5. MARCO DE REFERENCIA
5.1. DESCRIPCIÓN DEL ÁREA DE ESTUDIO
5.1.4. ASPECTOS HISTORICOS
An assessment protocol had been developed and tested in a pilot study (between June to October 1999). Minor corrections were made to this data form after evaluation o f the pilot study results (described in appendix A)
The sample o f volunteers was recruited from subjects who were attending a normal examination at the ENT unit o f the hospital and others were unpaid healthy volunteers (selected from the peer group o f the researcher and from the staff working at the hospital). All the patients who came to this unit (except the healthy volunteers who came on purpose for the study) had had a previous ENT examination and they were referred in order to obtain a more precise diagnosis by using laryngoscopy and/or stroboscopy. From the total number o f persons at the out-patient ENT unit the researcher randomly selected those who were sent with a voice complaint. Several difficulties arose in the subjects’ recruitment either because they simply refused to participate, were illiterate, or needed glasses to read and did not have them at the time o f assessment. Among potential healthy volunteers the major reason for refrisal was the need to submit to a stroboscopy examination.
Prior to the assessment each subject had a verbal explanation about the purposes and the procedures o f the experiment that was in progress. If he/she agreed to participate a consent form had to be signed (appendix B).
The assessment protocol included stroboscopic examination by two ENT surgeons (appendix C), individual history questionnaire (appendix D), acoustic and laryngographic recording and auditory voice assessment by the same speech and language therapist (the researcher), self-rating o f psychosocial consequences o f voice disorder (appendix E) and self-rating degree o f psychological stress (appendix F) by each subject.
All subjects were assessed, only once, with all the sources above explained, always during the morning (from 9 to one) to avoid, as much as possible, within-day variations in FO (Brown, Murry & Hughes, 1976; Garrett & Healey, 1987; Higgins, Netsell & Schulte, 1994).
5.1 - Subjects
One hundred and thirteen subjects participated in this study but the result was a set o f 109 subjects (or 96.5% o f all subjects recruited) whose data could be used. Only four subjects were eliminated based on the lack o f or unrehable quahty o f the Lx signal (two were 61 year old men with leukoplakia, a 19 year old woman with psychogenic aphonia and a 43 year old woman without voice and laryngeal problems).
Subjects had to meet tlie following criteria for inclusion in the study: (1) age over 18 years old and under 65. This lower limit o f 18 years was set based on the criteria that subjects had completed voice change and the upper limit of 65 years was selected to avoid the effects associated with the ageing process. That is, the same criteria used by several other researchers (Fitch, Holbrook & Tallahassee, 1970). However^ a 67 year old female, subject 84, without voice problems was accepted; (2) Portuguese speakers with fluent and intelligible speech, as rated by the researcher, (3) no history o f an upper respiratory tract infection (URTI) on the day o f the evaluation and no auditory handicap and, (4) reasonable reading skills. Although two subjects with hteracy difficulties were accepted (subjects 86 and 100) all their data from oral reading and subjects’ self-rating Voice Handicap Index (VHI) and Social Readjustment Rating Scale (SRRS) were not accepted for analyses (as will be mentioned in the data analyses section).
From the total 109 subjects, 82 were female subjects with a mean age of 42 years old (ranging fi-om 19 to 67 years old) and 27 were male subjects with a mean age o f 39 years old (ranging firom 20 to 63 years old).
The subjects were classified into two groups on the basis o f their voice quahty. For the purposes o f this study, a subject was included in the ‘dysphonie’ group when he/she presented with a voice problem on the day o f the assessment or history o f voice complaints and for which the SLT and the ENT surgeon found corroborative evidence. As history o f voice complaints meant only permanent or frequent episodic voice problems not related to upper respiratory tract mfcction or allergic situations. Subjects were included in the ‘control’ group when they did not report a voice problem on the day o f the assessment or history of frequent voice problems not related to upper
An electrolaryngographic study of dysphonie Portuguese speakers
Chapter V - M ethodology
respiratory tract infection or allergic situations and for which the SLT and the ENT surgeon did not find corroborative evidence.
The demographic information about the groups is presented in table V .l. The dysphonie group consisted o f 52 subjects (39 females and 13 males). The mean age was 44 years and the mean level o f education was eleven years. The control group consisted o f 57 subjects (43 females and 14 males). The mean age was 39 years and the mean level o f education was thirteen years. In both groups the majority o f subjects were Caucasian.
Table V.l - Subjects’ demographic information
Dysphonies (n=52)
Controls (n=57)
Age 44.06 ± 12.31* 38.53 ± 14.94*
Age range 19-64 years 20-67 years
Females’ age 43.62 ± 12.99* 40.12 ± 15.13* Males’ age 45.38 ± 10.38* 33.64 ± 13.69* % Females 75% 75.4% % Males 25% 24.6% Years of education 10.69 ±5.43* 12.5 ± 4.57* % Caucasian 96.2% 94.7%
* M ± SD (Mean ± Standard deviation)
Independent T-tests were performed to compare the mean age according to dysphonie and control groups and the results showed that there was a significant difference (t^2.116, df=105, p=û037). Nevertheless, when the mean age was compared according to gender between the groups the results show that there was no significant difference in age between the females’ groups (t= -l.l 18, df=80, p=0.267) but there is a significant difference m age between the males’ groups (t=-2.495, df=25, p=0.020) with the dysphonie males being older than control males (as might be seen in figure V.l).
Males Dysphonies
Group
Figure V.l - Age distribution according to gender and groups.
Also, no significant differences were found between the groups in the mean years o f education (t=1.878, df=100, p=0.063), in gender fi-equencies (%2=0.003, df=l, p=0.958) and in the fi-equencies o f Caucasians (x2=0.125, d f= l, p=0.724).
The social and voice habits history data (fi-equencies, means and standard deviation values) gathered during the interview is shown in the table below (table V.2) for both groups.
Table V.2 - Dysphonies and controls’ social and voice habits history
Dysphonies (n=52) M±SD* Controls (n=57) M±SD* % Lisbon natives 55.8% 61.4%
Number of years hving in Lisbon 29.8 ± 15.3 26.4 ± 14.7
% In employment 61.4% 75%
% Living alone 15.8% 15.4%
% Professional voice use 25% 22.8%
% Voice hobby 12.3% 7.7%
M ± SD (Mean ± Standard deviation)
No significant differences were found in the percentage of Lisbon natives (x2=0.768, df=l, p=0.381), number of years hving in Lisbon (t=1.176, df=107, p=0.242), in employment (x2=2.306, df=l, p=0.129), in living status (x2=0.003, df=l,
An electrolaryngographic study o f dysphonie Portuguese speakers
Chapter V - M ethodology
p=0.954), in professional voice use (%2=0.72, df=l, p=0.788) and voice hobby (%2=0.631, df=l, p=0.427) between the dysphonies and controls.
Table V.3 (below) shows the health history data gathered during the interview (the means and standard deviation values and frequencies) and from the ENT’s examiuation (nose, hypophaiynx and rhinophaiynx) and from the self-rated scale o f stress (SRRS).
Table V.3 - Dysphonie and control groups’ health history
Dysphonies (n=52) M ± S D * Controls (n=57) M ± S D * Height (cm) 1.62 ± .093 1.63 ± .087 Weight (Kg) 65.71 ± 13.38 63.23 ± 11.40 Medication consumption 0.79 ± 1.37 1.16 ± 2.52
Daily coffee intake 1.92 ± 1.64 1.35 ± 1.28
Daily water intake 1.08 ± 0 .6 6 0.84 ±0.51
Cigarettes a day 6.5 ± 11.64 1.64 ±4.05
% smokers 28.8% 22.8%
Surgery over last 3 years 11.5% 10.5%
Respiratory diseases 15.4% 8.8%
Gastroesophageal disease 26.9% 12.3%
Minor nose abnormalities 57.7% 52.6%
Minor rinopharynx abnormalities 21.2% 15.8%
Minor hypopharynx abnormalities 50% 42.1%
Stress (SRRS) Total 117.8 ± 8 8 .9 87.5 ± 63.9 Family 13.9 ± 2 5 .6 16.16 ± 29.2 Personal 52.9 ± 56.6 30.11 ±36.8 Work 14.4 ± 19.7 8.3 ± 14.4 Financial 36.7 ± 36.2 32.8 ±36.7
*M ± SD (Mean ± Standard deviation)
No significant differences were found in the mean heights (t=0.511, df=105 p=0.611), mean weights (t=-1.033, df=105, p=0.304), percentage o f smokers (%2=0.247, df= l, p=0.619), mean consumption o f prescribed medication per day (t=0.930, df=107, p=0.354), reported frequency o f use o f birth control pills (%2=2.022, df=l, p=0.155),
frequency o f reported respiratory diseases (%2=1.132, df=l, p=0.287), frequency o f surgery over last 3 years (%2=0.028, dfr=l, p=0.866), frequency o f minor nose abnonnalities (%2=2.819, df=2, p=0.244), frequency o f minor rhinophaiynx abnormalities (%2=2.908, df=2, p=0.234), and in the frequency o f minor hypophaiynx abnormalities (%2=3.279, df=2, p=0.194) between the dysphonies and controls.
Significant differences were found in the mean number o f cigarettes smoked a day (t=-2,850, df=62 p=0.006), mean quantity o f daily water intake (t=-2.143, df=107, p=0.034), mean number of coffees per day (t=-2.032, df=107, p=0.045), the frequency o f reported gastroesophageal disease (%2=3.748, df=l, p=0.045) and the total level o f stress (t=-2, df=90, p<.05) between the groups.
In summary, dysphonies are older than controls (just for the male subjects), present higher daily consumption habits (cigarettes, coffee and water) than controls and present more health related problems (gastroesophageal disease and stress) than controls.
In order to determine the effects o f group, age and gender on the results o f the SRRS scale a multiple repeated analysis was carried out. Results revealed that the main effect of group was statistically significant (F(l,103)=3.85, p<.05) and the within subjects’ effect also showed a statistically significant difference between the sub-scales scores (F(2,103)=94.4, p<.0001). A two way significant interaction was also foimd for SRRS-group interaction (F(2,103)=3.16, p<.05).
Post hoc comparisons were done so that significant differences between the SRRS sub-scales and group combinations could be assessed. Results showed significant differences for personal stress (t=-2.4, df=84, p<.05) and work stress (t=-1.8, df=89, p<.05) between the groups. Results shown in table V.3 confirm that dysphonies show higher scores in the subscales o f personal and work stress than those o f controls.
An electrolaryngographic study of dysphonie Portuguese speakers Chapter V - Methodology 120 100 CO or. or CO 40 ,„Group D ysphonies % Controls SRRS
Figure V.2 - SRRS means according to groups
Figine V.2 (above) shows the significant main effect o f group (by the gradient of the traces) and the significant interaction between SRRS subscale (family) and group (by the intersection of the traces).
According to the ENT results (figure V.3) in the dysphonies, 3 subjects (6%) present healthy larynxes and 1 subject (2%) shows minor laiynx variation. Tlie most fi-equent perturbation is the ftmctional disorder (24 subjects) followed by the mass lesion disorder (14 subjects) and by the tissue change disorder (10 subjects). In the control group, 24 subjects (42%) presented minor larynx variations that do not interfere in the vocal fold vibration (e.g. 13 females presented slight posterior glottal chink).
D y s p h o n ie s C o n tro ls
Figure V.3 - ENT larynx diagnosis -141-
The SLT voice diagnosis (figure V.4) shows that in the dysphonies’ group 4 subjects (7.6%) were diagnosed with ‘normal’ voice and the hoarse voice is the most fi-equent dysphonie pattern (19 subjects) followed by harsh voice (15 subjects) and breathy voice (14 subjects).
60 50 40 30 20 10 Dysphonies I I Controls 0
normal breathy harsh h o a rse
Figure V.4 - SLT voice diagnosis
According to the SLT dysphonia severity diagnosis among the 48 diagnosed dysphonie subjects, 35.4% show a severe degree of dysphonia, 35.4% a moderate degree of dysphonia whilst 29% show a shght degree of dysphonia.
Table V.4 - Dysphonia history
Dysphonies N = 5 2
Cause Onset Symptoms Duration
Work conditions - 25% Gradual - 79% Permanent - 73% Less than one year - 21% D ise a s e - 2 1 % Sudden - 19% Intermittent - 27% One to five years - 44% Psychological - 13.5% Unknown - 2% More than 10 years - 35% Tobacco - 1 3 . 5 %
Several others - 12% Unknown - 12%
An electrolaiyngographic study of dysphoric Portuguese speakers
Chapter V - Methodology
The subjects’ reported characteristics o f dysphonia (table V.4) shows that the most jfrequent reported cause o f dysphonia was work conditions followed by disease, psychological factors and tobacco, several causes, and unknown cause. The majority o f subjects reported a gradual onset o f the dysphonia and permanent symptoms of dysphonia. 44% o f the subjects reported dysphonia from one year to five years, 35% had had dysphonia for more than 10 years and 21% for less than one year.
5.2 - M aterials
The material used in this study was developed in or translated into Portuguese for the project and was based on materials used m previous studies investigating voice quahty. The rehability or vahdity o f the materials used was not formally tested. The translations into Portuguese were made by the researcher, back-translated by a Portuguese Enghsh teacher at secondaiy school in Portugal (who completed a Master’s degree in the University of London in 1998/99) and revised by a bilingual (Portuguese/ Enghsh) speech and language therapist. Some o f the investigator’s relatives and fiiends read the translated materials in order to check for the quahty o f the translation. All the materials were previously tested in a pilot study (appendix A).
There were different types of materials: for screening (an ENT assessment form, a standard screening questionnaire, an auditory voice assessment and a self-rating scale o f life events occurrence) and measurement (sustained vowels, reading, conversation and a self-rating questiormaire o f the psychosocial consequences of voice). Screening materials were developed to gathered information and classify the subjects into the different groups (dysphonie and control). Although they constitute an important form o f assessment for the purpose o f the present study only part o f the information was analysed.
The ENT assessment form is a modified version o f the existent nasolaiyngostroboscopy assessments (Hirano, 1981; Hirano and Bless, 1993), developed by the Portuguese ENT surgeons, to register their findings. It consists of personal details, clinical history, personal background, and family clinical history, laryngoscopic information about the nasal cavity, rhino- and hypopharynx and laiyngeal anatomy. The items for the laiynx observation are organised as: vibration (symmetrical or asymmetrical), glottal closure (complete, or incomplete), regularity o f the vocal fold
edge (regular or irregular), the amplitude o f vibration and mucosal wave ranging from normal to no visible movement, and non-vibrating portion ranging from none to total. The Enghsh and Portuguese version of the ENT assessment form is shown in appendix C.
A global auditory method o f voice assessment in accordance with some of the existing perceptual methods and commonly used in perceptual voice studies (Hirano, 1981; Haji et al., 1986; Eskenazi, Childers & Hicks, 1990; Laver, 1991) was organised by the researcher. It consists of 4 overall voice quahties: normal, harsh, hoarse and breathy (described in detail in chapter III) and a four-point severity scale ranging from normal voice to severe dysphonia with the following criteria; (1) ‘normal’ voice - auditory impression of a voice characteristic o f the speaker’s age and gender and in accordance with the speaking situation. The auditory impression o f non-dysphonic voice with pecuharities that are accepted socially and culturally was also classified as ‘normal’ voice because it is characteristic either o f laryngeal physiology (e.g. shghtly breathy voice associated with posterior glottal chink) or personal attitude (e.g. too low in loudness without giving the auditory impression o f hypokinetic voicing), or the result o f habits (e.g. habitual smoking is cormected with too low pitch; and primaiy teachers’ vocal load usually results in voices that are too loud); (2) shght dysphonia- auditory impression o f very shght, intermittent voice abnormahties or isolated pitch or loudness perturbations without difficulties in intelhgibihty; (3) moderate dysphonia - auditory hnpression of frequent voice abnormahties o f pitch or loudness with intermittent difficulties in intelhgibihty; (4) severe dysphonia - auditory impression o f constant voice abnormahties o f pitch or loudness with constant difficulties in inteUigibihty.
The screening questionnaire (appendix D) was developed according to the principal investigator’s professional experience and with reference to some pubhshed ‘history questiormaires’, (Koschkee & Rammage, 1997; Dworkin & Meleca, 1997; Sapir, Mathers-Schmidt, Larson, 1996; Sapir, Keidar & Mathers-Schmidt, 1993; Sataloff et al., 1994a) and foUowing similar procedures o f some voice studies (Fitch, 1990). The questiormaire gathers information about a subject’s personal details, social history, general health history, ENT history, voice history, smoking history, drinking history and food history. The purpose was to identify ah the factors that may contribute individually, or in combination, to dysphonia. It was decided to use open-ended questions without pre-coded responses as subjects’ answers may be complex,
An electrolaiyngographic study of dysphonie Portuguese speakers
Chapter V - M ethodology
unpredictable or too numerous to pre-code. Also, closed questions could carry the risk that subjects’ answers may be forced into inappropriate categories.
The Social Readjustment Rating Scale (SRRS), Holmes & Rahe (1967), consists o f a 43-item scale o f major hfe events either indicative o f the hfe style o f the individual (personal) or o f occurrences involving the individual (family, personal, work and financial). The emphasis is on major life events that could force die organism into illness or maladaptive behaviour, and not on psychological meaning, emotion or social desirabihty. The SRRS assigns points (through the use o f the magnitude estimation technique) to stressful events on the basis o f standards from the average individual. The scores on SRRS became weU known as the ‘resultant magnitude estimates’, later termed life change units (LCD’s) that reflect the amount o f change in life that one will have to adjust to. Holmes & Rahe (1967) showed a high degree o f consensus (correlation coefficients between 0.82 and 0.90) concerning the relative order and magnitude o f the items in the scale for a sample o f convenience composed o f 394 American adult subjects, as did Ruch & Holmes (1971) in a rephcation study using an adolescent sample (correlation coefficient o f 0.97).
Subjects are instructed to cross those items which they have experienced in a given period o f time (usually 6 month, one year or the hfetime). A quantitative amount o f change in ongoing hfe adjustment required for each subject in any given period of time was determined by Holmes & Rahe (Holmes & David, 1989) to define categories known as hfe crises with different magnitudes.
Since 1967, the original SRRS (used in the pilot study, appendix A) has undergone only two shght modifications. Firstly, doUar inflation makes items 20 and 37 inadequate (‘mortgage over $10,000’ and ‘mortgage or loan less than $10,000’ respectively) and they now read (‘mortgage or loan for a major purchase’ and ‘mortgage or a loan for a lesser purchase’ respectively), (used in the main study, appendix F). Secondly, the authors realised that the original SRRS with abbreviated wording for the hfe event items (decided by them to save space) was not as adequate as the complete wording used nowadays as a more accurate and helpful form (Hohnes & David, 1989).
The SRRS form (appendix F) was used here because it is one o f the most commonly used measures of general stress in the literature (Holmes & David, 1989; MiUer, 1989; Goldman et al., 1996; Rosen & Sataloff, 1997). Moreover, it does not place high demands on the subjects’ interpretation and time to complete it, and the
researcher needs no specific training. The stressors of the preceding six months were self-marked by each subject. Using six months as the time frame for subjects’ ratings reduces the problems associated with subjects’ recall o f hfe events.
The speech materials used for the laryngographic and acoustic measurement were: (1) sustained vowels (the Portuguese vowels [a], [i] and [u]), (2) reading aloud- ‘The Story o f Arthur the Rat’ developed by Abercrombie (1967) (appendix G) and, (3) conversation.
The rationale behind the choice o f the sustained vowels was: (1) sustained vowels provide a relative stable condition o f the phonatory system, therefore relevant to the evaluation o f the laiynx stabihty (Horn, 1979; Heiberger & Horii, 1982; Askenfelt & Hammarberg, 1986; Titze, Horii & Scherer, 1987; Fitch, 1990; Muny, Brown & Morris, 1995; Colton & Casper, 1996; Baken & Orlikoff, 2000); (2) these three vowels represent the range of physiological production (even in Portuguese as shown in table V.5) and include the two dimensions that have been mostly studied (vowel height and front/back) (Whalen & Levitt, 1995) (described in detail in chapter III) and (3) they have been used in other voice studies to which the data from this study can be compared (shown in chapter III).
Table V.5 - Portuguese vowels used in the study
. Degree o f mouth
opening Anterior Central Posterior Tongue height
Relatively closed [i] M High
(lips spread) (lips rounded)
Open Low
The rationale behind the choice of reading was: (1) it is a feasible procedure in order to obtain uniform samples (with the same linguistic content) of connected speech, (2) it provides repeatable and relatively consistent subject matter for cross-comparison