Prevalence and Risk Factors for Voice Problems Among Telemarketers

Objectives  To investigate whether there is an increased prevalence of voice problems among telemarketers compared with the general population and if these voice problems affect productivity and are associated with the presence of known risk factors for voice problems.

Design  Cross-sectional survey study.

Settings  One outbound telemarketing firm, 3 reservations firms, 1 messaging firm, 1 survey research firm, and 1 community college.

Participants  Random and cluster sampling identified 373 employees of the 6 firms; 304 employees completed the survey. A convenience sample of 187 community college students similar in age, sex, education level, and smoking prevalence served as a control group.

Main Outcome Measures  Demographic, vocational, personality, and biological risk factors for voice problems; symptoms of vocal attrition; and effects of symptoms on work.

Results  Telemarketers were twice as likely to report 1 or more symptoms of vocal attrition compared with controls after adjusting for age, sex, and smoking status (P<.001). Of those surveyed, 31% reported that their work was affected by an average of 5.0 symptoms These respondents tended to be women (P<.001) and were more likely to smoke (P = .02); take drying medications (P<.001); have sinus problems (P = .04), frequent colds (P<.001), and dry mouth (P<.001); and be sedentary (P<.001).

Conclusions  Telemarketers have a higher prevalence of voice problems than the control group. These problems affect productivity and are associated with modifiable risk factors. Evaluation of occupational voice disorders must encompass all of the determinants of health status, and treatment must focus on modifiable risk factors, not just the reduction of occupational vocal load.

APPROXIMATELY one third of the labor force in industrialized societies relies on voice as their primary work tool.¹ Occupational voice disorders might be the result of the repetitive movement or "collision" of the vocal folds.¹ Vocal attrition can be described as "the ‘wear and tear' of the vocal mechanism and the overall reduction in vocal capabilities associated with acute or chronic abuse of the phonatory system."²(p130) Recent studies have shown an association between voice problems and vocally demanding jobs such as teaching,^3-6 singing,^7-9 and aerobics instruction.^10,11

Vocal attrition and occupational voice disorders are also a result of the combined effects of vocational, personality (lifelong speech habits), and biological factors.^1,2 Vocational factors include vocal loading or the vocal demand of the job, background noise, room acoustics, speaking distance, air quality, posture, stress, and equipment design.¹ Speech-related personality factors include the tendency of a person to habitually speak loudly, excessively, and rapidly.² Biological factors include all of the factors that may affect the vocal mucosa or the respiratory ability to support speech, such as smoking,^7,12,13 hydration,¹² intake of caffeinated beverages,^12-14 medications,^12,13,15 respiratory illnesses, chronic allergic rhinitis, sinusitis,^1,10,12 gastroesophageal reflux,^12,13,16,17 and general physical condition.¹

Despite recognition of the existence of occupational voice disorders and vocal attrition, an objective measure of vocal function similar to the audiogram for hearing is not available. Furthermore, the ability to quantify the absolute amount of vocal dysfunction would not describe the impact of a voice problem on an individual's quality of life.¹⁸ Consequently, it is necessary to consider the nature of the voice symptoms reported by the target population and the impact of those symptoms on work and social interaction when defining the nature of a voice problem. There are currently 3 valid and reliable disease-specific outcome measures for use with populations who have an existing diagnosis of a voice disorder. These measures are the Voice Handicap Index,¹⁹ the Voice-Related Quality of Life Measure,¹⁸ and the Voice Outcome Survey.²⁰

There are no reliable data on the prevalence of voice disorders in the general adult population,¹⁸ and to our knowledge, a detailed examination of voice in the telemarketing environment has not been done. Titze et al²¹ report that telemarketers made up 2.3% of the patient volume in a voice clinic but comprised only 0.78% of the 1994 workforce; indicating that telemarketers were about 3 times as likely to seek help from a voice clinic than the general population. This research by Titze et al²¹ represents the only literature we found that specifically considers telemarketing in association with voice problems. We hypothesize that there is an increased prevalence of voice problems among telemarketers compared with the general population and that these voice problems affect productivity and are associated with the presence of known risk factors for voice problems. The 3 validated voice disorder outcome instruments do not meet our need for the collection of data on vocational, personality, and biological factors. To test these hypotheses we designed a 57-item "voice survey" that included questions regarding demographic, vocational, personality, and biological risk factors for voice problems; symptoms of vocal attrition; and effects of any voice problem on work and social interaction.

This cross-sectional survey research was accomplished by contacting human resources directors and managers at 19 separate firms in 2 midwestern cities. After discussing the research topic and hypotheses, 6 firms agreed to participate in the surveys. The firms who declined to participate cited a desire to avoid raising employee awareness of voice problems. The surveys were conducted between August 8 and October 4, 2000. The participating firms included 1 outbound telemarketing firm, 3 reservations centers, 1 messaging company, and 1 telephone survey research firm. The employee participants of these firms are collectively referred to as "telemarketers."

Random sampling was performed at small firms, and cluster sampling was performed at larger firms to avoid selection bias. Sampled employees received letters of invitation that described the survey and the time required for completion and informed them that a snack would be provided as compensation for their time, that all responses would be anonymous, and that the research was approved of by their employer and the institutional review board of the University of Nebraska Medical Center. Participation rates among the 6 firms varied from 50% to 100%. A total of 373 employees received letters of invitation and 304 completed the survey for an overall participation rate of 82%.

To test the hypothesis that telemarketers have a higher prevalence of voice problems than the general population, we also surveyed a convenience sample of 187 community college students similar to the telemarketers in age range, education level, and smoking status. The students were recruited in the commons areas of the 3 separate branches of a local community college in December 2000 and January 2001. Students read a similar letter of invitation and also received a compensatory snack for completing the survey. The student "control survey" was identical to the 57-item voice survey except that the 16 questions regarding the work environment were excluded.

The voice survey was developed from a review of the literature regarding occupational voice disorders and included recommendations from otolaryngologists, a speech and language pathologist, and the human resource and managerial professionals from the surveyed firms. The survey consisted of the following:

• 16 questions regarding the employee's telemarketing work history and environment

• 7 questions concerning respiratory health, medication use, hearing ability, smoking, acid reflux, and caffeinated vs noncaffeinated beverage intake

• 3 questions concerning amount, speed, and volume of social speech

• 2 questions concerning general activity level and participation in vocally demanding activities

• 1 question regarding previous vocal hygiene education

• 14 yes/no response questions concerning the current presence of various symptoms of vocal attrition (participants were instructed that these symptoms do not include sore throat or laryngitis associated with a common cold)

• 2 items to rank the quality of their voice at the beginning and end of a shift

• 2 questions regarding the employee's characterization of and feelings about the presence of any symptoms of vocal attrition

• 2 questions regarding the impact of any symptoms on the employee's work and social life

• 1 question regarding the relationship between season of the year and presence of symptoms

• 2 questions regarding previous treatment for symptoms of a voice problem

• 5 demographic questions

All raw data were entered into an Access database (Microsoft Inc, Redmond, Wash) to determine descriptive statistics. Analyses were performed using SAS software (SAS Institute Inc, Cary, NC). The end points of interest in comparing the telemarketers with the community college students are the prevalences of reporting 1 or more symptoms of vocal attrition and decreased social interaction due to symptoms. The end point of interest concerning the telemarketing group is the prevalence of impaired job performance due to symptoms of vocal attrition. The Fisher exact test and multiple logistic regression analysis were used to examine the association of telemarketing with the 2 end points of reporting 1 or more symptoms of vocal attrition and interacting less with family and friends. The Fisher exact test was also used to determine significant associations between biological, vocational, and personality factors and impaired telemarketing job performance due to symptoms of vocal attrition. All tests were 2-sided, and P≤.05 was considered statistically significant.

The age, sex, race, education level, and smoking status of the telemarketers and community college students are given in Table 1. The 2 groups are similar with respect to age and smoking status. However, the telemarketing group has more women than the student group (P = .007).

A total of 206 (68%) of the 304 telemarketers and 90 (48%) of the 187 community college students reported 1 or more symptoms of vocal attrition. This increased prevalence of the 14 symptoms of vocal attrition in the telemarketers is shown in Figure 1. Telemarketers are significantly more likely than the students to experience 8 of the 14 symptoms (P<.01). Multiple logistic regression analysis was used to find the odds of 1 or more symptoms of vocal attrition due to telemarketing while adjusting for age, sex, and smoking status. In this analysis, 4 observations were omitted owing to missing information. As given in Table 2, telemarketers are 2.1 times as likely to have 1 or more symptoms of vocal attrition than the students after adjusting for age, sex, and smoking status (P<.001). Smoking and sex, independent of telemarketing, also show a significant association with vocal attrition.

Of 282 telemarketers 8 (3%) reported that their voice problems caused them to interact less with family or friends, while 7 (4%) of the 187 community college students reported the same. The Fisher exact test did not indicate a significant association between telemarketing and interacting less with family and friends. However, female sex (P = .04) and each 1-year increase in age (P = .03) demonstrated an increased odds of interacting less with others because of a voice problem for both telemarketing and student participants. Women were 8.65 times (95% confidence interval, 1.1-67.6) as likely as men to report interacting less with family and friends because of a voice problem. Every 1-year increase in age results in a 4% increase in the risk of reporting decreased social interaction because of a voice problem.

Of the 304 voice participants, 94 (31%) responded that they had worked while having a voice problem and that this problem affected their productivity. These 94 participants were then asked to indicate whether 5 statements described the effect of the voice problem on their work. The frequency in which the 5 statements were chosen is given below.

Table 3, Table 4, and Table 5 summarize the differences between the participants who reported that their work was affected by the presence of symptoms and those whose work was not affected. Table 3 reports the association of demographic, vocational, and personality factors and work being affected by symptoms of vocal attrition. Female sex is the only demographic variable that is significantly associated with work being affected by symptoms. Such vocational factors as years worked in telemarketing, length of the shift worked, and hours worked per week are not significantly associated with reporting that work is affected by a voice problem. Those who indicated that they frequently experienced a stressful call and who described their work environment as dry or cold were significantly more likely to have indicated that their work was affected. Indicating a need to raise one's voice more than a few times each shift because of noise is associated with work being affected (P = .051). The personality factors "I talk more than others" and "I talk louder than others" are significantly associated with reporting that work is affected by a voice problem.

The biological factors investigated are given in Table 4. Reporting frequent colds, regularly clearing the throat, experiencing a dry mouth and throat, and taking drying medications are significantly associated with work being affected by symptoms. Those whose work is affected tend to drink more beverages per shift than those whose work is not affected. The association between having a dry mouth and throat and taking drying medications is statistically significant (P = .002). Additional significant biological factors for work performance being affected by symptoms are inactivity, smoking, and regularly experiencing acid reflux or "heartburn." Reporting "I don't participate in any other additional activities that might stress my voice" is significantly associated with work performance not being affected by symptoms.

The increased prevalence of the 14 symptoms of vocal attrition in telemarketers whose work was affected compared with those whose work was not affected is shown in Figure 2. Telemarketers whose work was affected are significantly more likely to experience 12 of the 14 symptoms than those whose work was not affected (P<.01). Table 5 reports the greater effect of symptoms of vocal attrition on telemarketers whose work was affected (n = 94) compared with those telemarketers with symptoms whose work was not affected (n = 112). Those whose work was affected averaged 5.0 symptoms of vocal attrition; 78% desired instruction about caring for their voices, 50% had missed work because of a voice problem, and 8% had interacted less with family and friends owing to their voice problem. Sixty-nine percent of those whose work was affected and 80% of those whose work was not affected reported that their voices were normal at the beginning of their work shift. By the end of the work shift, only 17% of those whose work was affected reported that their voices were normal, while 36% of those whose work was not affected considered their voices to be normal. Regarding whether a physician had been seen because of a voice problem, there was no statistical difference between those whose work was affected (14%) and those whose work was not (9%). Only 4 of the 304 telemarketers reported that they had seen a speech therapist because of a voice problem, and only 1 of those 4 reported that their work was affected by a voice problem.

The health status of an individual or population is a result of the interaction between behavior and lifestyle, heredity, environment, and medical care.²² Our cross-sectional survey of voice problems in telemarketers confirms the interaction between multiple factors in determining vocal health. Voice is the primary work tool for one third of those in industrialized societies, and previous studies have shown an association between voice problems and vocation.^3-11 Our study sought to evaluate the prevalence and risk factors for voice problems among telemarketers.

As expected, we found a significant increase in the prevalence of symptoms of vocal attrition in telemarketers compared with a control group of community college students. The telemarketing group was twice as likely to have 1 or more symptoms after adjusting for age, sex, and smoking status compared with the student group. Regression analysis found female sex and smoking status to be independent factors strongly associated with symptoms of vocal attrition. The finding that female sex is independently associated with vocal attrition is consistent with current research. In a cross-sectional study of 25 cadaveric vocal folds, Butler et al²³ found that women have less hyaluronic acid in the first 15% of depth (most superficial) of the lamina propria (subepithelial layer) compared with men. These authors conclude that this decrease in hyaluronic acid among women may provide less protection from vibratory trauma and overuse and may explain in part why women have more voice-related trauma than men.

Impaired work productivity due to voice problems occurred in 31% of the telemarketers surveyed. Impaired productivity was attributed to the communication difficulties of needing to repeat what was said and having to force the voice to be understood. These communication difficulties caused a decrease in call frequency and diminished enthusiasm for selling the product. Unexpectedly, within the telemarketing group, factors indicating workload were not significantly associated with prevalence of symptoms and impaired productivity, while half of the biological factors and 2 of 3 personality factors were significantly associated with symptom prevalence.

A major limitation of our study is the reliance on self-report, which is inherent in survey research. However, the number of people surveyed (N = 560) and the high participation rate of 82% contribute to the validity of the findings. An additional limitation is that the telemarketing group was surveyed during the summer and early fall, while the student group was surveyed during the winter. Consequently, the telemarketing group may have been influenced by seasonal allergies, while the student group may have been influenced by the dryness of central heating during a midwestern winter. A major strength of our study is the comparison between 2 groups similar in age and smoking status but differing in occupational vocal load. The advantage of this design is that the role of occupational vocal load in causing voice problems can be separated from other determinants of vocal health.

The clinical significance of these findings is that evaluation of occupational voice disorders must encompass all of the determinants of health status, and treatment must focus on modifiable biological, environmental, and personality factors and not simply reduction of the occupational vocal load. Although occupational voice disorders are a direct result of the repetitive collision of the vocal folds, additional risk factors are present when symptoms of vocal attrition affect work productivity and social interaction. Evaluation of occupational voice disorders should also include consideration of intrinsic laryngeal pathological conditions such as malignancy, vocal cord nodules and polyps, and laryngeal papillomatosis. Modifiable biological factors that should be considered include smoking status, chronic sinusitis and frequent upper respiratory tract infection, gastroesophageal reflux disease, the use of drying medications, hydration, and complaints of dry throat. Once the laryngeal pathological condition and modifiable biological factors have been addressed, complete management of occupational voice disorders should then consider vocational load, speech habits that are related to personality, and general activity level.

Addressing voice problems and the resultant impaired productivity in telemarketers will require education of employers and employees regarding the multifactorial nature of occupational voice disorders. Initiating a multidisciplinary treatment plan for patients with occupational voice disorders is facilitated by use of an in-take form designed to efficiently gather information regarding all behavioral, lifestyle, hereditary, environmental, and vocational risk factors. Speech and language pathologists in our voice clinic address the issues of workplace vocal load, vocal use patterns, vocal hygiene, and education.

Accepted for publication October 26, 2001.

This study was presented as poster R0050 at the annual meeting of the American Academy of Otolaryngology–Head and Neck Surgery Foundation/Association for Research in Otolaryngology, Denver, Colo, September 10-11, 2001.

Corresponding author: Katherine Jones, MS, 984350 Nebraska Medical Center, Omaha, NE 68198-4350 (e-mail: kjonesj@unmc.edu).

Reprints: Frederic Ogren, MD, ENT Physicians, PC, 17030 Lakeside Hills Plaza, Suite 200, Omaha, NE 68130.