and equity for all.
