Patient Motivation for Voice Therapy Increases After Stimulability Testing: Validation of a Voice Therapy Motivation and Voice Perception Inventory

Abstract

Objective:

Measuring patient motivation for voice therapy is an important component of determining prognosis and estimating compliance. Voice stimulability testing is assumed to influence the patient's motivation by improving their perception of the sound and/or feel of the voice. The purpose of this study was to validate a Voice Therapy Motivation and Voice Perception Inventory and to determine whether stimulability testing elicits improvements in motivation and patient-perceived voice status.

Method:

Seventy-five patients who underwent initial voice therapy evaluation completed a four-item inventory that queried their commitment to voice therapy, confidence that they could change their voice through voice therapy, perception of their voice severity, and ratings of their vocal effort. Inventories were completed shortly before and after stimulability testing. A psychometric analysis of the survey was conducted, including the internal consistency of each construct of the inventory along with content, predictive, and concurrent validity. Paired-samples statistical analyses were conducted to determine differences before and after stimulability testing, and Spearman correlations between voice and motivation ratings were conducted to determine relationships between the patient's perception of their voice and their motivation for voice therapy.

Results:

Internal consistency reliability was acceptable to good for each construct, particularly after stimulability testing. Motivation scales were predictive of returning for voice therapy sessions, subjective improvement of voice by the first therapy session, and likelihood of the patient practicing the home practice program. Patient ratings of voice severity and effort were positively correlated with clinician ratings of overall severity. Motivation and voice impairment perception improved following stimulability testing. Reduction in vocal effort was related to improved motivation, but change in voice severity was not.

Conclusions:

Voice stimulability testing can influence patient motivation for voice therapy and patient-perceived voice improvements. It should be considered a critical part of the voice evaluation.

Voice stimulability testing is a central component of the behavioral voice evaluation. The primary goal of stimulability testing is to determine whether a patient can immediately change their vocal quality, effort, or efficiency given a variety of facilitating tasks and levels of cueing (Dejonckere & Lebacq, 2001; Gillespie & Gartner-Schmidt, 2015). The use of stimulability testing by voice-specialized speech-language pathologists (SLPs) is nearly universal and highly valued (Toles & Young, 2023). Despite this widespread use, some factors associated with its use remain nebulous.

Qualitative responses from a recent clinical practice survey (Toles & Young, 2023) shed light on the complex role that stimulability testing plays in voice evaluation. Beyond its primary goal of assessing voice change, SLPs reported that the procedure can aid in therapeutic decision making, such as determining the patient's candidacy for therapy, identifying individualized therapy strategies, and estimating the patient's prognosis for improvement. The survey also found that SLPs use voice stimulability testing to assess cognitive and psychological components associated with the voice disorder. Specifically, stimulability testing was reported to aid in assessing the patient's awareness of their own voice and their motivation to take part in voice therapy. These components are related to the meta-therapy paradigm, described by Helou and colleagues (Helou, 2017; Helou et al., 2021). Petty et al. (2023) went on to describe how meta-therapy components can be applied to the initial voice evaluation to lay the foundation for therapy and to assess and support the patient's readiness for change. In part, they suggested that stimulability testing is one of the most important ingredients the clinician can use to target the patient's motivation for improvement.

Patient buy-in and motivation to participate in the voice therapy process are important. Attrition levels in voice therapy can be quite high, with up to half of the patients who are referred for voice therapy not initiating or completing their course of voice therapy (Hapner et al., 2009; Portone et al., 2008). There are many potential reasons for poor adherence or commitment to voice therapy, such as insurance denials or time constraints (Portone et al., 2008). However, it is often attributed, at least in part, to a lack of buy-in for the therapy process (van Leer & Connor, 2010). In turn, lack of buy-in can indirectly lead to poor outcomes due to limited confidence in the ability to make change or poor commitment to adhere to voice therapy recommendations. Nevertheless, motivation is considered a dynamic state versus a static trait and is susceptible to influence from various factors (Behrman, 2006). Understanding factors that influence patient motivation for voice therapy and implementing methods that directly target improving motivation could improve the effectiveness of voice therapy.

Motivation is closely aligned with the concept of self-efficacy. Self-efficacy is associated with the belief that one is capable of achieving a particular goal (Bandura et al., 1999). Bonilha and Dawson (2012) proposed that having a “mastery experience” during the initial evaluation could impact a patient's self-efficacy for achieving voice change. The vibration that occurs from the biomechanical interaction of the respiratory, laryngeal, and supralaryngeal structures during voice production is perceived by the individual via feedback from both the auditory and somatosensory systems (Larson et al., 2008; Smith et al., 2020; Titze, 2000). These systems are vital for monitoring and self-correcting disturbances in the vocal system (Weerathunge et al., 2022). Thus, individuals who are more aware of how their voice sounds and/or feels might be more likely to make immediate changes in the voice with stimulability testing and establish that “mastery experience” to feel confident that they can make changes with voice therapy (Bonilha & Dawson, 2012). Therefore, to understand how stimulability testing affects motivation and self-efficacy, it is important to measure not only a patient's motivation but also the patient's perception of the sound and feel of their voice.

The question at hand is “How can stimulability testing assess or influence motivation?” Although the clinical practice survey (Toles & Young, 2023) illuminated the concept of motivation as a component of stimulability testing, it is unclear how clinicians might actually gauge whether or not stimulability affects a patient's motivation for voice therapy. Two recent studies found no relationship between the level of stimulability (i.e., stimulable or not stimulable) and voice therapy attendance, but these studies did not measure motivation (McDowell et al., 2023; Weston et al., 2023). Currently, no direct measurements of patient motivation are standard practice for the behavioral voice evaluation, so clinicians might resort to judging these concepts anecdotally or based solely on clinical instinct. This could lead to unconscious biases affecting the clinician's judgment when recommending therapy or educating the patient. Therefore, it is necessary to establish a way to objectively measure motivation for voice therapy associated with fewer clinician biases. Having an objective estimate of the patient's motivation will support concordance between the patient and the clinician, facilitating the patient–clinician relationship when transitioning from evaluation to therapy.

One group has reported the use of a tool to measure motivation for practicing voice therapy exercises in several studies (van Leer & Connor, 2012, 2015; van Leer et al., 2017, 2021). For those studies, the “readiness ruler” (Miller & Rollnick, 2012) was modified for voice therapy. Items on the readiness ruler included the following: (a) “How confident are you that you can achieve this goal?” (b) “How committed are you to this goal?” and (c) “How important is this goal compared to other things you have to do this week?” These items were asked in reference to specific goals, such as “practicing voice therapy exercises 3–5 times per day” or “staying in your best voice throughout the day,” and were rated on a 0–10 interval scale. These items demonstrated predictive validity for the stated goals (van Leer & Connor, 2015). Because the questions are specific to those goals, they are not necessarily generalizable to the patient's overall motivation for voice therapy. Thus, the readiness ruler in its reported form might not be appropriate for measuring broad motivation for voice therapy but could be adjusted for that purpose. Modification and simplification of the scale could allow the measurement of broad motivation for voice therapy quickly and without much burden on the patient both before and after stimulability testing during the time-constrained evaluation with an SLP.

For the current study, we modified the readiness ruler to limit it to two items pertaining to motivation for voice therapy in general, and we also included two items pertaining to self-assessment of voice severity and vocal effort. Our first objective was to assess the reliability and validity of this brief Voice Therapy Motivation and Voice Perception Inventory. Then, we aimed to answer the following two questions: (a) Do perceptions change following stimulability testing? Specifically, we addressed patient perceptions of confidence in and commitment to voice therapy, self-ratings of voice severity and vocal effort, and clinician ratings of voice severity. (b) Are improvements in self-ratings of voice severity and effort related to increased confidence or commitment?

Method

Participants

The institutional review board at the University of Texas Southwestern Medical Center (UTSW; STU-2022-0388) approved this study. At UTSW, SLP voice evaluations are conducted either in a combined clinic with a laryngologist or as an independent SLP visit. For this study, we only included individuals who were seen for an independent SLP evaluation after being referred by their laryngologist (i.e., patients who were seen by the SLP and laryngologist simultaneously were not included) to limit extraneous variables (e.g., presence of extra individuals in the room, completion of stroboscopy, potential emotional reactions surrounding a new diagnosis). Medical charts were reviewed of 140 patients who underwent an independent behavioral voice evaluation with a voice-specialized SLP after being referred by their laryngologist at the UTSW Clinical Center for Voice Care between February 1, 2023, and November 30, 2023. All patients underwent a stimulability testing session during the voice evaluation. Only patients who had all self-report scale measures documented in their record were chosen for analyses. Patients were excluded from analyses if they did not complete any of the self-rating inventories, only completed a portion of the inventory, or completed only one time point or the SLP did not complete the second Overall Severity subscale. Patients who did not present with voice complaints (e.g., upper airway, dysphagia) were also excluded. Three SLPs collected self-report measures from their new voice patients. Each SLP completed the initial review of their own patient charts, and the first author completed a second review of all the charts. SLPs had clinical experience ranging from 2 to 21 years of practice assessing and treating patients with voice disorders as a primary part of their caseload. Of the total 140 patients who were reviewed, 75 met all inclusion criteria for analysis. All voice evaluations were conducted in person. Although all patients who come to UTSW are given the option of telehealth visits, only nine (17%) of the 54 patients who returned for the first therapy session were treated through telehealth. The remaining 83% chose to return for an in-person visit. Table 1 shows the demographics for the patient sample stratified by diagnosis. Diagnoses were binned into the following categories: muscle tension dysphonia (i.e., non-phonotraumatic vocal hyperfunction), benign mucosal vocal fold lesions (e.g., vocal fold nodules, polyps, cysts), other structural voice disorders (e.g., scar, sulcus, granuloma), and neurological voice disorders (e.g., vocal fold paralysis, vocal tremor, laryngeal dystonia).

Table 1.

Demographic information for the patient sample stratified by diagnosis.

Diagnosis	n	Age, years M (SD)	Female n (%)	Male n (%)
Combined groups	75	51.1 (18.5)	56 (74.7)	19 (25.3)
Muscle tension dysphonia	26	53.1 (16.2)	19 (73.1)	7 (26.9)
Benign mucosal vocal fold lesions	26	39.0 (16.5)	21 (80.8)	5 (19.2)
Other structural disorder	7	63.9 (21.6)	4 (57.1)	3 (42.9)
Neurological voice disorder	16	62.0 (11.6)	12 (75.0)	4 (25.0)

Data Collection

Rating Scales

Documentation of the patient's initial voice evaluation with the SLP and their first voice therapy session were reviewed. During the voice evaluation shortly prior to conducting stimulability testing, the SLP provided the patient with the four-item inventory. Patients rated each item on a 100-mm visual analog scale (VAS). The scales were described to the patient, and the patient was provided anchors as to what each end of the scale represented. The inventory, which can be seen in Figure 1, consisted of two questions about their motivation that were modified from the readiness ruler (van Leer & Connor, 2012, 2015; van Leer et al., 2017, 2021): (a) “How confident are you that you can make improvements in your voice through voice therapy?” and (b) “How committed are you to participating in voice therapy?” For these first two scales, 0 (i.e., the leftmost portion of the scale) represented not confident or committed at all, and 100 (i.e., the rightmost portion of the scale) represented extremely confident or committed. The next two items queried the patient's perception of their own voice: (c) “Rate the severity of your voice quality” and (d) “Rate the amount of physical effort it takes to produce your voice.” For these scales, 0 represented typical voice quality for their age/gender and minimal effort (respectively), whereas 100 represented severely affected voice quality and extreme effort (respectively). Simultaneously, the treating SLP rated the overall severity of the patient's voice quality on a VAS as they would rate the first item on the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V) auditory-perceptual scale (Kempster et al., 2009). Clinicians rated overall severity at the same time the patients completed their survey without looking at the patient's ratings to limit bias that could be associated with knowing how the patient rated themselves. Patients and clinicians rated the same items again shortly after completing stimulability testing. For the posttesting measures, patients were instructed to rate the items based on how they felt during and immediately following the trial tasks. Patients were not shown their prestimulability scores when rating the poststimulability inventory; however, clinicians were not blinded to their prestimulability overall severity scores.

Figure 1.

Voice Therapy Motivation and Voice Perception Inventory.

Stimulability Testing Procedure

Stimulability testing varied based on the patient's individual presentation and which SLP provided the assessment tasks. During independent SLP voice evaluations (i.e., those not conducted simultaneously during a laryngologist/SLP joint clinic), the SLPs at the UTSW Voice Center reserve a portion of their evaluation time for a dedicated stimulability testing session. During this session, the SLPs present a variety of facilitating tasks that target different aspects of vocal function during stimulability testing (McDowell et al., 2023). The facilitating tasks that are introduced include, but are not limited to, loudness or pitch modification, respiration modifications, forward resonance modifications, flow or breathy phonation, semi-occluded vocal tract techniques, and clear speech cues. The clinician might make various adaptations to any of the facilitating techniques to optimize the patient's vocal function and/or to aid in translating the improved voice into connected speech or conversation. The SLP often uses the task that was the easiest for the patient to produce or that optimized their vocal function the best to facilitate transferring improvements into connected speech. The SLP and the patient work through this process together to design the individualized home practice program or generalization technique to implement until the first follow-up session. Patients are always given a home practice program at the evaluation to practice until the first therapy session.

Statistical Analyses

Reliability

The first stage in our statistical analyses was assessment of the reliability and validity of the self-reported measures. Patients completed ratings once prior to and again after stimulability testing. Scores were expected to change between the first and second measurements due to the introduction of stimulability testing. This precluded the ability to obtain consistent repeated measurements to calculate test–retest reliability. Therefore, we calculated the internal consistency reliability of the four self-report items included in the inventory using Cronbach's alpha (Tavakol & Dennick, 2011), which is a measure that determines the extent to which items load onto a cohesive construct. Since there are two primary theoretical constructs in this four-item survey, we separated the four items based on the two constructs: motivation and voice perception. The motivation construct includes Items 1 and 2, which are related to confidence in and commitment to voice therapy. The voice perception construct includes Items 3 and 4, which queried the patient's perception of the severity of their voice quality and their vocal effort. Because there were only two items per construct, α values higher than .90 would indicate that the two questions are redundant and could be combined (Tavakol & Dennick, 2011). The α values lower than .50 would suggest that the two items are not measuring a similar construct. The α values ≥ .80 and < .90 were interpreted as good, values ≥ .60 and < .80 were considered acceptable, values ≥ .50 and < .60 were considered fair, and values < .50 were considered unacceptable. The α values were calculated using the baseline and the poststimulability scores separately to determine if the stage of rating affected the reliability of the survey. We hypothesized that the covariance of the two items within each construct would increase following stimulability testing.

Validity

Content validity was assessed from both the clinician perspective and the patient perspective. Nine voice-specialized SLPs (mean years of experience [SD] = 12.4 [10.0]) and nine patients who were participating in voice therapy were asked to provide feedback as to whether each item in the survey appears to test what it aims to test and whether each was essential to assess the motivation and voice perception constructs. Patients who participated in content validity rating had diagnoses of muscle tension dysphonia (n = 3), phonotrauma (n = 4), and neurological voice disorders (n = 2). Each SLP and patient rated whether the item was essential, useful but not essential, or not necessary for the given construct. The content validity ratio (CVR) was calculated for each item within each group. The CVR quantifies whether the item is essential to measure the construct by determining the agreement of “essential” ratings across raters. CVR values can range from −1 (perfect disagreement) to +1 (perfect agreement; Lawshe, 1975). The Content Validity Index (CVI), which is the mean value of CVR across items, was then calculated for each construct as well as for the entire inventory within each group. CVR scores were interpreted based on the critical value ≥ .548 for a group of nine raters, suggested by Wilson et al. (2012). Items greater than or equal to the critical value were interpreted as having adequate content validity.

Criterion-related validity was assessed through predictive and concurrent validity measures. We hypothesized that patients who felt more confident in and committed to voice therapy would be more likely to return for voice therapy sessions, practice the assigned exercises/strategies, and notice improvements in their voice. To test this hypothesis, a series of binary logistic regression models were conducted to determine whether the motivation construct (i.e., Items 1 and 2) was predictive of the following outcomes: (a) whether the patient returned for the first therapy session, (b) whether the patient practiced the recommended therapy exercises or strategies, and (c) whether the patient noticed improvements in their voice status upon returning for their first voice therapy session. These data points were taken from reports in the patient's medical chart. The first variable was documented as whether or not the patient attended the first therapy session scheduled after the evaluation, and the latter two variables were based on the subjective part of the first voice therapy session report. They were given binary “yes” or “no” ratings for each of these three outcome variables. Only attributes of the first therapy session were used when investigating the predictive validity of the scale to avoid confounding variables related to events from future therapy sessions, other life events, and so forth. The first model was conducted with the entire sample (N = 75), and the second and third models were conducted using data from only the patients who returned for their first voice therapy session (n = 54). Logistic regression models were conducted first using the baseline measurements of Items 1 and 2. To determine whether and to what extent the poststimulability testing measures predicted the outcomes, the logistic regression models were repeated using poststimulability testing scores. All models included diagnosis category as a covariate. To avoid overfitting the logistic regression models, age and sex were not initially included as covariates. However, models that had significant predictors were run again with the significant predictor along with age and sex to determine whether age and sex contributed to the outcome. Any models that had significant age or sex effects are reported.

To assess concurrent validity for the voice perception construct (i.e., Items 3 and 4), we investigated whether self-reported voice quality severity and vocal effort were related to clinician ratings of overall voice severity. We hypothesized that patients with higher overall severity scores by the clinician would also report higher voice severity and vocal effort. These relationships were investigated using a multiple regression model, with the patient-rated voice parameter being the dependent variable and the clinician rating of overall severity being the independent variable, controlling for diagnosis. Relationships were assessed with the standardized beta coefficients.

Assessing Changes in Perception After Stimulability Testing

To answer the question “Do perceptions change following stimulability testing?” we first assessed the normality of the distribution for each item in the survey and for the clinician-scored overall severity. For items that had a normal distribution, a paired-samples t test was implemented to determine within-person changes that occurred after stimulability testing. We judged the effect of the difference based on the Cohen's d effect size for paired-samples tests (i.e., the mean difference divided by the standard deviation of the mean difference; small = 0.20, medium = 0.50, large = 0.80; Cohen, 1988). Wilcoxon signed-ranks tests were conducted to determine pre/post changes for items that had a skewed distribution. Effects were judged using the Wilcoxon effect size r, which is calculated as the z statistic divided by the square root of the sample size. A Bonferroni correction was applied to account for multiple tests (α ≤ .01).

Our next question focused on whether self-reported voice severity and effort are related to confidence in or commitment to voice therapy. This question was addressed using Spearman's rho correlations conducted separately on baseline and poststimulability testing data. Correlations were also conducted between posttesting motivation scores and the change scores for patient-reported voice severity and vocal effort to determine whether the magnitude of change in the sound or feel of the voice was associated with poststimulability testing levels of motivation. Change scores were calculated by subtracting the posttesting score from the pretesting score. Strength of correlations was interpreted as follows: ≤ |0.19| as small, |0.20| to |0.49| as medium, and ≥ |0.50| as large (Cohen, 1988).

Results

Reliability

The four items in the survey were separated by constructs, with each construct consisting of two items. Cronbach's alpha for the motivation construct at the baseline time point indicated fair reliability (α = .53). Cronbach's alpha for the voice construct at baseline indicated good reliability (α = .80). Values were also calculated for each construct for the poststimulability testing time point. These calculations showed acceptable reliability for the motivation construct (α = .71) and good reliability for the voice construct (α = .82).

Content Validity

The nine voice-specialized SLPs unanimously agreed that each of the items appears to test what it aims to test and that items relate to the theoretical concepts for each construct. There was perfect agreement across raters that each item was either essential or useful to assess the given construct. The CVR is a quantitative measure that further tests expert rater agreement regarding whether the item is essential when assessing the construct (Lawshe, 1975). Table 2 shows CVR values for each item. All CVR values met or exceeded the critical value for content validity (Wilson et al., 2012). The CVI for the entire inventory was .72. Thus, the content validity for the inventory was considered acceptable from the SLP perspective.

Table 2.

Content validity ratio (CVR) scores for each item and Content Validity Index (CVI) scores for each construct and the total inventory separated by rater groups.

Inventory parameter	SLP ratings	Patient ratings
	CVR
Item 1: Confidence in voice therapy	.56	.56
Item 2: Commitment to voice therapy	1.00	.78
Item 3: Self-ratings of voice severity	.56	.56
Item 4: Self-ratings of vocal effort	.78	.56
	CVI
Motivation construct	.78	.67
Voice perception construct	.67	.56
Total inventory	.72	.61

Note. SLP = speech-language pathologist.

Similarly, patients had the opportunity to rate whether the items measured the theoretical constructs. The nine patients who participated unanimously agreed that the two items in the motivation construct were either essential or useful to measure the construct. For the voice perception construct, there was near-unanimous agreement that items were either essential or useful. One patient felt that the item related to rating the severity of one's own voice quality was not necessary or useful for measuring self-perception of voice. This patient had a diagnosis of muscle tension dysphonia and reported that how the voice feels is more important than how the voice sounds. Despite this one patient's rating, both items in the voice perception construct met the critical value for content validity (see Table 2 for more details). The overall CVI from the patient group was .61, suggesting that the inventory also demonstrated acceptable content validity in this group.

Predictive Validity

Prediction of return for the first therapy session. The logistic regression model using baseline measures of motivation significantly predicted whether the patient returned for the first therapy session over and above the intercept-only model, χ²(3) = 8.36, p = .0257. Confidence was not a significant predictor in the model (p = .9887), but commitment was a significant predictor, odds ratio (OR) = 1.045, Wald χ²(1) = 5.12, p = .0237. This result indicates that with each 1-mm increase on the 100-mm commitment VAS, the odds of the patient returning for the first therapy session increased by 4.5%.

The same model was run on the poststimulability testing scores. That model also predicted whether the patient returned for the first voice therapy session over and above the intercept-only model, χ²(3) = 9.75, p = .0226. Similar to the model using the baseline measures, confidence did not predict whether the patient returned for therapy (p = .9251), but commitment did, OR = 1.070, Wald χ²(1) = 4.40, p = .0359. The OR indicates that with each 1-mm increase on the commitment scale, the odds of the patient returning for the first therapy session increased by 7.0%. Diagnosis was not a significant predictor in either model (p = .3096 and p = .3158, respectively).

Prediction of voice improvements. The baseline model predicted improvements in the patient's voice status by the first therapy session greater than the intercept-only model, χ²(3) = 14.47, p = .0023. Confidence was a significant predictor in the model, OR = 1.080, Wald χ²(1) = 9.09, p = .0026, indicating that the odds of the patient improving increased by 8.0% for each 1-mm point increase on the confidence scale. Commitment was not a significant predictor (p = .1345).

The poststimulability testing model also significantly predicted improvements in the patient's voice status, χ²(2) = 11.24, p = .0105. Like the baseline model, confidence significantly predicted voice improvements, OR = 1.118, Wald χ²(1) = 5.24, p = .0220, indicating that the odds of the patient improving increased by 11.8% for each point increase in the confidence scale. Again, commitment was not a significant predictor in the model (p = .6743), nor was diagnosis (p = .7941 and p = .7906, respectively). The model was rerun, with only posttesting commitment, age, and sex included to determine whether age and sex contributed to the outcome. Age was not a significant factor, but sex was significant, OR = 5.82, Wald χ²(1) = 4.95, p = .0261. Improvements were reported in 63.6% of women but in only 27.3% of men.

Prediction of practice. The baseline model significantly predicted whether the patient practiced between the initial evaluation and the first therapy session, χ²(3) = 11.25, p = .0105. Confidence significantly predicted whether the patient practiced, OR = 1.065, Wald χ²(1) = 7.94, p = .0048, indicating that each increase of 1 point on the confidence scale was associated with a 6.5% increase in the odds that the patient practiced. Commitment was not a significant predictor in the model (p = .0905). Diagnosis was not a significant predictor in the model (p = .7064). Following stimulability testing, the overall logistic regression model was not statistically significant, χ²(3) = 6.71, p = .0818.

Concurrent Validity

Concurrent validity was assessed by investigating relationships between the patient's perception of voice severity/effort and the clinician's perception of the overall severity of voice quality, controlling for diagnosis. Table 3 shows the standardized beta coefficients and significance values resulting from the multiple regression models. Each multiple regression model was statistically significant. All relationships between patient and clinician ratings at each time point were statistically significant. Diagnosis was not a statistically significant variable in any of the models.

Table 3.

Standardized beta coefficients for the relationship between clinician-rated overall severity and patient-rated voice severity and vocal effort resulting from the multiple regression models, controlling for diagnosis.

Perceptual scale	Clinician-rated overall severity
	Baseline	Poststimulability testing
Patient-rated voice severity	.384**	.314*
Patient-rated vocal effort	.276*	.249*

^*p < .05.

^**p < .01.

Changes in Patient Perceptions

Each item was tested for distribution normality. The items in the motivation construct each demonstrated a distribution that was negatively skewed. Thus, nonparametric tests were conducted to answer our questions regarding whether scores on the motivation items changed following stimulability testing. The items in the voice construct and the clinician-rated overall severity scores were normally distributed, so parametric paired-samples t tests were conducted. Descriptive statistics for items at each measurement time point and mean differences are reported in Table 4. Figure 2 illustrates the within-patient changes across all items.

Table 4.

Means (standard deviations) of each rating scale both prior to (i.e., pretesting) and following (i.e., posttesting) stimulability testing and mean within-person differences (standard deviation of the difference).

Perceptual scale	Pretesting	Posttesting	Mean difference
Confidence	75.99 (21.59)	88.92 (14.58)	12.93 (16.61)
Commitment	86.03 (16.73)	92.20 (11.36)	6.17 (14.15)
Self-reported severity	52.12 (23.34)	43.80 (25.25)	−8.32 (20.18)
Vocal effort	54.81 (24.12)	47.49 (24.18)	−7.03 (22.04)
Clinician-reported overall severity	43.05 (20.53)	28.19 (18.95)	−14.87 (11.59)

Figure 2.

Box plots and visualization of within-patient changes from baseline ratings (pre) to poststimulability testing ratings (post) for (A) confidence in the ability to change through voice therapy, (B) commitment to voice therapy, (C) patient perception of voice severity, (D) patient perception of vocal effort, and (E) clinician ratings of the overall severity of dysphonia. Effect sizes are reported as r for nonparametric tests and Cohen's d for parametric tests. *p < .01, **p < .001.

Item 1 asked participants how confident they are that they can achieve improvements in their voice through voice therapy. There was a statistically significant increase in confidence following stimulability testing (z = 6.01, p < .0001), which represented a large effect (r = .69). Item 2 asked patients how committed they were to voice therapy. There was a statistically significant increase in commitment after completing stimulability testing (z = 4.32, p < .0001), which also represented a large effect (r = .50). In other words, both confidence and commitment increased after stimulability testing.

Patients rated the severity of their voice quality in Item 3. There was a statistically significant decrease in patient-reported voice quality severity (i.e., patients felt their voice quality improved) following stimulability testing, t(74) = 3.57, p = .0006, which represented a small-to-medium effect (d = 0.41). In Item 4, patients rated their vocal effort. There was also a statistically significant decrease in vocal effort following stimulability testing, t(74) = 2.76, p = .0073, which represented a small effect (d = 0.32). Finally, there was a statistically significant decrease in clinician-rated scores of overall severity of voice quality following stimulability testing, t(74) = 11.11, p < .0001, which showed a very large effect (d = 1.28). Although patients noticed a decrease in both voice severity and effort following stimulability testing that was statistically significant, clinicians noticed nearly twice as much improvement in voice quality compared to the patient's self-assessment.

Relationships Between Motivation and Voice

Spearman bivariate correlations revealed negligible/small nonsignificant relationships between both confidence in and commitment to voice therapy and both patient-rated voice severity and vocal effort. There were no changes in the strength or significance of the relationships following stimulability testing. However, there was a medium statistically significant negative correlation between the difference in vocal effort following stimulability testing and the posttesting rating of confidence (ρ = −.28, p = .0141). In other words, those who experienced a larger decrease in vocal effort after stimulability testing reported higher posttesting levels of confidence in their ability to make changes in their voice through voice therapy. See Table 5 for details.

Table 5.

Spearman rho correlations of poststimulability testing motivation scales with poststimulability testing voice perception scales and change scores in voice perception scales.

Perceptual scale	Posttesting confidence	Posttesting commitment
Posttesting self-rated voice severity	−.16	.19
Posttesting self-rated vocal effort	−.06	.08
Δ self-rated voice severity	−.20	.05
Δ self-rated vocal effort	−.28*	−.17

Note. Δ = pretesting-to-posttesting change.

^*p < .05.

Discussion

The primary objective of this study was to determine whether stimulability testing can facilitate improvements in the patient's motivation to participate in voice therapy. To further that objective, we set out to determine whether the patient's awareness of changes in their voice following stimulability testing influenced how motivated they were to participate in therapy. We hypothesized that both motivation and voice would improve following stimulability testing and that perception of their voice severity and effort would predict the patient's level of motivation. To achieve those objectives, we validated the Voice Therapy Motivation and Voice Perception Inventory. Results showed improvements in both motivation and self-reported voice status after stimulability testing.

Development of the Voice Therapy Motivation and Voice Perception Inventory

A readiness ruler is typically a single-item scale that queries a person's readiness for change on a particular goal and is often scored on a 0–10 interval scale (Miller & Rollnick, 2012). Readiness rulers have been used in psychological and health behavior literature to assess motivation to change (Heather et al., 2008; Hudson & Mac Neil, 2020; LaBrie et al., 2005). In the voice disorders literature, the readiness ruler has been used to investigate patients' commitment, confidence, and attributed importance to the goal of practicing voice exercises (van Leer & Connor, 2012; van Leer et al., 2017, 2021). We further modified those questions in the current study to take the form of asking the patient about how committed they are to voice therapy as a whole and how confident they are that they can achieve improvements in their voice through voice therapy. We also modified the rating method from a 0–10 scale to a 100-mm VAS. These decisions were made for three reasons. First, we wanted these questions to be simple and rapid for the patient to answer in the setting of a time-constrained SLP evaluation, particularly because we wanted to obtain scores at two time points during one visit, which is what prompted us to include only two questions versus the three questions in the previous studies (van Leer & Connor, 2012; van Leer et al., 2017, 2021). This is also what led us to refocus the questions to pertain more broadly to voice therapy as a whole as opposed to targeting specific goals. Second, we felt that a continuous variable (i.e., ratings from a 100-mm VAS) would provide more resolution and would not bind the patient to one of the limited options on an interval scale (Klimek et al., 2017). This is particularly relevant in light of the skewed results for the motivation items. We might not have been able to find robust effects if using an interval scale because patients would likely have clustered on one or two numbers, whereas the VAS allowed for a higher degree of precision. Third, we wanted the scales, particularly the self-report voice perception scales, to mirror the scoring method of the well-established CAPE-V auditory-perceptual analysis of dysphonia (Kempster et al., 2009).

Per our hypothesis, we expected that a patient's perception of the sound (i.e., voice severity) and feel (i.e., vocal effort) of their voice would impact their confidence in and commitment to voice therapy, particularly after stimulability testing. Therefore, the inventory we provided to patients included two items related to the perception they have of their voice. One asked the patient to rate the severity of their voice quality, which paralleled the CAPE-V Overall Severity subscale but from the patient's perspective. We also asked the patient to rate their vocal effort. Rating vocal effort on a 100-mm VAS has been conducted in past studies (Crocker et al., 2024; Marks et al., 2021; Shembel et al., 2023; Toles & Shembel, 2024), and we felt that this was the most straightforward way to quantify the feel of the voice.

Reliability Analysis

Even though some of the items in the inventory have been used in previous studies (Marks et al., 2021; van Leer et al., 2017), our modifications to the items for the purposes of this study necessitated a psychometric analysis of the inventory to confirm that the items are reliable and valid. Our first step was to assess the reliability of the inventory. Unfortunately, test–retest reliability was not possible to calculate with this data set. These questions were presented near the beginning of the clinical voice evaluation with the SLP and again shortly after stimulability testing was conducted during the same routine clinical visit. There was not an opportunity to obtain consistent measures because the scores were expected to change due to the introduction of stimulability testing. Therefore, we opted to assess the internal consistency reliability for each of the two constructs that comprised the inventory: motivation and voice perception. For this, we used Cronbach's alpha, which is known to be highly variable based on the number of items in the scale. Lower values were expected since the alpha calculations were conducted on only two items in each construct, but we felt that the brevity of the scale was important to be able to be administered rapidly at two time points in a clinical visit. The analysis found that the internal consistency reliability was fair for the motivation construct but improved to acceptable after completing stimulability testing. There was good internal consistency reliability for the voice construct regardless of time point. The scores for both constructs indicate that the two items within each construct are related to the construct but are not redundant (i.e., they are not assessing the exact same factor). Future work can assess test–retest reliability by potentially obtaining measurements at the visit with the laryngologist and again at the beginning of the SLP evaluation, assuming the SLP and the laryngologist are not conducting evaluations simultaneously (i.e., in a joint clinic format).

Validity of the Motivation Construct

The items included in each construct were judged to have good content validity for the construct being measured by both SLP and patient raters. All nine SLP raters universally deemed the commitment question essential to measure motivation. Across both the SLP and patient groups, raters agreed that both items were either essential or useful to measure the motivation construct. Raters who provided extra commentary on the items stated that they felt that commitment was the best way to measure motivation for voice therapy. Although confidence in the ability to make changes in the voice through voice therapy was not universally seen as an essential question for measuring voice therapy motivation, the CVR exceeded the critical value to be considered acceptable. SLPs who felt that the confidence question was useful but not essential stated qualitatively that “therapy can improve the patient's confidence if they are not (already confident)” and that confidence “can be built as therapy progresses.” These were insightful comments that were supported by our findings. For instance, the goal for this inventory was to assess tenants that could potentially impact the patient's compliance with therapy and to determine if stimulability testing was an ingredient that could target improving those tenants. Confidence significantly improved following stimulability testing, which is sometimes referred to as “trial therapy,” so it stands to reason that confidence would improve with the initial trial of therapy and continue to grow as therapy advances.

These ideas were further illuminated when investigating the predictive validity of the motivation construct. Both items predicted aspects of compliance and improvement. Commitment was a significant predictor of whether a patient would return for therapy. This likelihood increased further after stimulability testing. Based on the OR, for a 10-mm increase on the commitment VAS, the odds of the patient returning for therapy increased by 70%. Interestingly, confidence was not a significant predictor for returning for therapy. However, confidence predicted whether the patient practiced the exercises and/or used generalization strategies taught at the voice evaluation. Confidence also predicted whether the patient reported improvement in voice status by the first therapy session. Those who felt more confident that they can make changes to their voice at the evaluation noticed more improvement by the first therapy session. For each 10-mm increase in confidence following stimulability testing, the odds of the patient noticing improvement increased by 118%. This is a clinically meaningful likelihood, particularly because the average increase in confidence following stimulability testing was nearly 13 mm. Taken together, we feel that the findings suggest that the two items have predictive validity for concepts generally associated with motivation and, in turn, compliance.

Validity of the Voice Perception Construct

SLP and patient raters agreed that the two voice perception items had good content validity. Concurrent validity was assessed by investigating the relationships between clinician and patient perceptions. There were significant relationships between patient-perceived voice severity and clinician-perceived overall severity both at baseline and following stimulability testing. Similarly, there were also significant relationships between patient-perceived vocal effort and clinician-perceived overall severity, although these relationships had smaller effects. It is possible that vocal effort would map better to the Strain subscale of the CAPE-V as opposed to the Overall Severity subscale, but clinicians did not complete the entire CAPE-V subscales at both time points.

Despite the significant relationships, patients rated their voices on average as more severe than clinicians (i.e., average difference of 9 points at baseline). Both patients and clinicians noticed a decrease in severity on average following stimulability testing, but clinicians noticed a mean difference that was nearly twice as large as that noticed by patients (see Table 4). There could be a number of explanations for this discrepancy. The patient might be more critical of their voice than the clinician. Clinicians base their judgments on their internally calibrated system (Nagle, 2022), which involves experience with a range of voice severities. The patient is the one who is experiencing the voice disorder, so it could potentially feel and sound worse to themselves than others perceive. Conversely, the larger improvements that the clinician noticed following stimulability testing could be biased. Clinicians want there to be an improvement with the tasks and cues they provide, so it is possible that they perceived an improvement that was exaggerated compared to how others might perceive it. Similarly, because data were collected retrospectively from clinical practice, the clinicians were not blinded when rating posttesting overall severity. Since they potentially had access to prestimulability scores, they might have marked the posttesting score more favorably.

Other biases might stem from the patient. The patient could have potentially felt pressure to mark an improvement even if they did not feel or hear changes because the clinician was in the room with them. That might account for a smaller degree of change compared to the clinician's observation. It is also possible that a portion of the patients do not want therapy to be the answer (e.g., those who come in hoping for a “miracle pill”) and, therefore, underestimate the effect of stimulability testing. Even in the setting of these potential biases, the results point to the concurrent validity of the voice perception construct.

Perceptions After Stimulability Testing

Motivation Improvements

The next step was to determine whether stimulability testing influenced a change in motivation and/or self-reported voice perception. Our findings indicate that all scales in the inventory improved on average after stimulability testing. Improvements in confidence and commitment had large effects, suggesting that stimulability testing can be a key ingredient to improve a patient's commitment to voice therapy and confidence in their ability to improve their voice.

There was somewhat of a ceiling effect in each of the motivation items, however. For both scales, most patients only marked ratings on the upper half of the scale at each time point, and skewness values ranged from −1.3 to −2.1. For this reason, we decided to use nonparametric statistical tests to determine differences, which is what we would also recommend for future studies. The ceiling effect was most noticeable on the commitment scale. This could suggest that the patients who agreed to participate in the SLP voice evaluation already felt relatively committed to and confident in voice therapy. Those who are less committed and confident might not have attended the evaluation in the first place. Nevertheless, even with the high commitment and confidence scores at baseline, patients still reported even higher levels of commitment and confidence following stimulability testing.

Self-Reported Voice Improvements

A major part of stimulability testing is to facilitate the patient's awareness of their voice as a first step in the process of the patient regaining control over their voice. Our findings show that patients, on average, perceived improvements in the sound and feel of their voice after stimulability testing. This finding supports the hypothesis that stimulability testing results in immediate voice changes. However, the effect sizes were smaller than those for the motivation construct. This data set included a host of different diagnoses and severity levels. It is inevitable that some diagnoses and/or clinical presentations will not permit as much immediate change as others. For instance, someone with an immobile vocal fold with a large glottal gap will not likely be able to achieve as much improvement in the sound of the voice as someone with a functional disorder or even someone with an immobile vocal fold with a small glottal gap. A recent study by our group found that voice diagnosis and dysphonia severity were predictive of whether or not patients are stimulable for improvement (McDowell et al., 2023). That study showed that patients with lower overall severity of dysphonia scores and those with functional disorders were more likely to be stimulable. Considering that, it is possible that these factors are playing a role in how much change the patient will perceive.

It is also possible that some patients have better auditory or somatosensory perception than others (Weerathunge et al., 2022). Literature suggests that altered auditory and/or somatosensory feedback systems occur in a host of voice disorder populations, including individuals with both phonotraumatic (i.e., vocal fold mucosal lesions) and nonphonotraumatic (i.e., muscle tension dysphonia; Abur et al., 2021; Ziethe et al., 2019) vocal hyperfunction, laryngeal dystonia (Haslinger et al., 2005; Simonyan & Ludlow, 2010; Thomas et al., 2021), and dysphonia associated with Parkinson's disease (Abur et al., 2018). Those with aberrant internal feedback systems would be less likely to be aware of changes in the sound or feel of the voice. Similarly, interoception (i.e., awareness of one's internal physiological state and associated emotional and regulatory responses; Calì et al., 2015) could also be a factor in the precision of identifying physiological voice changes. A recent study purported that patients with functional voice and upper airway disorders (i.e., muscle tension dysphonia, paradoxical vocal fold motion disorder, and chronic cough) might be more likely to have lower attention regulation, less emotional awareness, and difficulty processing their body sensations than vocally healthy individuals based on scores on a self-report interoceptive awareness scale (Smeltzer et al., 2023). Therefore, we hypothesize that individuals with impaired auditory or somatosensory feedback systems or less interoceptive awareness, regardless of their diagnosis, would not be able to perceive as large of a change in the sound and/or feel of the voice. Future research should investigate the link between sensory feedback, interoceptive awareness, and voice stimulability, as the current study is insufficient to support this hypothesis.

Influence of Voice Change on Motivation

We hypothesized that improvements in patient-reported voice severity or effort would facilitate improvements in motivation for voice therapy. Surprisingly, there were no relationships between the voice and motivation items either at baseline or in posttesting. We then investigated whether the amount of change the patient noticed from baseline to posttesting would impact their posttesting motivation levels. This found a medium, statistically significant negative correlation between the change in vocal effort (i.e., posttesting scores subtracted from baseline scores) and the posttesting confidence score. This result indicates that patients who experienced a larger decrease in vocal effort following stimulability testing felt more confident that they would be able to make voice improvements through voice therapy. There was a small correlation between voice severity change and confidence, but it was not significant. Thus, it is perhaps more important for the patient to be able to feel the change in their voice to cultivate confidence.

There are a couple of potential reasons why confidence and commitment might improve after stimulability testing regardless of whether or not the patient notices a change in the sound of the voice. First, there is a possibility that stimulability testing affects the patient's locus of control. Patients who report higher voice-specific perceived control are more likely to experience less voice handicap and fewer concerns about the voice therapy process (Nguyen-Feng et al., 2021). Perceived present control is the degree to which one believes a particular stressor is currently under their control (Frazier et al., 2011). Considering this, combined with our findings, stimulability testing might influence the patient's present perceived control related to their voice. In other words, stimulability testing provides something on which the patient can rely that instills a feeling of control over their voice problem. Thus, stimulability testing has the potential to empower the patient, encouraging them that they can make a change through voice therapy. Second, the stimulability testing process requires the patient and the SLP to work closely together to be able to achieve improvements. This inevitably increases the rapport that the patient has with the clinician. This is an opportunity for the clinician to act as a partner or facilitator in the process. For some, having another person's support, encouragement, and praise is enough to believe in the ability to make changes (van Leer & Connor, 2010). This support might impact motivation regardless of whether the voice sounds or feels different.

Clinical Implications

The two constructs that comprise the inventory measure distinct concepts. We acknowledge that the two constructs could be used separately. We opted to combine the two constructs into one inventory because we felt that it was crucial to measure both motivation and perception of voice in an effort to understand how the patient's assessment of their own voice is related to their motivation. The psychometric analyses were conducted on each construct separately, which could be considered a benefit. Because each construct was found to be internally reliable and valid, the two constructs can be measured independently at the clinician's discretion. However, we recommend that clinicians use all items in the inventory to achieve the most insight regarding the patient's perspective.

Although not tested in this study, it is possible that this inventory could be beneficial to use serially throughout the course of therapy. It could be informative to the clinician to know whether the patient's motivation and perception of their voice status has waxed, waned, or plateaued. The clinician could use this information to assist in decision making when determining whether to continue therapy sessions. Future work could target use of these scales across therapy sessions to determine their ability to monitor therapy progress and inform decision making.

Commitment increased the likelihood of coming to therapy but not practicing at home. Confidence increased the likelihood of practicing at home but not coming to therapy. Thus, a “committed” person may show up but not practice. A “confident” person may practice at home but not feel the need to return for more instruction. Clinicians could consider emphasizing the importance of returning for additional instruction to the confident patient. Committed but not confident patients should be encouraged to practice at home. For these patients, stimulability testing is a prime opportunity to highlight mastery experiences with trial tasks to strengthen their confidence in practicing their home practice program and generalization techniques.

Limitations

As with any study, this study has its limitations. It is possible that there was a response bias that would influence the patient to answer in a way that seems more favorable to the clinician (Furnham, 1986). Because these inventories were completed with their treating SLP present, the patient might have felt pressured to indicate that they are highly motivated to participate in voice therapy in order to please the SLP. Furthermore, this bias might have encouraged the patient to report even higher confidence and commitment following stimulability testing. To minimize this bias, future work should incorporate blinding of survey responses, and surveys should be administered without the treating SLP in the room.

One limitation when studying stimulability testing is that it is difficult to control for informal clinician practices that might impact results. For example, the clinician is undoubtedly going to provide verbal affirmations to the patient based on how they sound during the stimulability testing process. This is likely to increase patient–clinician rapport. Although these informal clinician practices are not technically in the definition of stimulability testing, it is clear from prior work that they are thought to be a part of the process (Toles & Young, 2023). In fact, we argue that the informal clinician practices when presenting trial tasks might be more important than just trial therapy tasks alone in improving the patient's motivation and voice perception.

As for the validity of the motivation construct, we acknowledge that there are a multitude of reasons patients might not return for therapy sessions. Insurance denials, time commitment, and illness, among others, are potential reasons unrelated to motivation that a patient might decide not to pursue therapy (Weston et al., 2023). This study was limited by the fact that the medical charts for patients who canceled or did not attend their voice therapy session did not reliably provide reasons for the missed appointment. Despite the various reasons one might not attend their therapy session, the commitment question still predicted voice therapy attendance.

The aim of this study was to investigate patient and clinician perceptions. Thus, we did not include instrumental assessment methods, such as acoustic or aerodynamic measurements. This could be considered a limitation, as a more objective metric to correlate with the patient's perception of voice might provide a different perspective that is less prone to bias. Future studies could consider comparing objective instrumental measurements before and after stimulability testing to determine the benefits of stimulability testing on objective outcomes.

The current study did not investigate the clinician's subjective report of whether or not they felt the patient was stimulable. It did, however, investigate clinician ratings of overall voice severity before and after stimulability testing. Using those scores (visualized in Figure 2E), there was a very large effect indicating that the overwhelming majority of patients improved. Yet, not all patients improved. Thirteen patients had less than 5-mm decreases in overall severity (i.e., essentially no change) or slight increases in overall severity after stimulability testing. Despite minimal to no improvements, 11 of these 13 patients were still recommended for voice therapy sessions. The two patients for whom therapy was not recommended had lower scores on confidence following stimulability testing. Future work should explore how clinician ratings affect their clinical decision making and, in turn, whether the clinician's decisions impact the patient's confidence.

This study originally stemmed from a clinical idea. The treating SLPs administered these questions during routine clinical practice to improve objective documentation of motivation and self-assessment of voice and to determine how well they responded to stimulability testing. Prior to implementing this inventory, documentation included only subjective reports that had historically been conducted in the clinic. Therefore, data collection was conducted in a retrospective manner. This limited our ability to conduct test–retest reliability and did not control for biases. Although the SLPs at the UTSW Voice Center all have similar practice patterns, there was not a standardized stimulability procedure that was conducted across SLPs. Practice pattern variations could potentially influence how the patients respond to stimulability testing across clinicians. Furthermore, the retrospective nature of the study precluded our ability to include a control group, which is a confounding factor when considering the findings of the study and limits the ability to calculate test–retest reliability. It is possible that merely undergoing an assessment of any kind could affect the patient's sense of motivation or voice perception. Future work should initiate a prospective study to control for practice patterns, limit potential biases, investigate the effect of stimulability testing on long-term voice therapy outcomes, and include a control group that does not receive stimulability testing for comparison.

Conclusions

This study validated a Voice Therapy Motivation and Voice Perception Inventory that was used to determine changes in patient perceptions of motivation and voice status following stimulability testing. Results found that confidence in the ability to make changes through voice therapy and overall commitment to voice therapy improved following stimulability testing. There were also posttesting improvements in both the voice severity and the vocal effort perceived by the patient. Although there was a relationship between the degree of change in vocal effort and posttesting confidence scores, there were no relationships between patient-perceived voice severity and motivation. This suggests that motivation improves following stimulability testing regardless of whether or not the patient notices a change in the sound of the voice. Thus, stimulability testing is a key factor in the overall aim of improving motivation for voice therapy and should be considered a necessary component of the behavioral voice evaluation.

Data Availability Statement

The data sets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Funding Statement

This work was supported in part by National Institute on Deafness and Other Communication Disorders (NIDCD) Grant K23DC020758 (principal investigator: Laura E. Toles). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIDCD or the National Institutes of Health.