Early Insights Into the Role of Personality in Adherence to Voice Rest After Phonomicrosurgery

Summary:

Objectives/Hypothesis.

The purpose of this preliminary investigation was to explore the relationship between personality and adherence to voice rest in individuals following phonosurgery for benign mucosal or submucosal/lamina propria lesions. We hypothesized that adherence to voice rest would be inversely related to Extraversion and directly related to Conscientiousness, two superfactor traits of the Big Five personality model.

Study Design.

This was a prospective, longitudinal, cohort study.

Methods.

Twenty-five patients (eight men, 17 women; mean age = 46.25 years; $SD=14.38$) of Houston Methodist Hospital’s Department of Otolaryngology, Head and Neck Surgery underwent microflap phonosurgery. Each participant provided demographic information, completed the Ten Item Personality Inventory (a brief Big Five personality instrument), and wore a vocal dosimeter to objectively measure voice rest adherence during a seven-day, postoperative period.

Results.

The participants’ mean personality scores were higher than reported norms across all Big Five traits. Contrary to our hypotheses, (1) Extraversion related directly to voice rest adherence, and (2) no other personality traits (including Conscientiousness) related to voice rest adherence.

Conclusions.

Although extraverts have been described as talkative, assertive, and outgoing, we found that high Extraversion was related to less vocalizing during the voice rest period. Our results are compatible with the health psychology literature that describes a positive relation between Extraversion and adherence to preventive healthcare practices. Future research should employ a larger, single-sex sample, potentially with greater similarity between patient diagnoses. Clinically, such data could inform counseling approaches to facilitate voice rest adherence and potentially improve surgical outcomes.

INTRODUCTION

Postsurgical voice rest is frequently recommended for patients with benign vocal fold lesions to minimize the formation of scar tissue at the site of the lesion.^1,2 Voice rest recommendations are not standardized; they vary across practitioners in the number of days and amount of recommended voice use.^1–4 The two most common forms of voice rest are complete and relative.¹ Complete voice rest requires absolute silence for the prescribed duration.¹ Relative voice rest may entail a few days of complete voice rest and/or limited voice use during the recommended period.¹ The type and duration of voice rest varies based on the physician’s training and experience, location of the lesion on the vocal folds, and degree of surgical dissection.¹

Seven days of complete voice rest has been the most common recommendation. Animal studies have demonstrated the utility of complete voice rest for the first three days post surgery.^5–7 However, surgeon recommendations vary from zero to fourteen days for complete voice rest and zero to thirty-five days for relative voice rest.¹ Greater adherence to voice rest recommendations has been associated with better surgical outcomes^2,3,8 in relation to improved mucosal healing and minimized scarring.^8–11 Yet, poor adherence to voice rest recommendations has been reported, with adherence rates such as 39.5%,¹² 34.5%,¹³ and 0.0%.¹⁴

Barriers and facilitators to voice rest adherence

Adherence is defined as a person’s willingness and ability to execute recommended health changes.^15–17 There is limited evidence regarding barriers and facilitators to postsurgical voice rest adherence. Dong and Lin¹⁸ and Barna et al¹⁹ have both addressed this important topic. Dong and Lin¹⁸ reported four adherence facilitators in their qualitative study of patients who completed voice rest: (1) prenotifying contacts of voice rest, (2) using nonverbal communication, (3) understanding the rationale for voice rest, and (4) knowing the consequences of nonadherence. Barriers to voice rest adherence in that study were the social disconnect caused by voice rest and demands for voice use at work.¹⁸ Barna et al¹⁹ examined surgical outcomes of phonomicrosurgery and found that singers (performers, singing teachers, and singing students) adhered to voice rest more regularly than nonsingers. Synthesizing the results of these two studies, the role of occupation in adherence could differ between singing and nonsinging occupational voice users. Dong and Lin provided limited information on their participants’ occupations, but at least one participant had an office job (nonoccupational voice user), and at least one other individual worked in customer service (nonsinging occupational voice user). The participants stated that the office job was more amenable to voice rest, and the customer service job was less so. In contrast, Barna et al’s singers (singing occupational voice users) adhered with greater regularity than nonsingers.¹⁹ However, this comparison is cautiously interpreted because neither research group stated their criteria for adherence versus nonadherence; a one-to-one comparison cannot be established. In sum, postsurgical voice rest is frequently recommended and thought to positively impact outcomes but poorly adhered to, and there is little evidence regarding barriers and facilitators to adherence.

Personality and voice disorders

Personality factors may influence adherence to voice rest. In adults, personality traits are relatively stable and durable inclinations toward certain behavior.^20,21 Various personality theories exist, and two of the current leading personality theories are the Big Five²² and Big Three taxonomies.^23,24 The Big Five personality traits are Extraversion, Neuroticism, Conscientiousness, Agreeableness, and Openness, described as follows by John et al.²² Those who score high on Extraversion are typically talkative and assertive; those low on this trait are frequently quiet and reserved. High Neuroticism corresponds to feeling tense or anxious versus stable and calm (low Neuroticism). People with high Conscientiousness scores are likely organized and thorough, and those with low scores may be careless and disorganized. High Agreeableness characterizes one who is sympathetic and kind instead of fault-finding and cold (low Agreeableness). Finally, imaginative people with wide interests tend to have higher scores on Openness than those with narrow interests and poor imagination. The Big Three has been described using three higher-order dimensions, Positive Emotionality (PEM), Negative Emotionality (NEM), and Constraint (CONST).^23,24 PEM is analogous to Extraversion, NEM to Neuroticism, and CONST corresponds to Conscientiousness.²⁵ Each superfactor encompasses three or four lower-order traits, which are facets of the higher-order dimensions, such as Wellbeing and Social Potency for PEM, Stress Reaction and Aggression for NEM, and Control and Harm Avoidance for CONST.²³

Investigations grounded in the Theory of the Dispositional Bases of Functional Dysphonia and Vocal Nodules (ie, the Trait Theory)²⁶ have generally tied personality traits to vocal behavior for two patient populations. The first population is individuals with vocal fold nodules (VFN),^25–29 a pathology that fits within the umbrella category of phonotraumatic voice disorders (PVDs). The second is people with the voice disorder(s) variously labeled as functional dysphonia, primary muscle tension dysphonia, functional voice disorders, and/or non-phonotraumatic vocal hyperfunction (NVH).^30–39 People with PVDs/VFNs are prone to high Extraversion (ie, compelled to interact with others) and low Constraint (ie, are impulsive).^25,28,29 Conversely, people with NVH tend to have low Extraversion (ie, are behaviorally inhibited) and high Neuroticism (ie, are emotionally reactive).^25,28

Personality and adherence in voice rehabilitation

Despite the prominence of personality in the psychology-voice disorder literature, few studies have examined the role of personality in adherence to voice therapy or postsurgical voice rest. A recent investigation examined the influence of personality on general voice therapy adherence in a cohort of patients ($n=247$) with diagnoses such as vocal strain (43% of the sample), reflux (36%), or benign lesions (21%).⁴⁰ In these patients, Extraversion was directly tied to barriers to general voice therapy adherence (not related specifically to voice rest). This finding implicates personality as a potential factor in adherence to postsurgical voice rest. Yet, extant explorations of personality and voice rest are scarce. Van Leer and Connor⁴¹ studied people with a range of voice diagnoses to gather patient perceptions of facilitators of and barriers to voice therapy in general, not specifically voice rest. Two of their participants stated that being a “talker” was a barrier to adherence; the authors labeled this proclivity “Extraversion.” Friedman et al’s⁴² theoretical paper also addressed postsurgical voice rest in part. They theorized that high Extraversion and Neuroticism may interfere with adherence to both complete and relative voice rest. While not related to adherence, Toles et al²⁹ found that healthy singers’ vocal doses positively correlated with Wellbeing and Social Potency (subtraits of PEM/Extraversion).²³ This small corpus of personality-adherence data pertaining to general voice therapy, voice rest, or vocal dose seems to indicate that high Extraversion interferes with vocal rest. Yet, clearly, a gap remains in our understanding of the relationship between personality and adherence to voice rest. With more evidence, medical teams could adjust voice rest interventions to improve patient adherence. One potential adjustment is to counsel individuals high in Extraversion to greater awareness of their talkativeness triggers.

Personality and adherence in healthcare

The potential for personality traits to predict or influence adherence is widely acknowledged in other health-related disciplines.⁴³ Findings from these disciplines can inform hypotheses regarding possible relationships between personality and voice rest. For instance, in medicine protocols, good adherence was related to high Conscientiousness^44–48 and high Agreeableness.^44,46,48,49 Poor adherence to ingesting medicine correlated with high Neuroticism or high Extraversion.^{44,46,48–50} Surprisingly, greater Openness was tied to worse adherence.⁴⁴ Greater Conscientiousness and Openness correlated with better adherence to postsurgical limb rehabilitation.⁵¹ Perhaps most germane to postsurgical voice rest—a time when the vocal folds are especially vulnerable to further injury—personality has been associated with implementing preventive health behaviors. High Conscientiousness positively predicted performance of preventive behaviors in musculoskeletal health science.⁵² High Extraversion was related to frequent performance of self-care behaviors in people with diabetes⁵⁰ and participation in mammograms and breast self-examination.⁵³ Greater Agreeableness, Conscientiousness, and Extraversion scores characterized those with good adherence to COVID-19 protocols.⁵⁴ These are only a few of many examples from multiple health domains. They sharpen the impetus to explore personality in relation to voice rest adherence. Pertinent to hypothesis generation, the cited publications from medication and preventive health indicated that people with higher Conscientiousness and Agreeableness scores would be more adherent than those who scored lower in these traits, and that Extraversion’s impact could either be positive or negative.

Measures of voice rest adherence

Voice rest adherence has been measured using patient report and dosimetry. Generally, patient report measures are found to be less reliable than objective measures.^55,56 Related to general voice use (not voice rest), patients and healthy individuals who wore a dosimeter and also rated their amount of phonation on a visual analog scale (VAS) overestimated their voice use.⁵⁷ Related to voice rest, Misono et al¹⁴ used a VAS and dosimeter to measure voice rest in healthy participants. The measures were moderately correlated; some participants overestimated, and others underestimated. Other groups have only used a VAS^12,13 to measure adherence, and the lack of an objective measure could lead to validity concerns. Neither Dong and Lin¹⁸ nor Barna et al¹⁹ stated how they measured adherence. There is a clear need for dosimetry data in the study of voice rest adherence for improved understanding of this behavior.

Purpose of study

The purpose of this preliminary investigation was to explore the relationship between personality and adherence to voice rest following phonomicrosurgery for benign mucosal or submucosal/lamina propria lesions. Hypotheses were generated using the definitions of the Big Five personality traits, the Trait Theory of Voice Disorders, and personality-adherence literature from voice science and other health domains. The hypotheses were as follows: (1) Conscientiousness will have a direct relationship with voice rest adherence based on (a) the rule-following or disciplined nature of those who score high on this trait^20–22 and (b) consistent evidence that Conscientiousness correlates with adherence in other health disciplines. (2) Extraversion will have an inverse relationship with adherence due to (a) the drive to socialize or talk that characterizes individuals who score high on this trait,^20–22 (b) the correlation between high Extraversion and occurrence of PVDs (the Trait Theory), and (c) previous voice science evidence that Extraversion interferes with general voice therapy adherence and perhaps voice rest adherence.^40,41 However, as previously reported, Extraversion variably related positively or negatively with adherence in other disciplines. Hence, the alternative hypothesis for Extraversion is that it will positively impact adherence, as observed in preventive healthcare practice data. Because Agreeableness, Openness, and Neuroticism were not consistently relevant in the pertinent Trait Theory investigations,^25,28,29 previous literature on personality and adherence to voice therapy,^40–42 or personality-adherence evidence from other health disciplines (see above), they were not anticipated to be important factors in the present study.

MATERIALS AND METHODS

Participants

The study was approved by the Institutional Review Board at the Houston Methodist Hospital (IRB# PRO00025881) and the University of Houston (IRB# SITE00000272). Participants in the study were eight men and 17 women with a mean age of 45.4 years ($SD=12.0$, range: 31–70) and 47.1 years ($SD=15.3$, range: 21–69 years), respectively. As part of a previous study,⁵⁸ participants were randomly assigned to one of two voice rest groups (complete voice rest or relative voice rest, detailed below) controlling for sex (male or female) and lesion type. Table 1 is a breakdown of the participants according to voice rest group for the demographic categories age, sex, and occupational voice use. All participants were English-speaking patients of Houston Methodist Hospital’s Department of Otolaryngology, Head and Neck Surgery who received microflap phonosurgery for a primary diagnosis of a benign vocal fold lesion. Individuals with a neurological or cognitive diagnosis, significant hearing or vision impairment, or a history of voice rest were excluded. The groups were asked to complete either seven days of complete voice rest or three days of complete voice rest followed by four days of incrementally increasing voice use, which was labeled “relative voice rest” (Table 2). In the absence of a standard protocol for relative voice rest, we elected to use a time-based definition, as employed in other studies of relative voice rest.^59,60 Other researchers exploring this topic have defined relative voice rest with voice exercises¹¹ or other phonatory adjustments.^59,60 We chose to use time because it is a consistent unit of measurement that allowed us to balance the specificity of our instructions with practicality of patient execution of the recommendations. Patients in the relative voice rest group were asked to use a comfortable pitch and loudness level and maintain at most an arm’s length distance from their conversational partners. They were advised to limit their total speaking time to the day’s prescribed amount (eg, five minutes in the morning and five minutes in the evening of Day 4). Additionally, they were asked to refrain from vocalizations such as throat clearing and coughing that could further injure their vocal folds, and they were told to avoid whispering because both effortful whispering and whispered running speech may lead to vocal fold contact.^61,62 Finally, for purposes of feasibility, and to replicate clinical conditions, the patients were not required to monitor their duration of voicing in conversations in relation to the overall length of the conversation. This relative voice rest protocol was based on current clinical practice and existing evidence, but it was not standardized. Therefore, one should exercise caution when implementing it.

TABLE 1.

Demographic Data By Voice Rest Group

		Complete voice rest	Relative voice rest
All participants ($n=25$)		$n=12$	$n=13$
Mean age ($SD$)		46.0 years (13.47)	47.07 years (16.14)
Sex	Men	4	4
	Women	8	9
Pathology	Polyp	8	5
	Cyst	3	6
	Nodules	1	1
	Reinke’s edema	0	1
Occupational voice user status	Yes	8	4
	No	3	4
	No response	1	5

TABLE 2.

Voice Rest Instructions

Group	Voice rest schedule	Voice rest instructions
Complete voice rest	Days 1–7	No voicing
Relative voice rest	Days 1–3	No voicing
	Days 4–7	Day	Morning	Evening
		4	5 minutes of voicing	5 minutes of voicing
		5	10 minutes of voicing	10 minutes of voicing
		6	15 minutes of voicing	15 minutes of voicing
		7	20 minutes of voicing	20 minutes of voicing
Both	Days 1–7	Use alternate communication (paper and pen, text-to-speech app). No whispering, throat-clearing, or coughing (avoid as much as possible, substitute with soft glottal attack if necessary).

Notes: During prescribed voicing, patients using relative voice rest were asked to maintain conversational pitch and loudness at an arm’s length distance to the listener (at most) and avoid noisy environments.

Measures

Independent variables

Prior to surgery, the patients provided demographic information (age, sex, and occupation) and completed the Ten Item Personality Inventory (TIPI),⁶³ which is a brief measure of the Big Five personality traits^20,21,64 of Extraversion, Emotional Stability (the inverse of Neuroticism⁶⁵), Conscientiousness, Agreeableness, and Openness using two items per trait. Each item completes the statement, “I see myself as…” For instance, one Extraversion item is, “I see myself as extraverted, enthusiastic.” A complete item for Conscientiousness is, “I see myself as dependable, self-disciplined.” Higher scores on Emotional Stability correspond with lower Neuroticism. Participants indicate their degree of agreement with each item by choosing a rating on a 1–7-point Likert scale. Choosing “1” means “disagree strongly,” and selecting “7” means “agree strongly.” Higher scores indicate greater levels of the corresponding personality trait. The TIPI has good convergent and discriminant validity with three longer personality tests: the 44-item Big-Five Inventory,^22,66,67 the 60-item NEO Personality Inventory, Revised,^68,69 and the 100-item trait descriptive adjectives questionnaire by Goldberg.^63,70 The TIPI approximates these longer questionnaires in test-retest reliability, patterns of external correlates, and convergence between self and observer ratings.⁶³ It was normed using scores from 175,981 women and 130,737 men with 15 different primary cultural or racial identifications located in 242 different countries.⁷¹

Dependent variable

Duration of voice use quantified adherence to voice rest protocols. Participants wore a vocal dosimeter [VocaLog^™ Vocal Activity Monitor (Griffin Laboratories, Temecula, California)] to objectively assess duration of voice use during the seven-day voice rest period. This dosimeter measures voice behavior in both normal and disordered voices^14,72–76 by registering the presence of phonation once per second. It captures the intensity and duration of voice use⁷⁷ directly from the signal without retaining raw speech or voice data, which protects participant privacy. Daily calibration is not required, and the device can store data for up to 21 days,¹⁴ making it ideal for this longitudinal study. The dosimeter does not measure vibration dose or differentiate between phonation for speech and vegetative sounds. However, vibration dose was not a variable of interest for adherence, and participants were asked to refrain from vegetative sounds when using voice rest during the data capture time frame.

Procedures

Participants were recruited during the initial clinical assessment with a laryngologist following a phonomicrosurgery recommendation. Prior to surgery, participants provided demographic information and completed the TIPI. At this time, an investigator demonstrated placement and use of the dosimeter, observed and corrected the participants’ placement of it, and answered all questions. Immediately following surgery, dosimeter instructions were reviewed, and the participant received an information sheet with a photograph of the dosimeter placement and specific voice rest protocol (Table 2). The patients were instructed to wear the device during all waking hours, except when engaging in water-related activities such as showering and swimming. Dosimeter activity could not be tracked in real time, but patients self-reported wearing the dosimeter as instructed.

Data extraction

TIPI questionnaires were scored using Gosling’s instructions⁷¹: (1) recode reverse-scored items and (2) average the standard and recoded items for each personality trait. Voice use data were extracted from the dosimeters using the VocaLog2^™ proprietary software.

Data preparation

For a previous study, the complete and relative voice rest groups’ data were tested for between-group differences in duration of voice use.⁵⁸ However, no personality hypotheses were investigated. No significant differences in duration of voice use or clinical outcomes (auditory-perceptual, acoustic, and aerodynamic) were found between the groups ($\alpha <0.05$). A detailed discussion of the dosimetry data is beyond the scope of this paper. However, the data indicated that the complete voice rest group (1) used their voices (against recommendations) and (2) vocalized in quantities comparable to the relative voice rest group. Therefore, for the current study, we collapsed the complete and relative voice rest groups into a single group. We created a new variable, “nonadherence,” to account for the different prescribed amounts of voicing in the complete versus relative voice rest groups. To calculate nonadherence, we subtracted each day’s amount of prescribed voicing from the participant’s dosimeter readings. For example, if relative voice rest Participant A voiced for 11 minutes on Day 4 (10 minutes of prescribed voicing, Table 2), the “nonadherence” variable would be 1 minute, interpreted as 1 minute of voicing in excess of the prescribed voicing amount. If nonadherence values were negative, they were rounded to zero. For example, if relative voice rest Participant B voiced for 7 minutes on Day 4 (10 minutes of prescribed voicing, Table 2), the raw transformed dosimeter outcome would be −3 minutes.

This value would be rounded to zero and interpreted as adherent to recommendations. Any duration of voicing in the complete voice rest group was considered “nonadherence” due to their voicing prescription.

Data analysis

First, descriptive statistics were calculated for the TIPI scores. Then, inferential tests were computed. Because each patient’s voice use was measured daily, the data structure consisted of daily measurements nested within individual participants (Figure 1). Given the nested data structure (Figure 1), an intercept-only model was calculated to assess the proportion of variability that could be attributed to individual differences. The interclass correlation of the intercept-only model was 0.5978, indicating that 59.78% of the total variance in voice use could be attributed to between-person differences. To further assess individual variability, we examined the distribution of nonadherence between male and female participants (Figure 2). Two male participants had substantially greater voice use durations compared with the rest of the cohort. Our consideration of these participants’ data is described later in this section. The significant variance component justified using hierarchical linear modeling (HLM) to account for the hierarchical data structure. HLM was used to analyze changes in adherence levels with the five personality traits entered independently as predictors in separate models ($p<0.05$). The use of HLM allowed the analysis to explain individual differences in adherence levels while considering the entire sample.⁷⁸ Maximum likelihood estimation was used to manage missing data.^79,80 The PROC MIXED procedure in the SAS software (version 3.81) was used for data analysis. This analytical approach aligned with the purpose of this research, ie, to examine whether variations in adherence could be explained by personality traits.

FIGURE 1.

Data structures and predictors at each level.

FIGURE 2.

Violin plot of voice rest nonadherence, expressed as duration of voice use, by sex.

We employed the random intercept model with effect coding of days. The sex variable was dummy-coded; men were the reference group. Each day of the voice rest protocol was effect-coded, and Day 3 was the reference day. This parameterization allowed us to isolate the effect of each specific day and contrast it against the grand mean (the average voice use per patient during the seven-day protocol). The coefficient for each day represented the difference between the voice use that day and the grand mean. The model implicitly estimated Day 3 (reference day), which was calculated using the following formula:

$${\text{Day}}_3=\text{intercept}+\sum _{k=1,k\ne 3}^7(-1)\cdot {\text{Day}}_k$$

The level 2 predictor personality traits were centered at their grand means.⁸¹ We examined the assumption of homogeneity of regression and found that Extraversion influenced voice use differently across sexes. Specifically, the difference in intercepts for male and female participants was 81.98 minutes ($p<0.05$), indicating that, on average, male participants used their voices 82 minutes more than female participants. The interaction between sex and Extraversion was significant ($\beta =-46.30,p=0.05$), suggesting that Extraversion influenced adherence levels more strongly in men than women. Because two male participants had greater voice use compared with the remainder of the cohort, we conducted a sensitivity analysis by re-running the models both with and without these two men (participants 7 and 21). Overall, trends in the data remained consistent, though on a smaller scale, which indicated that the sex-based differences in nonadherence were not solely driven by these two participants. Further, hour-by-hour dosimeter readings confirmed that both participants recorded actual voice use, which validated their data. Therefore, these two men were retained in the final analysis. Given these findings, we included both sex and the interaction term sex · personality in Model B and the supplementary models. This approach allowed us to account for the differing effect of personality traits on voice use behavior across the sexes. It removed the variability in the data due to mean differences across all participants from day to day.

Model A (Reference Model):

Adherence for each patient over time with no consideration of personality traits.

Model A-Level 1:

$$Y_{id}={\beta }_{0i}+\sum _{k=1}^{d-1}{\beta }_k\cdot {\text{Day}}_{ik}+r_{id},$$

$$\text{where}k=\left\{\begin{array}{c}1,\text{for}\text{day}k\\ 0,\text{for}\text{day}\ne k\text{and}k\ne 3\\ -1,\text{for}\text{day}=3\end{array}\right.$$

Model A-Level 2:

$${\beta }_{0i}={\gamma }_{00}+u_{0i}$$

Model A was designed to analyze variations in average voice use across different days. The intercept of the models was the overall nonadherence over seven days (the grand mean). In Model A-Level 1, $Y_{id}$ was the observed voice use for patient $i$ on Day $d.{\beta }_{0i}$ denoted the random intercept for patient $i$, showing the grand mean for voice use over seven days for this patient. ${\beta }_k$ represented the fixed effect of days, where each ${\beta }_k$ was the difference between the mean voice use on Day $k$ and the grand mean across all days, excluding Day 3 (the reference day). ${\text{Day}}_{ik}$ was the effect-coded variable for Day $k$. The residual term $r_{id}$ captured the degree to which observed nonadherence for patient $i$ on Day $d$ differed from what was predicted for the patient on the day given the grand mean of voice use (${\beta }_{0i}$) and the effect of time (${\beta }_k\cdot {\text{Day}}_{ik}$). The $r_{id}$ variable was assumed to be normally distributed with constant variance (${\sigma }^2$) and was assumed to be independent across time within- and between-persons.

In Model A-Level 2, ${\gamma }_{00}$ represented the overall average voice use across all participants on Day 3, and $u_{0i}$ indicated the difference between patient $i$’s expected voice use on Day 3 and expected use across all patients (${\gamma }_{00}$). The $u_{0i}$ were assumed to be normally distributed with mean zero and variance of $t_0^2$. The two sources of random variation in the data, $u_{0i}$ and $r_{id}$, were further assumed to be independent of one another.

Model B:

Adherence for each participant over time with two personality traits as predictors.

Model B-Level 1:

$$Y_{id}={\beta }_{0i}+\sum _{k=1}^{d-1}{\beta }_k\cdot {\text{Day}}_{ik}+r_{id},$$

$$\text{where}k=\left\{\begin{array}{c}1,\text{for}\text{day}k\\ 0,\text{for}\text{day}\ne k\text{and}k\ne 3\\ -1,\text{for}\text{day}=3\end{array}\right.$$

Model B-Level 2:

$${{\beta }_{0i}=\gamma }_{00}+{\gamma }_{01}(\text{Extraversion}{)}_i+{\gamma }_{02}(\text{Sex})+{\gamma }_{03}(\text{Sex}\cdot \text{Extraversion}{)}_i+u_{0i}$$

$${{\beta }_{0i}=\gamma }_{00}+{\gamma }_{01}(\text{Conscientiousness}{)}_i+{\gamma }_{02}(\text{Sex})+{\gamma }_{03}(\text{Sex}\cdot \text{Conscientiousness}{)}_i+u_{0i}$$

Per our hypotheses, we prioritized Extraversion and Conscientiousness to explain the variability in the random intercepts. Given the sample size, Extraversion and Conscientiousness were entered in separate models to preserve power.

The Model B-Level 1 equation remained the same as Model A-Level 1. In Model B-Level 2, ${\gamma }_{00}$ was the expected amount of voice use across seven days for a female participant with average scores of Extraversion or Conscientiousness. Coefficient ${\gamma }_{01}$ represented the effect of Extraversion or Conscientiousness on expected nonadherence. ${\gamma }_{02}$ demonstrated the effect of being male. ${\gamma }_{03}$ examined the effect of the interaction between sex and Extraversion or Conscientiousness. $u_{0i}$ was the random effect, capturing unexplained variability across participants. We calculated the impact of personality traits on adherence by comparing Model B with Model A.

Supplementary Models:

Adherence for each participant over time with Emotional Stability/Neuroticism, Agreeableness, and Openness analyzed individually in separate models.

The supplementary models were similar in structure to Model B.

Supplementary Model-Level 1: Fixed effects analysis.

$$Y_{id}={\beta }_{0i}+\sum _{k=1}^{d-1}{\beta }_k\cdot {\text{Day}}_{ik}+r_{id},$$

$$\text{where}k=\left\{\begin{array}{c}1,\text{for}\text{day}k\\ 0,\text{for}\text{day}\ne k\text{and}k\ne 3\\ -1,\text{for}\text{day}=3\end{array}\right.$$

Supplementary Model-Level 2: Random effects analysis.

$${\beta }_{0i}={\gamma }_{00}+{\gamma }_{01}(\text{Neuroticism}{)}_i+{\gamma }_{02}(\text{Sex})+{\gamma }_{03}(\text{Sex}\cdot \text{Neuroticism}{)}_i+u_{0i}$$

$${\beta }_{0i}={\gamma }_{00}+{\gamma }_{01}(\text{Agreeableness}{)}_i+{\gamma }_{02}(\text{Sex})+{\gamma }_{03}(\text{Sex}\cdot \text{Agreeableness}{)}_i+u_{0i}$$

$${\beta }_{0i}={\gamma }_{00}+{\gamma }_{01}(\text{Openness}{)}_i+{\gamma }_{02}(\text{Sex})+{\gamma }_{03}(\text{Sex}\cdot \text{Openness}{)}_i+u_{0i}$$

RESULTS

Descriptive outcomes

TIPI norms for men and women aged 41–50 years were used as a referent due to the mean ages of the men and women in this study, respectively 45.4 and 47.1 years. For all five personality traits, the participants’ mean scores were higher than the male and female norms (Table 3).⁷¹ For men, the widest range was Extraversion (2.50–7.00), and Conscientiousness was the narrowest range (5.50–7.00). In women, Neuroticism was the widest range (1.00–7.00). Conscientiousness was also the narrowest range for the women (4.00–7.00). The Extraversion range was intermediate in women (3.00–7.00). The TIPI has no range norms.

TABLE 3.

Ten Item Personality Inventory Descriptive Data

Personality trait	Extraversion	Neuroticism/Emotional stability	Conscientiousness	Agreeableness	Openness
Men ($n=8$)
Mean ($SD$)	5.06 (1.76)	5.31 (0.75)	6.38 (0.58)	4.81 (1.07)	5.75 (1.04)
Range	2.50–7.00	4.50–6.50	5.50–7.00	3.00–6.50	4.00–7.00
Women ($n=17$)
Mean ($SD$)	5.18 (1.41)	4.59 (1.49)	5.65 (1.10)	5.50 (0.95)	5.79 (1.23)
Range	3.00–7.00	1.00–7.00	4.00–7.00	3.50–7.00	3.00–7.00

Inferential outcomes

Estimates of the fixed effects and random effects from Model A and Model B are displayed in Tables 4, 5, and 6. The initial reference model (Model A) indicated an estimated average daily voice use duration of 57.19 minutes during the voice rest program ($SE=21.42,p<0.01$). While there was fluctuation of voice use duration from Day 1 to Day 7, the differences were not significant, as indicated by the estimated fixed effect of these days in Table 4. Model B introduced personality traits (Extraversion and Conscientiousness) and sex as level 2 predictors. Extraversion significantly predicted adherence to voice rest protocols. Higher Extraversion scores were associated with lower durations of voice use. The slope estimations showed that the effect of Extraversion on adherence was different for men and women. A significant sex-based difference (${\beta }_{\mathit{\text{diff}}}=-47.86,SE=24.09,p=0.04$) indicated a stronger effect of Extraversion on men (${\beta }_{\mathit{\text{male}}}=-57.02$) than women (${\beta }_{\mathit{\text{female}}}=-9.15,SE=15.27,p=0.54$). Conscientiousness did not significantly predict voice use duration (${\beta }_{\mathit{\text{female}}}=-11.02,SE=23.46,p=0.63;{\beta }_{\mathit{\text{diff}}}=-26.35,SE=70.43,p=0.71$), indicating that variability in voice rest could not be reliably attributed to differences in Conscientiousness levels. In intercept estimations, sex was a significant predictor, with male participants exhibiting greater voice use than female participants, after controlling for average levels of Extraversion and Conscientiousness (Model B-Extraversion: $\beta =82.11,SE=37.19,p=0.01$; Model B-Conscientiousness: $\beta =110.31,SE=55.41,p=0.04$). For supplementary models, the fixed-effects estimates showed that Neuroticism (${\beta }_{\mathit{\text{female}}}=8.07,SE=17.23,p=0.64;{\beta }_{\mathit{\text{diff}}}=-39.48,SE=54.59,p=0.47$), Agreeableness (${\beta }_{\mathit{\text{female}}}=-0.57,SE=25.15,p=0.98;{\beta }_{\mathit{\text{diff}}}=68.01,SE=42.83,p=0.11$), and Openness (${\beta }_{\mathit{\text{female}}}=-17.76,SE=20.31,p=0.38;{\beta }_{\mathit{\text{diff}}}=-57.50,SE=41.44,p=0.16$) were not significant predictors of adherence. The estimations of fixed effects and random effects of supplementary models are displayed in Table 7.

TABLE 4.

Fixed Effects of Models A and B

Fixed Effect	Model A			Model B-Extraversion			Model B-Conscientiousness
	Value	$SE$	$p$	Value	$SE$	$p$	Value	$SE$	$p$
Intercept (grand mean)	57.19	21.35	0.01**	29.01	21.12	0.18	25.79	25.77	0.32
Day 1	−12.92a	16.08	0.42	−12.26	16.04	0.44	−13.14	16.09	0.41
Day 2b	10.97	15.74	0.48	11.62	15.71	0.46	10.94	15.75	0.48
Day 4	10.15	16.61	0.54	9.73	16.53	0.55	10.74	16.61	0.51
Day 5	−1.71	16.56	0.91	−2.00	16.53	0.90	−1.64	16.57	0.92
Day 6	−0.62	16.89	0.97	−1.06	16.86	0.94	−0.94	16.91	0.95
Day 7	−33.88	17.31	0.05	−34.41	17.28	0.04*	−34.00	17.32	0.05
Sex (Male)	-	-	-	82.10	37.19	0.02*	110.31	55.42	0.04
Extra. (Female)	-	-	-	−9.15	15.27	0.54	-	-	-
Sex · Extra.c	-	-	-	−47.86	24.09	0.04*	-	-	-
Con. (Female)	-	-		-	-	-	−11.02	23.46	0.63
Sex · Con.	-	-	-	-	-	-	−26.35	70.43	0.71
AIC	1774.50			1734.90			1739.80

Notes: “-” indicates that the predictor was not included in the model. Extra. = Extraversion. Con. = Conscientiousness. AIC = Akaike information criterion.

^aEtimate for ${\text{Day}}_k(k=\mathrm{1,2},4-7)$ is the difference between day $k$ and the grand mean (intercept).

^bEstimate of Day 3_modelA = 85.20. Estimate of Day 3_{modelB.Extraversion} = 57.39. Estimate of Day 3_{modelB.Conscientiousness} = 53.83.

^cThe estimate of slope indicated a sex-based difference in the effect of Extraversion.

p < 0.05*,

p < 0.01**,

p < 0.001***.

TABLE 5.

Random Effects of Model A

Parameter	Model A
	Estimate	$SE$	$p$ value
Patient (Random)	9398.56	3234.56	0.001***
Residual	6707.73	853.95	< 0.0001****

p < 0.001***.

p < 0.0001****.

TABLE 6.

Random Effects of Model B

Parameter	Model B-Extraversion			Model B-Conscientiousness
	Estimate	$SE$	$p$ value	Estimate	$SE$	$p$ value
Patient (Random)	6330.54	2262.67	0.002**	9398.56	3234.50	0.001***
Residual	6633.62	848.46	< 0.0001****	6653.95	853.29	< 0.0001****

p < 0.01**.

p < 0.001***.

p < 0.0001****.

TABLE 7.

Fixed Effects and Random Effects of Supplementary Models

Fixed effect	Supplementary Model-Neuroticism/Emotional Stability			Supplementary Model-Openness			Supplementary Model-Agreeableness
	Value	$SE$	$p$	Value	$SE$	$p$	Value	$SE$	$p$
Intercept (gm)	30.56	25.38	0.24	28.56	24.24	0.25	28.73	24.04	0.24
Day 1	−12.89	16.09	0.42	−12.95	16.09	0.42	−12.36	16.07	0.44
Day 2	11.03	15.74	0.48	11.11	15.75	0.48	11.61	15.73	0.46
Day 4	10.42	16.61	0.53	10.71	16.61	0.52	10.00	16.61	0.54
Day 5	−1.78	16.57	0.91	−1.81	16.57	0.91	−2.17	16.55	0.93
Day 6	−0.91	16.89	0.97	−0.80	16.90	0.96	−1.30	16.88	0.93
Day 7	−33.88	17.31	0.05	−33.41	17.32	0.05	−34.31	17.32	0.05
Sex (M)	103.69	52.19	0.04*	90.61	42.67	0.03*	118.38	44.26	0.008**
Pers.	8.07	17.23	0.64	−17.36	20.31	0.38	−0.57	25.15	0.98
Sex (M) · Pers.	39.48	54.59	0.47	57.50	41.44	0.16	68.01	42.83	0.11
Patient (RE)	9395.10	3228.67	0.0018**	8719.92	3036.60	0.002**	8097.72	2808.65	0.002***
Residual (RE)	6651.67	852.72	< 0.0001****	6655.84	853.80	< 0.0001****	6641.62	850.38	< 0.001***

Abbreviations: gm = grand mean. M = male. Pers. = personality. RE = random effects

p < 0.05*

p < 0.01**

p < 0.001***

p < 0.0001****

DISCUSSION

The purpose of this preliminary investigation was to explore the relationship between the Big Five personality traits and adherence to voice rest following phonomicrosurgery for benign mucosal or submucosal/lamina propria lesions. Participants completed the TIPI prior to surgery and wore a vocal dosimeter during the seven-day voice rest period. Participants were randomized to either a complete voice rest or a relative voice rest protocol. We gave them a time-based guideline for using their voices. The complete voice rest group were instructed not to use their voices for the duration of the study. The relative voice rest group were to remain silent for the first three days and then increase their voice use in five-minute increments, morning and night, over Days 4–7. To support the task’s feasibility and external validity, we did not require the participants to account for the duration of their voicing in a conversation versus the duration of the entire conversation. Overall, there was no statistically significant difference between the groups in the duration of voice use or clinical outcomes, as detailed in our previous paper.⁵⁸ Therefore, we combined the two groups into a single group.

We hypothesized that adherence would vary inversely with Extraversion and directly with Conscientiousness. Somewhat surprisingly, our hypothesis that elevated Extraversion (and presumably greater talkativeness and sociability) would lead to reduced voice rest adherence was not supported. Instead, our results aligned with health psychology literature related to preventive health interventions wherein greater Extraversion has often been found to vary directly with adherence. Also, contrary to general expectations, our participants’ mean scores for all five of the personality traits exceeded the normative values. High Extraversion scores could be expected because most of our participants’ lesions could be classified as PVDs, and the Trait Theory postulates that people with PVDs tend to have high Extraversion scores. However, high Conscientiousness scores in particular (combined with high values of the other personality superfactor traits) have not been observed in people with PVDs. We will contextualize these preliminary results with the preventive healthcare and medication adherence literature as well as a discussion of the nature of Extraversion.

According to the preventive health literature, people with diabetes were more likely to perform self-care activities if they had high Extraversion scores.⁵⁰ High Extraversion also predicted likelihood of participation in mammograms and breast self-examination⁵³ and adherence to COVID-19 protocols.⁵⁴ In contrast, medication adherence was almost unanimously predicated upon high Conscientiousness and Agreeableness.^44–48 Therefore, the significance of the various personality traits to health behaviors may depend on the health behavior’s specific demands. For example, preventive health practices, ingestion of medication, and voice rest could all involve awareness of a negative outcome if the action is not performed. If a woman does not perform a breast self-examination, she may not identify a breast tumor in a timely manner. If someone does not ingest blood pressure medicine, a stroke may occur. If someone does not maintain voice rest, the vocal folds could be injured. In this way, these preventive, medicine-related, and voice rest behaviors are similar. However, many differences between them remain, such as behavior-dependent cognitive demands. A breast self-examination could require more awareness, judgment, and duration of attention than ingesting a pill. Similarly, voice rest’s cognitive demands may exceed those of medication ingestion because voice rest could require a greater degree of attention, inhibition, and problem-solving skills to maintain silence and use alternate means of communication. Because high Extraversion has been found to mediate both performance of voice rest and preventive health measures, it is possible that some lower-order trait of Extraversion specifically facilitates achievement of cognitively taxing health activities. To shed light on this possibility, we explore the Big Three’s lower-order traits and how the TIPI conceptualizes Extraversion.

The TIPI measures Extraversion broadly using two prompts, ie, how “extraverted, enthusiastic” or “reserved, quiet” the participant is. The latter prompt is reverse-scored. The Big Three instrument, the Multidimensional Personality Questionnaire (MPQ),²⁴ also delineates lower-order traits for Extraversion. The MPQ and its brief form (MPQ-BF)²³ are relevant because they are two of the most frequently used personality instruments in the Trait Theory literature.^{25,29–31,33,34,39} They examine personality at both the broad trait level [PEM/Extraversion (E)] and the lower-order, primary trait levels, which for PEM/E are Achievement, Wellbeing, Social Closeness, and Social Potency.²³ The primary traits function as facets of Extraversion, and several of them suggest aspects of self that could support adherence to voice rest. Primary traits associated with the PEM/E dimension often predict positive adjustment and adaptive behavior change,²³ especially in situations requiring resilience and high endurance.^82,83

Postsurgical adherence to a voice rest regimen likely constitutes such a challenge, and Achievement and Social Potency are agentic aspects of PEM/E that may be particularly relevant.²³ Individuals who possess a strong sense of agency feel as if they are in the driver’s seat when it comes to their actions. So, the primary trait Achievement may be relevant to voice rest adherence. Someone high on this trait is “ambitious, enjoys effort, likes challenging tasks, perfectionistic, persistent, works hard [emphasis added].”^23(p154) Voice rest can be a difficult task, especially for occupational voice users.¹⁸ As noted, Dong and Lin¹⁸ found that facilitating voice rest adherence required challenging work such as (1) prenotifying contacts of voice rest, (2) using nonverbal communication, (3) understanding the rationale for voice rest, and (4) knowing the consequences of nonadherence. Someone high in Achievement may enjoy the effort and challenge that voice rest presents. This person may want to have perfect voice rest adherence and persist in this goal until it is achieved. Social Potency refers to one who “enjoys visibility, dominance, likes to be in charge” or who is “persuasive, strong, a leader.”^23(p154) At first glance, it might be difficult to imagine how Social Potency impacts adherence, but individuals who score high on this trait also possess a strong sense of agency. They tend to view life not simply as happening to them but instead believe that they are in charge of their own actions. This strong sense of agency (ie, being in control of their actions, being resilient) could override social/communal tendencies and contribute to better adherence (at least in the short term). Alternatively, being on voice rest could make one stand out among peers and lead to a perception of elevated visibility.

Wellbeing is another lower-order trait of Extraversion potentially relevant to our findings. One who scores high on Wellbeing “does fun things, has a happy disposition, has interesting experiences, and is optimistic and hopeful [emphasis added].”^23(p154) Voice rest is not likely to be fun, but it is a unique time in one’s life and could be an interesting experience. A happy disposition, optimism, and hopefulness could buoy a patient’s mood during the voice rest period, thereby supporting adherence. Indeed, high optimism has been repeatedly connected to superior physical health via flexible coping strategies,^84–86 and patients likely benefit from flexible coping skills during voice rest. Finally, one scoring high on Social Closeness is “sociable, values close relationships, warm and affectionate, and welcomes support.”^23(p154) Again returning to the study by Dong and Lin,¹⁸ individuals with greater adherence tended to prenotify contacts of planned voice rest and use nonverbal communication to maintain sociability, participate in close relationships, and receive the support of others.

Given the nature of Achievement, Social Potency, Wellbeing, and Social Closeness, it seems that qualities of optimism, hopefulness, and agency are most likely to positively influence adherence to voice rest. However, while the above discussion identifies possible facets of Extraversion that might explain our findings, it is admittedly speculation and awaits experimental verification. Therefore, discussion of the clinical applicability of these findings may be premature. Future studies that explore personality and adherence to voice rest will need to employ personality instruments like the MPQ that permit closer examination of lower-order traits. Until then, our results related to Extraversion and voice rest adherence suggest that personality deserves attention for both theoretical and practical reasons. Findings from this study, in the context of established evidence on personality and health behavior, provide us with testable hypotheses for our next steps.

LIMITATIONS

We selected the TIPI as the Big Five instrument for this preliminary exploration to provide early insights into personality as a factor in adherence to voice rest. However, because each superfactor trait was measured using only two items, we could not perform analyses at the lower-order trait level. Future research could use a comprehensive Big Three personality instrument such as the MPQ/MPQ-BF or a Big Five instrument like the NEO Personality Inventory, Revised⁶⁹ to further understand the effect of personality on adherence to voice rest at both the superfactor and lower-order trait levels. Our study was limited by the sample size. We used HLM to adjust for some of those limitations. However, future studies should be performed in a larger group of participants.

As suggested by our data, the differing sexes of our participants could have impacted our personality results. These findings were incidental but suggest future research studies (see “Future Directions” below). The Trait Theory primarily relates to PVDs (and specifically VFNs) and not other benign vocal fold lesions, so the participants’ differing vocal diagnoses might have played a role in the findings.

FUTURE DIRECTIONS

Because different personality traits have been significant for adherence in diverse disciplines, it is possible that Extraversion is only significant in voice rest and not in other voice therapy behaviors. Future research could investigate which personality traits (such as Conscientiousness and Agreeableness) are important in voice therapy adherence behaviors such as attending sessions or completing home practice activities.

To date, data from investigations grounded in the Trait Theory pertain to women. In our study, the male participants were more impacted by Extraversion and talked more (ie, were less adherent to voice rest) than the female participants. This suggests that personality traits may differentially impact voice behaviors in men versus women. Future researchers may focus on an all-male participant sample or compare male and female groups. Finally, with a larger sample size, statistical analysis could include occupation (singing occupational voice users versus nonsinging occupational voice users versus nonoccupational voice users) as a factor that may provide more nuance to the role of personality in adherence.

CONCLUSIONS

Consistent connections have been documented between personality and voice behavior in relation to the onset and maintenance of voice disorders. Our findings encourage further investigation into personality and adherence to the behavior of voice rest. Through understanding the significant personality traits, our field may tailor specific supports to increase voice rest success and tissue healing. These improved outcomes may prevent recurrence of benign mucosal or submucosal/lamina propria lesions and help patients to avoid additional surgeries.