Predicting Communicative Participation in Adults Across Communication Disorders

Abstract

Purpose

The purpose of this study was to explore the extent to which communicative participation differs across diagnoses and if there are common predictor variables for communicative participation across diagnoses.

Method

Survey data on self-report variables including communicative participation were collected from 141 community-dwelling adults with communication disorders due to Parkinson's disease, cerebrovascular accident, spasmodic dysphonia, or vocal fold immobility (VFI). Analysis of covariance was used to determine communicative participation differences between diagnoses, with age, sex, and hearing status as covariates. Sequential entry linear regression was used to examine associations between communicative participation and variables representing a range of psychosocial constructs across diagnoses.

Results

The VFI group had the least favorable communicative participation differing significantly from Parkinson's disease and spasmodic dysphonia groups. Self-rated speech/voice severity, self-rated effort, mental health, perceived social support, and resilience contributed to variance in communicative participation when pooled across diagnoses. The relationship between communicative participation and the variables of effort and resilience differed significantly when diagnosis was considered.

Conclusions

The findings suggest that communicative participation restrictions may vary across some diagnoses but not others. People with VFI appear to differ from other diagnosis groups in the extent of participation restrictions. Effort and resilience may play different roles in contributing to communicative participation in different disorders, but constructs such as social support, severity, and mental health appear to have consistent relationships with communicative participation across diagnoses. The findings can help clinicians identify psychosocial factors beyond the impairment that impact clients' communication in daily situations.

Speech-language pathologists (SLPs) have sought to understand how communication disorders impact people's lives, including their participation in preferred activities (Baylor et al., 2013; Dykstra et al., 2007; Eadie et al., 2006; Kovarsky et al., 2001; Ma et al., 2007; Washington et al., 2013). The term communicative participation has been defined as “taking part in life situations where knowledge, information, ideas, and feelings are exchanged” (Eadie et al., 2006). Restrictions in communicative participation can have consequences for the well-being of the individual, including social isolation, loss of employment, loss of meaningful relationships, and difficulty in accessing necessary services such as health care (Eadie et al., 2006).

This focus on participation stems from the International Classification of Functioning, Disability, and Health (ICF) presented by the World Health Organization (2001), which has enhanced clinical appreciation of biopsychosocial contributors to, and consequences of, disability that include personal, environmental, activity, and participation factors in the experience of the health condition beyond physical impairments to the body. In particular, the ICF emphasizes that participation in daily activities is influenced by a combination of external and internal factors in addition to the presence of the health condition (Perenboom & Chorus, 2003). Given that participation in life activities can be shaped by a variety of factors, it is possible that people with different physical impairments may have similar participation experiences based on influential factors beyond their diagnosis and clinical symptoms. The focus of this study is to examine variables that are associated with communicative participation and to explore similarities and differences in participation across communication disorders.

The Communicative Participation Item Bank

Patient-reported, participation-focused outcome (PRO) measures allow clinicians to better support client-centered decision making in therapy by recognizing the impact of communication disorders on clients' unique lifestyles and to develop targeted therapy that addresses the specific issues that impact participation. PROs are often regarded as a critical component of clinical assessment in order to appreciate the client's perspective and responsiveness to treatment that might not otherwise be directly observed through clinical measures (Deshpande et al., 2011; Gotay et al., 2008; Revicki et al., 2008; Yorkston et al., 2014). The Communicative Participation Item Bank (CPIB) is a PRO that measures the extent to which the respondent's health condition interferes with communicative participation in a variety of everyday conversational situations (Baylor et al., 2013). While the CPIB was designed for and has been used with a variety of communication disorder populations, currently, there is limited evidence comparing how communicative participation differs in people with different communication disorders, or if similar factors are associated with participation restrictions in these different populations.

Predictors of Communicative Participation in Communication Disorders

A series of prior studies have identified variables that appear to be associated with, and thus are likely factors contributing to restrictions in communicative participation. In people with multiple sclerosis (MS), restrictions in communicative participation were associated with increased fatigue, more severe self-reported dysarthria symptoms, more severe depression, problems thinking, no paid employment, and reduced perceived social support (Baylor et al., 2010). Similar results were reported in a later study with MS participants suggesting the association of lower levels of speech usage, higher levels of education, cognitive-communication–related difficulties, reduced physical abilities, and increased self-reported speech impairment symptoms with restricted communicative participation (Yorkston et al., 2014).

A study with head and neck cancer (HNC) survivors reported more severe self-reported speech symptoms and cognitive problems, history of laryngectomy, and shorter time since diagnosis to be variables that were significantly associated with more communicative participation restrictions (Bolt et al., 2016). A related study reported the association of lower levels of perceived social support, presence of depression, and lower resilience with poorer communicative participation in this population (Eadie et al., 2018).

For people with Parkinson's disease (PD), poorer clinician-measured cognitive status and intelligibility in conversation (based on acoustic and listener analyses of speech presentation) were significant predictors of restrictions in communicative participation (Barnish et al., 2017). When people with PD from the United States and New Zealand were compared, more severe self-rated speech impairment was the strongest predictor of reduced communicative participation, with other significant predictors including lower levels of speech usage, increased fatigue, presence of cognitive and emotional problems, and swallowing difficulties (McAuliffe et al., 2017). These relationships were also influenced by demographic factors within the study including age, sex, and country of residence.

In people with amyotrophic lateral sclerosis, poorer self-rated speech and swallowing abilities, and lower speech usage were associated with worse communicative participation (Yorkston, Baylor, & Mach, 2017). Lower levels of self-esteem, self-efficacy, and social support were predictors of reduced communicative participation in people who stutter after controlling for demographic and speech-related variables (Boyle et al., 2018).

While direct comparisons across the studies summarized above are difficult due to differences in methods and measures, some trends have emerged. Self-rated speech symptom severity appears to be the most common significant predictor of CPIB scores as evidenced in studies with people with MS (Baylor et al., 2010), HNC (Bolt et al., 2016; Eadie et al., 2016, 2018), PD (McAuliffe et al., 2017), and amyotrophic lateral sclerosis (Yorkston, Baylor, & Mach, 2017), suggesting that physical symptoms have a highly influential role in communicative participation. Other physical manifestations of the impairment such as increased speaking effort is also a salient symptom that has been shown to impact the quality of interactions in people with spasmodic dysphonia (SD; Eadie & Stepp, 2013), stuttering (Ingham et al., 2009), laryngectomy (Searl & Knollhoff, 2018), and PD (Yorkston, Baylor, & Britton, 2017). Broader physical and mental characteristics of the communication disorder have also been shown to contribute to communicative participation in PD (McAuliffe et al., 2017), MS (Baylor et al., 2010), and HNC survivors (Eadie et al., 2018). Beyond physiologic status, other psychosocial variables including social support and resilience have been found to be associated with CPIB scores in MS (Baylor et al., 2010) and HNC survivors (Eadie et al., 2018). These findings suggest the potential for similar constellations of factors shaping communicative participation in different disorders despite limitations in direct comparisons across disorders.

Evidence is also emerging from qualitative research supporting similarities in communicative participation experiences by people with different communication disorders. For example, in a study including people with dysarthria of different etiologies, participants reported that the heightened awareness of their speech presentation lead to reduced engagement in unnecessary conversations, limited amount of speaking, and avoidance of words and communication situations that they found difficult. Barriers to successful communication attempts noted by participants included increased embarrassment, reduced self-confidence, and social isolation. Overall, the experience of dysarthria, regardless of type or etiology, had affected the participants' sense of identity and self-perception (Walshe & Miller, 2011). In a study that compared environmental barriers experienced by people with dysarthria of two different etiologies, common factors restricting their communication that were brought up by both groups were background noise and familiarity of communication partners (Whitehill, 2010).

Another study including people with a variety of communication disorders related to speech, voice, language, and hearing impairments suggested that there are common factors that acted as barriers to work integration. These included personal factors such as self-perception, speech symptoms, and general fatigue, but also extended to environmental factors such as noise, number of conversation partners, and attitudes of others (Garcia et al., 2002). As a final example, in a qualitative study that explored the self-reported barriers to participation across communication disorders including dysarthria, dysphonia, and stuttering, common themes on the effect and sources of interference with communication participation emerged (Baylor et al., 2011). Participants felt the need to strategize in communicating their messages, yet even with efforts to compensate, they reported limitations in their interactions that had both functional and emotional consequences. Despite differences in the participants' communication disorder symptoms and diagnoses, there was a shared sense of lack of control, which similarly restricted participation.

Purpose of the Study

Prior work has suggested that there are common themes of participation restrictions across different communication disorders (Baylor et al., 2011; Garcia et al., 2002; Walshe & Miller, 2011; Whitehill, 2010). However, quantitative exploration of this premise is sparse. Prior surveys of SLPs have shown that there is increased interest in participation-focused treatment resources in clinical practice but barriers exist in making therapy targets related to participation accessible and evidence based (Collis & Bloch, 2012; Gauvreau, Le Dorze, Croteau, et al., 2019; Gauvreau, Le Dorze, Kairy, et al., 2019; Laliberté et al., 2016; Miller et al., 2011; Torrence et al., 2016). A quantitative analysis of the self-report variables that most strongly relate to communicative participation across disorder types will inform clinical practice and future research in identifying areas for intervention that may optimize participation outcomes for clients. Although impairment-focused interventions necessarily vary across different types of communication disorders, it is possible that aspects of participation-focused interventions that address a broader range of socio-environmental factors may not need to be re-invented for different communication disorders, which would improve efficiency in developing clinical resources (Baylor & Darling-White, 2020). Having a better understanding of similarities and differences in communicative participation across disorders will help SLPs to identify when and how resources and strategies for improving participation can be used effectively and efficiently with clients with different diagnoses, and when diagnosis-specific strategies might be needed.

The purpose of this study is to investigate whether communicative participation differs across communication disorder diagnoses characterized primarily by motor speech or voice impairments, and whether the associations between selected self-reported psychosocial variables and communicative participation differ across diagnoses. The research questions are as follows:

1. Does communicative participation differ based on the type of communication disorder diagnosis?

2. What self-report variables are associated with communicative participation, and does the relationship between these variables and communicative participation differ between diagnosis groups?

Method

All procedures for this study were approved by the institutional review board at the University of Washington. Written informed consent in English was collected from all participants. This study was a survey design using data collected as a part of a larger study exploring the stability of CPIB scores over time in people not receiving new intervention and the sensitivity of the CPIB to change in people who were receiving standard-of-care intervention. In the larger study, participants with motor speech disorders due to PD and cerebrovascular accident (CVA), and voice disorders due to SD and vocal fold immobility (VFI) were seen for two data collection sessions. The analyses reported here included only data from the first time point.

Participants

Participants were recruited from a variety of sources including community SLP clinics, the University of Washington Medical Center Voice Clinic, and the University of Washington Speech and Hearing Clinic. Participants were also recruited from community support groups for their respective diagnoses. Finally, some participants with PD were recruited from the Washington Parkinson's Disease Registry for Research. Inclusion criteria included age 18 years or older, community-dwelling status, self-reported speech- or voice-related symptoms associated with their diagnosis, ability to pass a cognitive screening, and English fluency sufficient to participate comfortably in the research methods.

The sample of participants consisted of 141 adults with motor speech or voice disorders. This included 48 participants with dysarthria due to PD, 18 participants with dysarthria due to CVA, 44 participants with SD, and 31 participants with VFI. The sample consisted of individuals with a range of histories of intervention for their communication disorder, with some participants having had no prior treatment, but many participants having had prior treatment. The type of treatments that participants received varied based on their individual clinical assessment and specific needs but reflected standard clinical practice for each diagnosis. Treatment was not administered or controlled as part of this study.

Measures

All measures are PRO measures obtained from participant surveys. These variables were selected from those available to represent a range of psychosocial issues based on the World Health Organization ICF categories that might influence communicative participation including measures of impairment/activity (disorder symptom severity, speaking effort, and overall physical health), environment (perceived social support), and personal factors (resilience and mental health).

Communicative Participation

The dependent variable was communicative participation measured using the CPIB. The general short form questionnaire used in this study consisted of 10 items that all began with the item stem, “Does your condition interfere with…” followed by a typical communication encounter such as “asking questions in a conversation” (Baylor et al., 2013). Participants rated their responses on a 4-point Likert-type scale from not at all to very much. The CPIB was developed and calibrated using item response theory for community-dwelling adults with a range of communication disorders. Scores are reported as T-scores with a mean of 50 and an SD of 10 (Baylor et al., 2013). The range of possible T-scores on the general short form is 24.20–71.00, where higher scores indicate more favorable participation.

Self-Rated Speech/Voice Severity and Effort

Participants were asked to rate the severity of their speech/voice disorder symptoms (how their speech/voice sounded to them), as well as their speech/voice effort exerted during speaking for a typical day. Speech/voice severity was rated on a 100-mm visual analog scale (VAS) with 0 = speech/voice being normal as before onset of symptoms and 100 = severe problem as perceived by the participant. A 100-mm VAS was also used to measure speech/voice effort with 0 = no effort and 100 = extreme effort as perceived by the participant.

Physical and Mental Health-Related Quality of Life

Health-related quality of life was measured using the Patient Reported Outcomes Measurement Information System Global Health measures (Hays et al., 2009). Subscores for both physical and mental health of the Global Health measure were calculated and converted into T-score values with a mean of 50 and an SD of 10. Higher scores are more favorable.

Perceived Social Support

Perceived social support can be defined as the tangible (e.g., physical or financial resources) or intangible (e.g., informational or emotional provisions) assistance the individual believes to have available from their social network (Holt-Lunstad & Uchino, 2015; Zimet et al., 1988). Perceived social support was measured using the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet et al., 1988), which indicated the participant's perception of social support available from significant others, family members, and friends. The 12 items of the MSPSS are rated on a 7-point Likert-type scale ranging from 1 (very strongly disagree) to 7 (very strongly agree). Higher scores represent greater social support with a possible summary score range of 12–84.

Resilience

Resilience is defined as resistance to negative impacts and ability to cope with adversity. Resilience was measured using the 10-item Connor–Davidson Resilience Scale (Campbell-Sills & Stein, 2007; Connor & Davidson, 2003) on which participants rate how they respond to possible adverse events on a 5-point Likert-type scale. The range of possible scores is 0–40 points, with higher scores indicating better resilience.

Demographic Variables

Participants reported their demographic information including diagnosis, age, sex, marital status, race/ethnicity, education level, employment status, living arrangement (alone, with family, etc.), time since diagnosis, history of hearing loss, presence of other significant medical conditions, and annual household income.

Data Collection

Participants were mailed questionnaire booklets prior to their in-person appointment and were asked to complete the surveys prior to the session. Questionnaire booklets were collected and reviewed for completion during the in-person session, with participants invited to complete any surveys that were not completed prior to their arrival. Participants completed the CPIB, as well as the speech/voice severity and effort ratings during the in-person session. Data were later entered into the REDCap database for storage with double entry by two research assistants to assess reliability of data entry.

Data Analysis

All data were analyzed using IBM SPSS Statistics (Version 24) predictive analytics software (IBM SPSS Statistics, 2016). Descriptive analyses included means, standard deviations, and ranges for continuous variables, and frequency counts for categorical variables.

Group Differences in Communicative Participation

Analysis of covariance (ANCOVA) was conducted to identify significant differences between diagnosis groups on the CPIB scores while controlling for potential confounding factors between groups. The CPIB short form T-scores were the dependent variable. Diagnosis was the independent variable with four levels (PD, CVA, SD, and VFI). Effects of potential confounding factors of age, sex, and hearing status were partialed out in the analysis as covariates. Demographic data (see Table 1) showed variations in age and sex based on diagnosis groups (e.g., the PD group was predominantly male and older in age, whereas the SD group was predominantly female and younger in age). A previous study also suggested an effect of age and sex on communicative participation (McAuliffe et al., 2017). Hearing status was included as covariate because hearing function has an impact on communicative interactions and hearing loss becomes more prevalent in older ages. Roughly one third of the participants across all diagnoses self-reported a history of hearing loss, but the prevalence was higher in the PD group than in the other diagnostic groups. Age was retained as a continuous variable measured in years, and sex was a categorical measure with two categories (male or female). Hearing status was in a categorical measure with two categories, no hearing loss or self-reported hearing loss. Participants who use and do not use hearing aids or other assistive devices were grouped in the self-reported hearing loss category given the small number of participants who reported the use of hearing aids or assistive devices. To ensure that statistical assumptions were tenable, normality, linearity, and homogeneity of variance were examined for each model. In addition, homogeneity of regression slopes was tested using an Aptitude-by-Treatment Interaction analysis, which tests for interactions between the independent variable (diagnosis) and the covariates (age, sex, and hearing status). Post hoc pairwise comparisons with Bonferroni adjustment were used to compare differences in CPIB T-scores between diagnosis groups.

Table 1.

Participant demographic data separated by diagnosis and for the sample as a whole.

		Diagnosis groups
Variable	Statistic/category	PD (n = 48)	CVA (n = 18)	SD (n = 44)	VFI (n = 31)	Total (n = 141)
Age (years)	M (SD)	68.79 (9.07)	63.78 (14.13)	60.56 (11.78)	55 (16.47)	62.52 (13.44)
	Range	49–84	33–84	29–85	20–87	20–87
	No response	1	0	1	0	2
Sex	Male	30 (62.5%)	6 (33.33%)	8 (18.18%)	19 (61.29%)	63 (44.68%)
	Female	18 (37.5%)	12 (66.67%)	36 (81.82%)	12 (38.71%)	78 (55.32%)
Marital status	Married/committed relationship	43 (89.58%)	10 (55.56%)	34 (77.27%)	21 (67.74%)	108 (76.60%)
	Single/divorced/widowed	5 (10.42%)	8 (44.44%)	10 (22.73%)	10 (32.26%)	33 (23.40%)
Racial/ethnic group	White or Caucasian	46 (95.83%)	15 (83.33%)	40 (90.91%)	30 (96.77%)	131 (92.91%)
	African American	0	1 (5.56%)	0	0	1 (0.71%)
	Asian	1 (2.08%)	0	1 (2.27%)	1 (3.23%)	3 (2.13%)
	More than one	1 (2.08%)	2 (11.11%)	3 (6.82%)	0	6 (4.26%)
Highest education completed	High school	0	1 (5.56%)	2 (4.55%)	2 (6.45%)	5 (3.55%)
	Vocational school	0	1 (5.56%)	4 (9.09%)	0	5 (3.55%)
	Some college	6 (12.5%)	4 (22.22%)	10 (22.73%)	10 (32.26%)	30 (21.28%)
	College	19 (39.58%)	9 (50%)	13 (29.55%)	11 (35.48%)	52 (36.88%)
	Postgraduate education	23 (47.92%)	3 (16.67%)	15 (34.09%)	8 (25.81%)	49 (34.75%)
Employment status	Paid employment	8 (16.67%)	5 (27.78%)	20 (45.46%)	14 (45.16%)	47 (33.33%)
	No paid employment a	20 (41.67%)	8 (44.44%)	6 (13.64%)	10 (32.26%)	44 (31.21%)
	Retired	20 (41.67%)	5 (27.78%)	18 (40.91%)	7 (22.58%)	50 (35.46%)
Living arrangement	With family	43 (89.58%)	12 (66.67%)	34 (77.27%)	23 (74.19%)	112 (79.43%)
	Other b	5 (10.42%)	6 (33.33%)	10 (22.73%)	8 (25.81%)	29 (20.57%)
		Diagnosis groups
Variable	Statistic/category	PD (n = 48)	CVA (n = 18)	SD (n = 44)	VFI (n = 31)	Total (n = 141)
Time since diagnosis c	< 1 year	1 (2.08%)	4 (22.22%)	10 (22.73%)	18 (58.06%)	33 (23.40%)
	1–5.9 years	10 (20.83%)	7 (38.89%)	11 (25%)	6 (19.35%)	34 (24.11%)
	6–9.9 years	17 (35.42%)	4 (22.22%)	3 (6.82%)	1 (3.23%)	25 (17.73%)
	10+ years	17 (35.42%)	3 (16.67%)	13 (29.55%)	2 (6.45%)	35 (24.82%)
History of hearing loss	No hearing loss	27 (56.25%)	13 (72.22%)	32 (72.73%)	23 (74.19%)	95 (67.38%)
	Self-reported hearing loss d	20 (41.67%)	5 (28.78%)	12 (27.27%)	8 (25.81%)	45 (31.91%)
Presence of other significant medical conditions	Yes	19 (39.58%)	6 (33.33%)	14 (31.82%)	15 (48.39%)	54 (38.30%)
	No	28 (58.33%)	12 (66.67%)	30 (68.18%)	16 (51.61%)	86 (60.99%)
Annual household income e	Less than $25,000	3 (6.25%)	3 (16.67%)	4 (9.09%)	3 (9.68%)	13 (9.22%)
	$25,000–$40,000	3 (6.25%)	4 (22.22%)	4 (9.09%)	4 (12.90%)	15 (10.64%)
	$41,000–$60,000	6 (12.5%)	3 (16.67%)	4 (9.09%)	4 (12.90%)	17 (12.06%)
	$61,000–$80,000	7 (14.58%)	1 (5.56%)	4 (9.09%)	3 (9.68%)	15 (10.64%)
	$81,000–$100,000	8 (16.67%)	2 (11.11%)	7 (15.91%)	5 (16.13%)	22 (15.60%)
	More than $100,000	12 (25%)	2 (11.11%)	16 (36.36%)	8 (25.81%)	38 (26.95%)

Note. Percentage may not add up to 100% due to no response from participants.

^a Home maker, caregiver, or unable to work due to medical conditions.

^b Alone, living with a roommate, assisted living facility/adult family home.

^c Calculated at the time of enrollment.

^d Included participants with and without hearing aids or assistive devices; given the small percentage of participants reporting use of hearing aids, groups were combined.

^e Data collected 2015–2018.

Variables Associated With Communicative Participation

To explore the relationships among the predictor variables and the CPIB, Pearson correlations were first calculated. Then, sequential entry multiple linear regression analyses were used to determine the strength of association between each of the six predictor variables (speech/voice severity, speech/voice effort, physical health, mental health, perceived social support, and resilience) and CPIB T-scores. Sequential predictor entry specifically allows for testing incremental variance accounted for as predictor(s) are added to the model. A separate regression analysis was carried out to examine the relationship of each of the six different predictor variables and CPIB due to power restrictions as well as interest in how each variable impacted communicative participation individually.

Since the sample consisted of participants with different communication disorder diagnoses, diagnosis was incorporated into the model to avoid nonindependence of residuals. Confounding factors of age, sex, and hearing status were also included in the model to remain consistent with the ANCOVA analysis. Normality, linearity, and homoscedasticity of residuals were examined to ensure that linear regression model assumptions were tenable. Block 1 included diagnosis groups (with PD serving as the reference group) and confounding variables (age, sex, and hearing status). Block 2 included main effects of one of the six predictor variables. Block 3 included interaction terms between the predictor variable and each diagnosis group to examine whether diagnosis influenced the relationship between the variable of interest and the CPIB. z-score derivations for each of the six predictor variables were used in the regression analysis to standardize the comparison of scores.

Difference in the variance accounted for in the models without and with the predictor variable (Blocks 1 vs. 1 + 2, respectively) revealed how strongly the predictor variable uniquely contributed to communicative participation with age, sex, and hearing status held constant and when pooled across diagnoses. The unstandardized coefficient of the predictor variable informed the degree and direction of its association with communicative participation. F tests were used to compare the model without and with the Diagnosis × Predictor Variable interaction terms (blocks 1 + 2 vs. 1 + 2 + 3, respectively) to test for evidence that the relationship between the CPIB and the predictor variable was modified by diagnosis. Significance in the F test indicated that the addition of interaction terms significantly improved the model prediction, therefore suggesting that the relationship between the predictor variable and the CPIB varied across diagnoses.

Results

The results section summarizes demographic and descriptive data followed by the two research questions.

Demographic Variables

The mean age of the 141 participants was 62.52 years (SD = 13.44) with a range from 33 to 84 years. The VFI group had the lowest mean age at 55 years (SD = 16.47) and the PD group had the highest mean age at 68.79 years (SD = 9.07). The sample had slightly more females (55.32%) overall than males, but the distribution varied in each diagnosis group (female = 37.5%, 66.67%, 81.82%, 38.71% in PD, CVA, SD, and VFI groups, respectively). The majority of the sample self-identified as White or Caucasian (92.91%), had some college education or higher (92.91%), and lived with family (79.43%). Roughly two thirds (67.38%) of the sample reported no hearing difficulties, whereas the remaining one third reported symptoms of hearing loss. Additional details of demographic information are available in Table 1.

Descriptive Data

CPIB T-scores and scores for the six predictor variables for each diagnosis group and for the sample as a whole are presented in Table 2. The mean CPIB T-score for all participants was 45.23 (SD = 9.46) and is consistent with CPIB scores reported in previous literature (Baylor et al., 2010; Bolt et al., 2016; Boyle et al., 2018; Eadie et al., 2014, 2018; Yorkston et al., 2014; Yorkston, Baylor, & Britton, 2017). Within each diagnosis group, the mean CPIB score ranged from the VFI group being the lowest at 39.76 (SD = 8.02) to the SD group being the highest at 48.06 (SD = 11.37). For each of the six predictor variables, the VFI group had the least favorable mean score in five of the variables (self-rated speech/voice severity, effort, mental health, perceived social support, and resilience), while the PD group had the least favorable mean score in physical health. The SD group had the most favorable mean score in four of the six variables (physical health, mental health, perceived social support, and resilience), while the CVA group had the most favorable mean score in self-rated speech/voice severity and effort.

Table 2.

Descriptive data for communicative participation and predictor variables separated by diagnosis and for the sample as a whole.

	Diagnosis groups
	PD n = 48	CVA n = 18	SD n = 44	VFI n = 31	Total n = 141
Communicative participation	46.89 (7.89)	43.34 (6.15)	48.06 (11.37)	39.76 (8.02)	45.23 (9.46)
Speech/voice severity	n = 46				n = 139
	50.43 (22.50)	54.61 (23.92)	47.70 (31.13)	76.23 (16.04)	55.86 (26.79)
Speech/voice effort	n = 45				n = 138
	40.11 (19.99)	50.67 (28.59)	37.80 (29.13)	57.03 (26.60)	44.55 (26.69)
Physical health	n = 46		n = 43		n = 138
	43.74 (7.30)	46.83 (7.16)	52.71 (7.65)	46.41 (10.26)	47.54 (8.85)
Mental health	n = 46	n = 17		n = 30	n = 137
	46.19 (8.25)	45.91 (7.05)	48.84 (10.34)	45.82 (9.01)	46.92 (9.01)
Social support	n = 45	n = 17	n = 43	n = 30	n = 135
	69.04 (15.70)	68.18 (13.53)	69.72 (16.58)	68.07 (16.59)	68.93 (15.78)
Resilience	n = 47		n = 43		n = 139
	28.45 (5.62)	29.94 (5.99)	31.95 (6.54)	28.19 (5.47)	29.67 (6.09)

Note. Descriptive statistics include mean and standard deviation (in parenthesis). All participants accounted for unless alternative n stated in each section due to missing data. Disorder groups (PD = Parkinson's disease; CVA= cerebrovascular accident; SD = spasmodic dysphonia; VFI = vocal fold immobility); Communicative Participation = Communicative Participation Item Bank (CPIB) short form (T-score); Severity = self-rating of speech/voice severity on a typical day (raw score); Effort = self-rating of overall speech/voice effort on a typical day (raw score); Physical health = Patient-Reported Outcomes Measurement Information System (PROMIS) physical health scale (T-score); Mental health = PROMIS mental health scale (T-score); Social support = Multidimensional Scale of Perceived Social Support; Resilience = Connor–Davidson Resilience Scale 10.

Group Differences in Communicative Participation

Assumptions of normality, homogeneity of variance, linearity, and homogeneity of regression slopes were satisfactory. Specifically, the assumption of homogeneity of regression slope was satisfied as the interaction effect between the independent variable (diagnosis) and the covariates (age, sex, and hearing status) were not significant (p = .652, p = .913, p = .935, respectively). After adjusting for covariates, ANCOVA results showed that there was a significant effect of diagnosis on communicative participation, F(3, 131) = 5.97, p = .001, r ² = .14. This effect size is large according to Cohen's standards (Cohen, 1988).

Post hoc pairwise comparisons using Bonferroni's correction were conducted for six contrasts (corrected p = .008). Results showed that the PD group had significantly higher communicative participation (M = 46.58, 95% CI [43.78, 49.38]) than the VFI group (M = 39.50, 95% CI [36.08, 42.92]), p = .002, d = 0.78. The SD group (M = 48.65, 95% CI [45.77, 51.52]) also had significantly higher communicative participation than the VFI group, p < .001, d = 1.01. No other group differences were significant. Figure 1 presents the group differences in CPIB T-scores adjusted by age, sex, and hearing status.

Figure 1.

CPIB T-scores by diagnosis groups, adjusted for age, sex, and hearing status. The possible CPIB range is 24.2–71.0 with higher scores being more favorable. Error bars indicate ± 1 SE. Brackets indicate statistically significant contrasts between diagnosis groups. CPIB = Communicative Participation Item Bank (short form T-scores); PD = Parkinson's disease; CVA = cerebrovascular accident; SD = spasmodic dysphonia; VFI = vocal fold immobility.

Variables Associated With Communicative Participation

A zero-order correlation matrix of communicative participation and all predictor variables is presented in Table 3. Better communicative participation was most highly associated with less self-reported speech severity and effort, followed by better mental health, better resilience, better physical health, and better perceived social support.

Table 3.

Zero-order correlations of CPIB and predictor variables.

Measure	1.	2.	3.	4.	5.	6.	7.
Outcome
1. Communicative Participation	—
Predictor variables
2. Speech severity	–.72**	—
3. Speech effort	–.54**	.62**	—
4. Physical health	.20*	–.21**	–.26**	—
5. Mental health	.34**	–.27**	–.31**	.61**	—
6. Social support	.20*	–.13	–.17	.16	.44**	—
7. Resilience	.24**	–.22**	–.22**	.52**	.6**	.42**	—

Note. Bold indicates that a higher score of the measure indicates a worse outcome (i.e., higher severity and increased effort). Other measures not bolded indicate higher scores are better outcomes. Communicative Participation = Communicative Participation Item Bank (CPIB; short form T-scores); Severity = self-rating of speech or voice severity on a typical day (raw score); Effort = self-rating of overall speech or voice effort on a typical day (raw score); Physical health = Patient-Reported Outcomes Measurement Information System (PROMIS) physical health scale (T-score); Mental health = PROMIS mental health scale (T-score); Social support = Multidimensional Scale of Perceived Social Support scale (T-score); Resilience = Connor–Davidson Resilience Scale 10.

^* p < .05.

^** p < .01.

Results for the sequential entry multiple linear regression models for each of the six predictor variables, the amount of variance in CPIB scores accounted for by the variable, degree and effect of the variable, and evidence for effect modification between diagnosis group and the variable are presented in Table 4. To understand the nature of the interactions, in Figure 2, predicted values of the CPIB T-scores were plotted for each diagnosis group by three levels of each of the predicted variables in z scores, –1 SD, mean, and +1 SD.

Table 4.

Sequential entry multiple linear regression model results for the relationship between predictor variables and communicative participation across diagnosis groups.

Variables	Relationship between variable and CPIB when pooled across diagnosis groups				Interaction between variable and diagnosis
	Change in variance accounted for by variable across diagnoses (R 2 change)	Unstandardized coefficient (b)	95% Confidence interval	Effect size (sr2)	F-test change significance for comparing model with and without interaction terms (p)
Severity	.41	–6.80***	[–8.05, –5.54]	.41	.154
Effort	.21***	–4.70***	[–6.15, –3.25]	.21	.005**
Physical health	.02	1.35	[–.41, 3.12]	.02	.902
Mental health	.09***	2.98***	[1.48, 4.48]	.09	.318
Social support	.03*	1.68*	[.17, 3.19]	.03	.107
Resilience	.03*	1.86*	[.22, 3.50]	.03	.043*

Note. Severity = self-rating of speech or voice severity on a typical day (raw score); Effort = self-rating of overall speech or voice effort on a typical day (raw score); Physical health = Patient-Reported Outcomes Measurement Information System (PROMIS) physical health scale (T-score); Mental health = PROMIS mental health scale (T-score); Social support = Multidimensional Scale of Perceived Social Support; Resilience = Connor–Davidson Resilience Scale 10.

^* p < .05.

^** p < .01.

^*** p < .001.

Figure 2.

Interactions between CPIB and predictor variables by diagnosis groups. y-axis indicates predicted CPIB T-scores based on regression model analysis adjusted for age, sex, and hearing status. The possible CPIB T-score range is 24.2–71.0 with higher scores being more favorable. x-axis indicates changes in standardized predictor variables at –1, 0, and +1 SDs. For predictor variables, high scores are better except for severity and effort (low scores are better). CPIB = Communicative Participation Item Bank (short form T-scores); PD = Parkinson's disease; CVA = cerebrovascular accident; SD = spasmodic dysphonia; VFI = vocal fold immobility.

Severity

Self-reported speech/voice severity accounted for an additional 41% of the variance in communicative participation when pooled across diagnosis and after controlling for confounding variables. Severity had a unique negative effect on communicative participation, where there was an estimated mean decrease of 6.80 points in CPIB T-scores for every standard deviation increase in self-reported severity, holding all else constant. There was no significant evidence for change in the model by considering the interaction between diagnosis and speech/voice severity.

Effort

Self-reported speech/voice effort on its own accounted for 21% of variance in communicative participation when pooled across diagnosis and after controlling for confounding variables. Effort had a unique negative effect on communicative participation where there was an estimated mean decrease of 4.70 points in CPIB T-scores for every standard deviation increase in effort, holding all else constant. The addition of the interaction terms revealed statistically significant evidence that the relationships between CPIB and effort were modified by diagnosis.

Physical Health

Physical health accounted for 2% of the variance in CPIB scores when pooled across diagnosis and after controlling for confounding variables. This effect was not significant. There was no significant evidence for change in the model by considering the interaction of diagnosis and physical health.

Mental Health

Mental health status on its own accounted for 9% of variance in communicative participation when pooled across diagnosis after controlling for confounding variables. It had a unique positive effect on communicative participation where there was an estimated mean increase of 2.98 points in CPIB T-scores for every standard deviation increase in mental health, holding all else constant. There was no significant evidence for change in the model by considering the interaction of diagnosis and mental health.

Perceived Social Support

Perceived social support on its own accounted for 3% of the variance in communicative participation when pooled across diagnosis after controlling for confounding variables. Social support had a unique positive effect on communicative participation where there was an estimated mean increase of 1.68 points in CPIB T-scores for every standard deviation increase in perceived social support, holding all else constant. There was no significant evidence for change in the model by considering the interaction between diagnosis and social support.

Resilience

Resilience on its own accounted for 3% of variance in communicative participation when pooled across diagnosis after controlling for confounding variables. Resilience had a unique positive effect on communicative participation where there was an estimated mean increase of 1.86 points in CPIB T-score for every standard deviation increase in perceived change in resilience, holding all else constant. The addition of interaction terms showed statistically significant evidence that diagnosis modified the relationship between resilience and CPIB scores.

Discussion

The purpose of this study was twofold. The first aim was to examine if communicative participation differs in people with different communication disorders. The second aim was to investigate what self-reported variables are most strongly associated with communicative participation and if those associations differ across communication disorder diagnoses.

Communicative Participation Differences Based on Diagnosis

The results of this study suggest that there are differences between some diagnosis groups but not others on communicative participation scores, with an overall significant and large effect of diagnosis on communicative participation. In particular, of the four diagnosis groups examined, the VFI group had the least favorable CPIB scores and these were statistically different from both PD and SD groups. One metric for interpreting clinical significance is half a standard deviation difference in scores (Norman et al., 2003), which is 5 points on the CPIB T-score. If this metric is used, the results suggest that the difference between the VFI group and each of the PD and SD groups on CPIB scores is clinically significant as well as statistically significant. In examining possible reasons for these findings, a descriptive review of Table 2 showed that the VFI group reported the least favorable scores in self-rated speech/voice severity, speech/voice effort, mental health, perceived social support, and resilience compared to the other diagnosis groups. The VFI group also reported a higher presence of other significant medical conditions (48.39%, compared to 31.82%–39.58% in other groups). It is possible that some constellation of these factors contributed to the lower CPIB scores for the VFI group. Differences between diagnosis groups on CPIB scores could also be attributed to other differences between the groups, such as their treatment status. While the PD, CVA, and SD groups all contained a mixture of participants who had had prior treatment and those who had not, the VFI group was the only group where participants had no previous treatment related to their voice condition and had more recent diagnoses (58.06% had < 1 year since diagnosis; see Table 1). This may suggest that people who are recently diagnosed with a communication disorder without clinical intervention may experience higher levels of restriction in their communicative participation. Thus, while these findings point to evidence that the extent of communicative participation restrictions may vary across diagnoses, it is not entirely clear if those differences can be attributed to the speech/voice characteristics themselves or to other characteristics of the sample.

Self-Report Variables Associated With Communicative Participation

This study investigated the association of six self-report predictor variables representing a range of psychosocial constructs with the CPIB. All of the predictor variables individually had significant unique associations with CPIB scores except for physical health. The highest levels of variance accounted for in CPIB scores were by speech/voice severity and effort (41% and 21%, respectively). The pattern of self-rated speech/voice severity being most strongly associated with participation has been previously reported when communicative participation was examined in people with MS (Baylor et al., 2010), HNC (Bolt et al., 2016; Eadie et al., 2018), laryngectomy (Eadie et al., 2016), and PD (McAuliffe et al., 2017). The current findings also support existing evidence that effort associated with speaking is a commonly reported symptom of the communication disorder in people with speech and voice disorders (Eadie & Stepp, 2013; McKenna & Stepp, 2018; Searl & Knollhoff, 2018; Stepp et al., 2012; Yorkston, Baylor, & Britton, 2017).

In this study, mental health, as measured by the Patient Reported Outcomes Measurement Information System Global Health instrument, accounted for 9% of the variance in communicative participation. This finding generally supports other literature examining the relationship between mental or emotional health and participation, although the measures used in the studies differed. For example, Baylor et al. (2010) found a strong association between higher levels of depression, as measured with the Center for Epidemiological Studies Depression Scales, and lower communicative participation in people with MS. In a sample of people with HNC, lower depression as measured with the Hospital Anxiety and Depression Scale was associated with more favorable communicative participation (Eadie et al., 2018). A prior qualitative study on communicative participation across diagnosis groups also reported accounts from participants that negative emotions experienced due to their communication difficulties contributed to participation restrictions (Baylor et al., 2011).

The positive relationship between perceived social support and communicative participation observed in this study is consistent with similar findings in people who stutter (Boyle et al., 2018), survivors of HNC (Eadie et al., 2018), and people with MS (Baylor et al., 2010). In these studies, better perceived social support as measured by the MSPSS was associated with higher scores on communicative participation but the extent of that relationship varied across studies.

Resilience is a newer area of investigation in terms of its relationship with communicative participation, and initial findings have been mixed. The current study found a statistically significant relationship between resilience and communicative participation, and similar connections between resilience and communication have been found for people who stutter (Craig et al., 2011). The study of resilience is an emerging topic in health care research that focuses on positive health, but its exact contribution to rehabilitation is currently confounded with other psychosocial constructs (Craig et al., 2011; Eadie et al., 2018; White et al., 2008). Future research is needed to single out the effect of resilience on communicative participation.

Differences in Relationships Between Predictor Variables and Communicative Participation Across Diagnoses

The results of this study showed that the relationship between the predictor variables and communicative participation was modified by diagnosis for two of the six variables—speech/voice effort and resilience, but the relationships between communicative participation and each of the constructs of speech/voice severity, physical and mental health, and social support may not vary across diagnosis. Due to limited statistical power, further statistical analysis of this relationship was not explored in this article.

Visual inspection of the graphs in Figure 2 comparing the association of speech/voice effort and CPIB scores in the four diagnosis groups showed that while there was a general negative relationship between effort and CPIB (more effort associated with worse CPIB scores), the trend was more marked for the SD group. Given that vocal effort exerted during speaking is a characteristic trait of SD (Baylor et al., 2005; Eadie & Stepp, 2013; Nagle et al., 2015), it is possible that effort has a more pronounced effect on communicative participation in the SD population than the other diagnoses accounted for in this study.

One unexpected finding from the visual inspection of Figure 2 showed that higher resilience was associated with lower CPIB scores in the VFI group. While prior research is limited, a qualitative study of participants' experience with unilateral vocal fold paralysis suggested that resilience emerged as a positive factor that modified their voice experiences (Francis et al., 2018). One difference between the two studies is that 78% of the participants in the Francis et al. (2018) study had received treatment for their voice while the participants in the current study were recently diagnosed and had not yet received any intervention. This is an area in need of future research to achieve a better understanding of the relationship between communicative participation and resilience, and how interventions directed at voice or each of these constructs may impact one another. With propositions that resilience is a factor that can be facilitated through intervention, its effect on communication outcomes could be modified positively if it is appropriately addressed as part of therapy goals (White et al., 2010).

Clinical Implications

All communication disorder diagnosis groups in this study reported restrictions in communicative participation; therefore, clinical understanding of what impacts communicative participation is important in order to address the participation experiences of the client. The findings from this study contribute to clinical practice by guiding SLPs to identify psychosocial factors beyond the physical impairment that could impact how clients use their speech or voice in daily activities. Comparing different communication disorders showed that the experience of communicative participation is influenced by a multitude of factors regardless of the diagnosis for the most part, but clinical expertise of each diagnosis and its associated symptoms is still essential in determining the best course of intervention for the clients' unique experience of their conditions. In order to optimize therapy outcomes and encourage generalization of practiced skills, SLPs are encouraged to address the clients' speech/voice disorder symptoms as well as the environment in which they communicate. Clinicians should keep in mind of the multidimensionality of the health condition as addressed in the ICF, and that addressing personal and environmental factors is also relevant to optimize communication outcomes.

Limitations and Future Directions

This sample was a convenience sample recruited from community sites that impacted size and characteristics of the sample. Related to characteristics of the sample, participants represented a largely Caucasian and highly educated sample, which may limit generalization to a more diverse population. It should be recognized that this is a secondary analysis of data collected for a larger study and is exploratory in nature. This may lead to issues related to treatment status being a confound in interpretation.

Another issue warranting consideration is the measurement of self-rated speech/voice severity. The same VAS and anchors were used across diagnosis groups to obtain an indication of how participants rated the severity of their communication disorder symptoms. However, given that the actual symptoms would differ across groups (e.g., different types of voice symptoms, or voice symptoms vs. other speech symptoms), it is unknown whether these ratings across diagnostic groups represent the same or different constructs. Given that self-rated speech/voice symptom severity consistently emerged as a strong predictor of communicative participation across disorders, this topic may warrant further investigation to better understand what parameters impact client self-ratings of symptom severity.

Finally, the decision to examine each predictor variable and its relationship to communicative participation separately across diagnoses reflects the statistical power limitations. Testing of more complex models incorporating all predictor variables and confounding factors requires a larger sample size, but this study may inform variables to include in a more rigorous study later. Acknowledging these limitations, the purpose of this study was to examine communicative participation and related variables in a sample representing a range of severities and experiences. Further research should also examine other psychosocial variables for their roles in impacting communicative participation.

Funding Statement

This work was supported by the National Institute for Deafness and other Communication Disorders (1R01DC012510-01A1; PI Baylor), and the Department of Rehabilitation Medicine at the University of Washington.