Community Dent Oral Epidemiol 2003; 31: 275±84

Printed in Denmark. All rights reserved

Validation of an oral health

S. Tubert-Jeannin1, P. J. Riordan1,
A. Morel-Papernot2, S. Porcheray1
and S. Saby-Collet1

quality of life index (GOHAI) in

1
Department of Public Health, School of
Dentistry, University of Auvergne, Clermont
Ferrand, France, 2Regional health-fund of®ce

France
for self-employed persons (Auvergne),
ChamalieÁres, France

Tubert-Jeannin S, Riordan PJ, Morel-Papernot A, Porcheray S, Saby-Collet S.

Validation of an oral health quality of life index (GOHAI) in France. Community
Dent Oral Epidemiol. 2003; 31: 275±84. ß Blackwell Munksgaard, 2003

Abstract ± Oral health has traditionally been de®ned in terms of disease. Today, health
is seen in a wider context: taking into account its impact on everyday living. Several
indices attempt to capture this dimension. The Geriatric Oral Health Assessment Index
(GOHAI) has been adapted for general use and tested on adult samples, mainly in
North America. Language, cultural norms and the health care system in France are
different and this raises the need to validate the GOHAI in France before it receives
widespread use. Objectives: The purpose of this study was to test the validity of a French
version of the GOHAI. Methods: The GOHAI is based on responses to a 12-item
self-administered questionnaire. The items were translated into French, back-translated
and compared with the original. After pilot testing and minor modi®cations, the
French version was administered to a group of low-income persons bene®ting from the
national health insurance system (n ˆ 260, 18±45 years). Measures for stability and
internal consistency were calculated. Concurrent and discriminant validity were
assessed. Results: Cronbach's a (0.86) showed a high internal consistency and Key words: access to dental care; GOHAI;
homogeneity between items. Item-scale correlations varied between 0.40 and 0.78. health insurance validation; oral health-related
Repeat administration of the GOHAI to 32 participants gave weighted kappa quality of life
coef®cients from 0.51 to 0.87 and a Pearson's correlation coef®cient of 0.87. Low SteÂphanie Tubert-Jeannin, DeÂpartement de
GOHAI scores were associated with perceptions of poor oral and general health, low Sante Publique,
UFR d'Odontologie, 11 Bd Charles de Gaulle,
satisfaction with oral health and a perceived need for dental care. There were signi®cant
63000 Clermont Ferrand, France
relationships between the GOHAI score and most objective measures of dental status Tel: ‡33 4 7317 7326
except FT. Younger, well-educated and higher income respondents were more likely to Fax: ‡33 4 7317 7309
have a high GOHAI score. Conclusion: The French version of the GOHAI exhibits e-mail: [email protected]
satisfactory psychometric properties but two items (one about swallowing, the other with Submitted 17 October 2001;
complex sentence structure) had poor stability. accepted 10 December 2002

The contemporary de®nition of (oral) health rejects social perspective, incorporation of patients' point of
the notion that health is equivalent to the absence of view, and qualitative measurements. Perception of
physical disease (1). It places an individual's experi- oral health depends on the subject's understanding of
ence of his health in the context of physical, psycho- what normal oral health is and of the speci®c symp-
logical and social well-being. The measurement of toms he may have experienced, of cultural values, of
oral health is historically derived from the disease- past experiences with the health care system, of
based model and oral diseases were measured with general health and of psycho-social well-being (2).
objective and quantitative indicators. Thus, the pre- Self-reported measures of impacts of health con-
valence of oral diseases has been described in various ditions (generic health status and disease speci®c)
samples of adults but less is known about how the on quality of life have expanded rapidly in the
diseases and symptoms affect adults' daily activities medical literature. Speci®c measures of Oral
and quality of life. Currently, conceptual models of Health-Related Quality of Life (OHRQL) are likely
oral health and instruments are focusing on a psycho- to be more sensitive than generic health status mea-

275
Tubert-Jeannin et al.

sures because oral health is perceived as a distinct by Atchison & Dolan (9). The GOHAI has been
dimension of overall Quality of Life (QOL; 3). Over validated and widely used in North America. It
the past 10 years, a number of investigators have has been shown that reliability (internal consistency)
developed and tested the performance of measures was satisfactory and all hypotheses designed to
designed to assess the functional, social and psycho- assess concurrent and construct validity were con-
social outcomes of oral disorders (4). These mea- ®rmed. The GOHAI, with only 12 items, is fairly
sures were reviewed at an international meeting in compact. A short questionnaire would be more
1996 (5) and descriptions of additional instruments likely to get complete, considered responses from
have been published since then (6). The majority of disadvantaged adults. A possible alternative to
these measures have been shown to have adequate GOHAI was the OHIP-14 (14) which has also been
reliability and validity but there are differences. widely used and validated. This index, however, has
Some are organised into subscales, others consist a bias towards more serious conditions (for example,
of a single scale. The number of items varies from 3 the extent to which oral health impacts on work) and
(7) to 49 (8). Another difference is whether the items thus, in a population in which there is little serious
constituting the measure are weighted (Oral Health impairment, responses will be skewed towards zero
Impact Pro®le, Dental Impact on Daily Living) or not impact. The GOHAI has been reported to be more
(GOHAI, Subjective Oral Health Status Indicator) successful than the OHIP-14 at detecting the impacts
(8±11). Some indicators are composed of a battery of of oral disorders in a population of younger, rela-
indicators while others consist of an index from tively healthy adults (15). The GOHAI was initially
which an overall score can be derived (12). designed to assess the oral health of older adults but
In 2000, a regional health fund in the Auvergne it has been used recently with disadvantaged His-
(France) decided to conduct a survey to evaluate the panic and African-American populations of all ages
oral health of a population of disadvantaged adults in the USA (16). The GOHAI, in consisting of a single
insured by the fund after the introduction of new index that incorporates different dimensions of oral
medical and dental coverage. It was decided to health, avoids problems connected with weighting,
collect information on oral health quality of life which introduces complications and subjectivity
because the health fund was interested in the impact (12).
of oral diseases on the overall comfort of this speci®c For all these reasons, it was decided to develop a
population. version of the GOHAI in the French language for use
The ®rst step required a choice of indicator of in France. The aim of this study was to test its
OHRQL. All the existing OHRQL indicators have validity.
initially been validated in English-speaking coun-
tries and no indicator of OHRQL has been validated
in French. Transferring such indicators from one
country to another presents problems at two levels.
Materials and methods
Direct translations may present linguistic problems The GOHAI
because some words and phrases have no direct The 12-item Geriatric Oral Health Assessment Index
translation and questions conceived in the context (GOHAI) was developed in the USA in 1990 (9) and
of one language may not be understood in the same was later renamed as the General Oral Health
way in the other language. There are many examples Assessment Index (17). It has been validated pri-
of this between French and English. Further, lan- marily in the United-States in an elderly Caucasian
guages exist within social and cultural frameworks sample and later in a mixed-age adult sample of
that are frequently unique and some questions Hispanics and African-Americans (9, 16). A Chinese
therefore may become different or meaningless in version has been published recently (18). The 12
a different culture and location. While one OHRQL items assessed the dimensions of physical functions
(the OHIP) has been translated into French for use in (eating, speaking and swallowing), psycho-social
Canada (13), the resulting questions in French functions (worry or concern about oral health, dis-
would not be well understood in France. It was, satisfaction with appearance, self-consciousness
therefore, decided to choose an indicator that has about oral health, avoidance of social contacts
been validated in English, and then to translate it because of oral problems), and pain or discomfort
into French and validate it in France. (use of medication to relieve pain, oral discomfort).
A decision was made to use the General Oral Subjects were asked if they have always, often,
Health Assessment Index (GOHAI), ®rst described sometimes, seldom or never experienced any of

276
 GOHAI validation in France

those problems in the previous 3 months. Questions islation. The insured mostly access care through
were worded sometimes in a positive direction, private dentists and pay on a fee-for-service basis,
sometimes in a negative one, to require respondents according to a ®xed schedule. Patients are partly
to consider their answers. Responses were scored on reimbursed by the fund. Insured persons can extend
a scale ranging from 1 to 5. When data were trans- their cover by subscribing to and paying for an
ferred to computer, the responses were re-coded, if optional additional plan. This cost can be a problem
necessary, so that responses indicating good condi- for disadvantaged persons. In 1999, a new law
tions and no problems carried the highest scores. (N899±641, 27/07/1999) to reduce the disadvantage
Thus, the scale score was a sum of values so that a of poorer people came into force. It guarantees total
low value indicates an oral health problem. A sum- reimbursement of health (including dental) care
mary score (Add-GOHAI) ranging from 12 to 60 was costs for low income earners, monthly income cur-
calculated for each subject with a higher score indi- rently below s535. In this report, disadvantaged
cating better oral health. A simple count score was adults are thus de®ned.
also calculated (SC-GOHAI) by counting the num- All the participants in this study (n ˆ 260) were
ber of items with responses `sometimes', `often' or people who had been called for medical or dental
`always (reversed for questions 3, 5 and 7; 15). examinations by one of these three health insurance
funds in the department of Puy de DoÃme. Partici-
The translation process and pilot study pants were selected in two ways. Some (n ˆ 84) were
The GOHAI was translated into French. We sought invited to participate while they were waiting for
to use straightforward colloquial French that would medical examination in the main examination centre
be easily understood by most people in France. The of the health funds in Clermont-Ferrand in the last
process involved translation from English to French quarter of 2000. Others (n ˆ 176) were invited to
by three bilingual individuals whose ®rst language attend for dental examination in connection with a
was French and then backward translation from different study (20). These groups, which were
French to English by bilingual individuals whose deemed to come from the same population, were
®rst language was English. Once the translations merged for the purposes of this report. Disadvan-
were completed, comparisons between the original taged adults were over-represented in the group and
English and back-translated versions and among the the actual participants were aged 18±45 years. For
three French versions were made (19). The transla- the purpose of this study, it was not necessary that
tion process involved compromises; we encountered the participants be a random sample.
examples of how some concepts are not easily ren-
dered into a different language. For example, Data collection
GOHAI item 3 (Table 1) uses in English the word The data for this report came from a self-adminis-
`comfortably' which we translated using `conforta- tered questionnaire and a clinical oral examination.
blement' but this is not a word people in France Besides the GOHAI items, the questionnaire included
would choose to describe the swallowing process; it socio-demographic data such as age, sex, educational
is more related to comfort as in how things (furni- level, marital status, employment, rural or urban resi-
ture, cars) are experienced rather than physiological dence and income. Subjects were also asked about
processes. We tried the word `facilement' (translates their perception of their general and oral health
as `easily') but ®nally settled for `confortablement' status, whether they were satis®ed with their dental
because, while it is not ideal, it seems less likely to be condition and their feelings about their need for
misunderstood. The ®nal French version was then dental treatment. All oral examinations were per-
pilot-tested on a sample of adults (n ˆ 90) after formed the same day as the questionnaire was admi-
which further minor language modi®cations were nistered either in dental or medical examination
made. For example, in Item 3, we omitted the sup- rooms in premises belonging to one of the insurance
plementary item `sans geÃne' that we had added to funds or, using portable lamps and equipment, in
make the meaning clearer, because it seemed to add general of®ces belonging to the same agencies. Dis-
a negative concept to a positive question. posable or prepackaged instruments were used, and
as the examinations took place in seven different sites
The population study in the department, portable lighting was used. The
The French health insurance system covers most of examinations, based on 28 teeth, used WHO criteria
the population. It operates through three indepen- (21) to register decayed, missing and ®lled teeth
dent health funds established under the same leg- which were the principal dental status measures.

277
Tubert-Jeannin et al.

Table 1. GOHAI items and frequency distribution of the responses

Ces trois derniers mois Jamais Rarement Parfois Souvent Toujours

In the past three months Never Seldom Sometimes Often always Total

1 ± Avez-vous limite la quantite ou le genre d'aliments 170 17 39 18 9 253

que vous mangez en raison de probleÁmes avec vos (67.2%) (6.7%) (15.4%) (7.1%) (3.6%)
dents ou vos appareils dentaires?/How often did you
limit the kinds or amounts of food you eat
because of problems with your teeth or denture?
2 ± Avez-vous eu des difficulteÂs pour mordre ou 128 29 53 24 20 254
mastiquer certains aliments durs tels que de la viande ou (50.4%) (11.4%) (20.9%) (9.4%) (7.9%)
une pomme?/How often did you have trouble
biting or chewing any kinds of food, such as a
firm meat or apples?
3 ± Avez-vous pu avaler confortablement?/How often 22 6 16 27 176 247
were you able to swallow comfortably? (8.9%) (2.4%) (6.5%) (10.9%) (71.3%)
4 ± Vos dents ou vos appareils dentaires vous ont-ils 196 14 24 7 9 250
empeÃcheÂ(e) de parler comme vous le vouliez?/How often (78.4%) (5.6%) (9.6%) (2.8%) (3.6%)
have your teeth or dentures prevented you from speaking
the way you wanted?
5 ± Avez-vous pu manger de tout (sans ressentir une 33 25 50 35 109 252
sensation d'incomfort)?/How often were you able to eat (13.1%) (9.9%) (19.8%) (13.9%) (43.3%)
anything without feeling discomfort?
6 ± Avez-vous limite vos contacts avec les gens aÁ cause 192 21 24 11 3 251
de l'eÂtat de vos dents ou de vos appareils dentaires?/ (76.5%) (8.3%) (9.6%) (4.4%) (1.2%)
How often did you limit contacts with people because
of the condition of your teeth or denture?
7 ± Avez-vous eÂte satisfait(e) ou content(e) de l'aspect 42 49 34 53 70 248
de vos dents, de vos gencives ou de vos appareils (16.9%) (19.8%) (13.7%) (21.4%) (28.2%)
dentaires?/How often were you pleased or happy with
the looks of your teeth, gums or dentures?
8 ± Avez vous pris un (des) meÂdicament(s) pour 98 38 72 39 5 252
soulager la douleur ou une sensation d'inconfort (38.9%) (15.1%) (28.6%) (15.5%) (1.9%)
dans votre bouche?/How often did you use medication
to relieve pain or discomfort from around
your mouth?
9 ± Vos probleÁmes de dent, de gencive ou d'appareil 58 48 73 48 26 253
dentaire vous ont-ils inquieÂteÂ(e) ou preÂoccupeÂ(e)?/How (22.9%) (19.0%) (28.8%) (19.0%) (10.3%)
often were you worried or concerned about the
problems with your teeth, gums or dentures?
10 ± Vous eÃtes vous senti(e) geÃneÂ(e) ou mal aÁ l'aise aÁ 112 32 45 44 19 252
cause de probleÁmes avec vos dents, vos gencives ou vos (44.4%) (12.7%) (17.9%) (17.5%) (7.5%)
appareils dentaires?/How often did you feel nervous or
self-conscious because of problems with your teeth, gums
or dentures?
11 ± Avez vous eÂprouve de l'embarras pour manger 173 24 29 16 12 254
devant les autres aÁ cause de probleÁmes avec vos dents (68.1%) (9.5%) (11.4%) (6.3%) (4.7%)
ou vos appareils dentaires?/How often did you feel
uncomfortable eating in front of people because of
problems with your teeth or dentures?
12 ± Vos dents ou vos gencives ont-elles eÂte sensibles 37 32 103 57 24 253
au froid, au chaud ou aux aliments sucreÂs?/How often (14.7%) (12.6%) (40.7%) (22.5%) (9.5%)
were your teeth or gums sensitive to hot, cold
or sweets?

A total of 254 questions.

Add-GOHAI score: mean ˆ 46.43 (9.46), minimum ˆ 17, maximum ˆ 60.
SC-GOHAI score: mean ˆ 4.46 (3.17), minimum ˆ 0, maximum ˆ 12.

278
 GOHAI validation in France

Examinations were conducted by ®ve experienced to report their oral health as fair or poor, more likely
dentists who had been calibrated in a 2-h discussion to report a self-perceived need for dental care and
session by experienced examiners. The presence of more likely to report dissatisfaction with their oral
removable prostheses was also recorded. health. Discriminant validity was evaluated by com-
paring the participants' item responses, and their
Ethics GOHAI scores, with their objectively assessed den-
Letters to participants were sent by the insurance tal status. It was hypothesised that there should be a
funds and names were not transmitted to those modest correlation between the GOHAI score and
conducting the examinations, to maximise the ano- the number of teeth present, the DMFT or the pre-
nymity of participants. The sheets on which data sence of removable prostheses. In addition, GOHAI
were registered at clinical examinations were iden- scores were compared with the socio-demographic
ti®ed by numbers, not names. The protocol for this variables.
study was submitted to the Comite National Informa-
tique et Liberte in Paris which gave permission and set
conditions for the collection of personal information
for the study.
Results
Characteristics of the subjects
Data analysis The total sample consisted of 260 subjects. Sixty-®ve
The general approach in this study involved an percent of the individuals were 30 years of age or
assessment of the reliability of the GOHAI measures older. Forty-nine percent of the subjects were
and an assessment of the principal aspects of the female. The subjects represented a disadvantaged
construct validity. group of adults with 76% of the individuals having
an income below s535 per month. Twenty-nine
Reliability percent of the subjects reported having completed
Reliability was assessed using the internal consis- high school education. Almost 60% of the respon-
tency approach. Pearson's correlation coef®cient dents lived in an urban area (population >5000
was used to measure inter-item and item-score cor- inhabitants). Forty-one percent of the subjects were
relation. Cronbach's a was calculated to assess the married or living together as a couple while the
degree of internal consistency and homogeneity others were single (with or without children).
between the items (22). Another aspect of reliability, Almost half (47%) of the sample was not employed.
stability, was assessed by repeating the administra-
tion of the GOHAI. On the ®rst occasion, partici- Acceptability and responses to the GOHAI
pants themselves completed the questionnaires on French version
the occasion of the clinical examination. Three weeks GOHAI items (in French and the original English
later, the same questionnaire was sent to a subset of versions) and distribution of the responses for this
these participants (n ˆ 40) and 32 of them (80%) sample are presented in Table 1. Six subjects did not
completed it again at home and returned it by post. complete the GOHAI questionnaire. For each item,
Stability was measured by using the Pearson's cor- the response rate was above 95%. The respondents
relation coef®cient for the cumulative GOHAI score reported some functional problems. Twenty-six
and the kappa score modi®ed by Fleiss for each of percent of the respondents limited (sometimes,
the 12 items (23). often or always) the kind or amounts of food eaten
(Q1), 38% had problems chewing foods such as ®rm
Validity meat or apples (Q2) and 42.8% had discomfort
Given that a pertinent criterion-related estimator of when eating (Q5). More than 80% of the respon-
prediction was not available nor was there a well- dents were always or often able to swallow com-
de®ned domain of content for determining validity, fortably (Q3) and to speak easily (Q4). The subjects
it was decided to use construct validation (24). also reported psycho-social concerns. Fifty-eight
Concurrent validity was evaluated by examining percent were worried or concerned (sometimes,
the association between the GOHAI score and the often or always) about their oral health (Q9),
global oral health rating questions. It was hypothe- 42.9% were nervous or self-conscious because of
sised that those reporting functional problems, psy- problems with their teeth or dentures (Q10), 15.2%
cho-social impacts, pain or discomfort and thus limited social contacts because of the condition of
having a low GOHAI score would be more likely their teeth or dentures (Q6) and 22.4% felt uncom-

279
Tubert-Jeannin et al.

Fig. 1. Distribution of the Add-GOHAI

scores (range 17±60, 11 categories)
among the 230 respondents who ad-
dressed each of the 12 items.

fortable eating in front of others (Q11). Only 50% of cient of 0.87 between the two successive GOHAI
the sample was always or often satis®ed with their scores. Concerning items 3 and 5, respectively, 4 and
oral appearance (Q7). Almost 73% of the subjects 5 respondents reversed completely their answers
reported (sometimes, often or always) sensitivities between the two administrations leading to a kappa
with their teeth (Q12) and 46% took medication to score of, respectively, 0.18 and 0.35.
relieve oral pain (Q8). Altogether 230 persons com-
pleted all the GOHAI questions and were examined Validity
clinically. The distribution of the GOHAI scores is Concurrent validity was evaluated by examining the
displayed in Figs 1 and 2. association between the GOHAI score and the global
(oral) health rating questions (Table 3). Low GOHAI
Reliability scores were associated with perceived fair or poor
Table 2 provides data on the reliability of the French oral health, perceived fair or poor general health,
version of the GOHAI. The Cronbach's a (0.86) low level of satisfaction with oral health and with the
showed a high degree of internal consistency and perception of dental care needs. GOHAI scores were
homogeneity between items. Item-scale correlations also compared among groups known to have differ-
varied between 0.40 and 0.78. Test±retest reliability ent levels of oral health (Table 3). It was demonstrated
was assessed by repeating the administration of the that mean GOHAI score decreased with ageing. No
GOHAI to 32 participants. For the 12 items (except difference of mean GOHAI score was observed
item 3 and 5), the weighted kappa coef®cients varied between sex, rural or urban residence or marital
from 0.51 to 0.87 with a Pearson's correlation coef®- status (Mann±Whitney, P > 0.05). Respondents with

Fig. 2. Distribution of the SC-GOHAI

scores (range 0±12, 13 categories)
among the 230 respondents who ad-
dressed each of the 12 items.

280
 GOHAI validation in France

Table 2. Inter-item and item-scale correlations for the GOHAI items (Add-GOHAI)

GOHAI item 1 2 3 4 5 6 7 8 9 10 11 12

Item-scale correlation 0.727 0.745 0.400 0.647 0.520 0.607 0.659 0.560 0.672 0.787 0.736 0.533

Cronbach's a ˆ 0.86.

a low level of education, a low income and those life outcomes in a group of French disadvantaged
who did not have a professional activity experienced adults aged 18±45 years. The GOHAI was originally
lower levels of GOHAI score. Data in Table 4 show developed and tested on samples of American, well-
that the GOHAI did not discriminate between parti- educated older adults (9). It has been shown that it
cipants according to the number of ®lled teeth. Sub- can be used in younger samples and in poorly
jects with a high GOHAI score had a higher number educated populations (16). Problems of validity
of natural teeth and a lower number of carious and stemming from cultural or language differences
missing teeth. Participants wearing one or more appear more complex and populations with differ-
removable prostheses had lower GOHAI scores than ent cultural backgrounds may respond differently to
participants who did not use a denture. the GOHAI items. Thus, it is important that the
GOHAI be tested in diverse populations in terms
of culture, language and geography. In this study,
the ®rst step consisted in using a standardised
Discussion translation process (19). Translation and back trans-
This study examined the ability of a French version lation were conducted to ensure the accuracy and
of the GOHAI to detect oral health-related quality of interpretability of the questions. This allowed the

Table 3. Concurrent validity of GOHAI scores and certain groups of questions

Variable
(standard deviation) Mean Add-GOHAI score Test

Self-rating of oral health

Poor, very poor n ˆ 49 38.06 (9.20) Kruskall±Wallis
Fairly poor: n ˆ 66 43.77 (8.57) P < 0.0001
Good: n ˆ 97 51.52 (6.51)
Very good-excellent: n ˆ 17 52.05 (6.27)
Self-rating of health
Poor, very poor n ˆ 14 36.14 (10.83) Kruskall±Wallis
Fairly poor: n ˆ 31 42.25 (7.93) P < 0.0001
Good: n ˆ 150 47.64 (8.50)
Very good-excellent: n ˆ 34 50.15 (9.18)
Satisfied with oral health
Yes n ˆ 105 51.99 (6. 44) Mann±Whitney
Not n ˆ 125 41.77 (9.08) P < 0.0001
Perception of dental care needs
Yes n ˆ 160 43.46 (9.30) Mann±Whitney
No n ˆ 70 53.22 (5.53) P < 0.0001
Age
20±29 years: n ˆ 54 51.18 (8.08) Mann±Whitney
30±45 years: n ˆ 171 44.80 (9.41) P < 0.001
Sex
Male n ˆ 103 46.24 (9.77) Mann±Whitney
Female n ˆ 127 46.59 (9.23) NS
Income
>s534/month n ˆ 54 48.87 (6.88) Mann±Whitney
<s534/month n ˆ 176 45.69 (10.02) P < 0.01
Professional situation
In work n ˆ 118 47.89 (9.11) Mann±Whitney
Not in work n ˆ 111 44.82 (9.61) P < 0.05
High school education
Completed n ˆ 66 49.28 (7.47) Mann±Whitney
Not completed n ˆ 156 45.45 (9.75) P < 0.01

281
Tubert-Jeannin et al.

Table 4. Discriminant validity of GOHAI-based on certain objective measures of oral health

Mean Mann±Whit-
Variable Comparison Add-GOHAI score (SD) ney test

Number of teeth <20 teeth (n ˆ 25) vs. 36.12 (8.58)

20 teeth (n ˆ 202) 47.81 (8.76) P < 0.001
Number of carious teeth DT < 2 (n ˆ 166) vs. 48.57 (8.36)
DT  2 (n ˆ 61) 40.95 (10.0) P < 0.001
Number of missing teeth MT < 4 (n ˆ 159) vs. 48.72 (8.46)
MT  4 (n ˆ 68) 41.38 (9.74) P < 0.001
Number of filled teeth FT < 10 (n ˆ 143) vs. 46.18 (10.24)
FT  10 (n ˆ 84) 47.10 (8.00) P > 0.050
Presence of a removable prosthesis Yes (n ˆ 21) 42.09 (9.55)
No (n ˆ 209) 46.87 (9.36) P < 0.050

creation of a French version which exhibits satisfac- perception of dental care needs was very high while
tory psychometric properties. the objective dental status was not different from the
As did Locker et al. (15), we found that the GOHAI French general population (27). This indicates that
was very successful at detecting oral disorders with the perception of oral health and the level of accep-
few participants having a very high score of 55±60. tance of oral conditions may vary according to the
This is because of the fact that the GOHAI gives a country and the socio-economic status irrespective
great weight to functional limitations and pain or of the objective dental status (18).
discomfort which are more immediate and more These results show that the French version of the
common outcomes of oral disorders in mixed age GOHAI exhibits satisfactory psychometric proper-
adults than psycho-social impacts. Nevertheless, it is ties. The analysis reported here indicates that the
noticeable that the proportion of subjects reporting French GOHAI demonstrates good internal consis-
problems to do with the ability to swallow comfor- tency. The Cronbach's a coef®cient was higher than
tably was very low. This has also been observed by values obtained in previous surveys which varied
Dolan (25) in a longitudinal survey on older adults. from 0.74 to 0.81 (9, 15, 18, 26). This good internal
The item `ability to swallow' presented other limita- consistency supports the use of the GOHAI in a
tions such as low item to scale and item to item variety of samples. Results concerning stability indi-
correlations and bad reproducibility. This item was cate good reproducibility concerning the global
originally developed to help measure the problems score of the GOHAI (r ˆ 0.87) as in the Chinese
people with xerostomia might encounter. Xerosto- version where 47 elderly persons were re-inter-
mia is much more common in older adults. It is viewed after 1 week (18). Calabrese et al. (28)
likely that dif®culty in swallowing would be less obtained a lower score (r ˆ 0.61) on 23 older adults.
relevant in a sample of younger individuals (16), so The subjects ®lled in the GOHAI administered by a
perhaps the inclusion of this item in the GOHAI dentist for the ®rst administration and 8 weeks later
score should be reconsidered. by a physician. A longer period between the two
Socio-economic data suggest that the sample was administrations and a change in the way the ques-
characterised by a low educational achievement and tionnaire was administered between the two ses-
low income. Self-rating of oral health was particu- sions might explain this lower correlation
larly poor and perception of dental care needs was coef®cient. We also tested the reproducibility of
very high indicating a substantial negative impact of the GOHAI for each item. The weighted kappa
oral conditions. This is in accordance with previous coef®cients observed in the present study were
®ndings showing that populations with lower socio- not satisfactory for two items (3, 5). In the Chinese
economic status experienced a greater negative version (18) those two items are negatively worded,
impact of oral conditions on functioning and well- leading to better reproducibility (k, 0.44, 0.42) but
being (26). It can be advocated that there might be a two other items (4, 10) appeared to be not easily
bias toward more perceived needs for treatment understandable in Chinese (k, 0.26, 0.33). This
given that the participants voluntarily attended a demonstrates the dif®culty of cultural adaptation
free dental screening, but it can be noticed that self- of the OHRQL indicators and suggests it may be
rating of oral health was particularly poor and necessary to explore reproducibility more carefully

282
 GOHAI validation in France

by conducting other test±retest procedures with a The French version of the GOHAI seems to pro-
greater number of subjects. vide valuable information about oral symptoms,
Regardless, it seems that poorly educated respon- psycho-social and functional problems in disadvan-
dents had some dif®culties in understanding the taged adults. Nevertheless, there may be a need to
direction of the answers, given that the 12 questions review some of the items used in the original
are alternatively positively or negatively formu- GOHAI questionnaire. The effect of wording some
lated. This problem of understanding was increased items in a positive format and others in a negative
by the fact that the questionnaire was self-adminis- format should be studied. It would be also interest-
tered, unlike the studies of Atchison et al. (16) on ing now to use the French version with other popu-
ethnic groups and Wong et al. (18), where the ques- lations such as older adults or people of different
tionnaire was administered by an interviewer. The social classes, to assess its generalisability, which at
problem of item 3 has been discussed previously. this stage is not established. Further research is also
The low kappa value observed for item 5 may be needed to examine the stability of the GOHAI over
explicable by the complex sentence structure time through the test±retest procedure. It would be
employed: the sentence has both a positive `able interesting to study the nature, the magnitude and
to eat anything' and a negative `without feeling the direction of change in oral health as measured by
discomfort' statement. Poorly educated subjects the GOHAI overtime. This has been done already
may have had some dif®culty in interpreting its among older English speaking adults (30) but
direction. We think that it would be useful to refor- further evaluation is needed. It is yet to be demon-
mulate this question. All hypotheses relating to strated that self-reported oral health outcomes and
concurrent validity were met. particularly the GOHAI (25) have the potential to
The literature contains examples of good correla- evaluate the appropriateness and effectiveness of
tion (4, 9, 18), and poor correlation (15) between the dental therapies. Thus, it would be of great interest
GOHAI and clinical observations. Except for the to develop the use of self-reported oral health out-
number of ®lled teeth, our study found statistically comes such as the French version of the GOHAI to
signi®cant relationships between the GOHAI and evaluate dental therapies or public health programs.
clinical measures (Table 4). Locker (15) reported
contrary ®ndings, in a population with a mean
age of 83 years. In our population of younger adults, Acknowledgements
the functional and psycho-social self-evaluation of
oral health seems to have been more highly in¯u- This study was conducted with the help of several organi-
sations: UFR d'Odontologie de Clermont-Ferrand, Union
enced by the actual clinical status. Except for a single reÂgionale des caisses d'assurance maladie Auvergne,
item concerning the self-rating of general health, the Caisse Maladie ReÂgionale des Professions IndeÂpendantes,
GOHAI was not compared with other measures of Mutualite Sociale Agricole du Puy de DoÃme, Caisse Pri-
general psychological well-being and general life maire d'Assurance Maladie du Puy de DoÃme, et Echelon
Local du Service Medical, ISBA (centre d'examen de santeÂ
satisfaction. Thus, it would be necessary in further de Clermont-Ferrand), and the Centre Hospitalier Uni-
research to explore more carefully this aspect of versitaire de Clermont-Ferrand. We would like to thank
construct validity. There was an important differ- the staffs of these organisations and all those who con-
ence in GOHAI scores for the two age groups. This is tributed to this survey.
consistent with previous ®ndings which note
declines in oral health with age (26), possibly
because of an increase in untreated disease. More- References
over, in the present study, subjects aged less than 25± 1. World Health Organisation. Definition of health.
30 years may have bene®ted from ¯uoride in tooth- http://www.who.int/aboutwho/en/definition.html.Accessed
paste during infancy while older subjects did not. 19 September 2001.
Fluoridated toothpastes have been freely sold in 2. Cruz GD, Galvis DL, Kim M, et al. Self-perceived oral
health among three subgroups of Asian-Americans in
France since 1975 (29). Income and employment also New York City: a preliminary study. Community
emerged as important factors in¯uencing the Dent Oral Epidemiol 2001;29:99±106.
GOHAI score, as found in previous studies (15, 3. Broder HL, Slade G, Caine R, Reisine S. Perceived
26). The differences in GOHAI scores observed impact of oral health conditions among minority
adolescents. J Public Health Dent 2000;60:189±92.
between groups well known for their different oral
4. Slade GD, editor. Measuring oral health and quality
health demonstrate the sensitivity of the GOHAI to of life. Chapel Hill: University of North Carolina,
within-sample differences. Dental Ecology 1997.

283
Tubert-Jeannin et al.

5. Kressin NR. Symposium on self-reported assessments health and quality of life. Chapel Hill: University of
of oral health outcomes. Introduction. J Dent Educ North Carolina, Department of Dental Ecology, 1997 .
1996;60:485±7. 18. Wong MC, Liu JK, Lo EC. Translation and validation
6. McGrath C, Bedi R, Gilthorpe MS. Oral health-related of the Chinese version of the GOHAI. J Public Health
quality of life ± views of the public in the United Dent, 2002;62:78±83.
Kingdom. Community Dent Health 2000;17:3±7. 19. Guillemin F, Bombardier C, Beaton D. Cross-cultural
7. Gooch BF, Dolan TA, Bourque LB. Correlates of self adaptation of health-related quality of life measures:
reported dental health status upon enrolment in the literature review and proposed guidelines. J Clin
Rand Health Insurance experiment. J Dent Educ Epidemiol 1993;46:1417±32.
1989;53:629±37. 20. Morel Papernot A, Moulin R, Tubert-Jeannin S, et al.
8. Slade GD, Spencer AJ. Development and evaluation Etat bucco-dentaire reÂel et percËu des adultes beÂneÂ-
of the Oral Health Impact Profile. Community Dent ficiaires de la CMU. Inf Dent 2003;85:133±41.
Health 1994;11:3±11. 21. World Health Organisation. Oral health surveys,
9. Atchison K, Dolan TA. Development of the geriatric indices and methods for measurement of dental
oral health assessment index. J Dent Educ 1990;54: diseases, 4th edn. GeneÁve: WHO 1997.
680±7. 22. Cronbach LJ. Coefficient alpha and the internal
10. Locker D, Miller Y. Evaluation of Subjective Oral reliability of tests. Psychometrica 1951;16:297±334.
Health Status Indicators. J Public Health Dent 1994; 23. Fleiss JL, Cohen J, Everitt BS. Large sample standard
54:167±76. errors of kappa and weighted kappa. Psychol Bull
11. Leao A, Sheiham A. Relation between clinical status 1969;72:323±7.
and subjective impacts on daily living. J Dent Res 24. Singleton RA, Straits BC. Approaches to social
1995;74:1408±13. research, 3rd edn. Oxford: Oxford University Press
12. Allen PF, Locker D. Do item weights matter? An 1999.
assessment using the Oral Health Impact Profile. 25. Dolan TA. The sensitivity of the Geriatric Oral Health
Community Dent Health 1997;14:133±8. Assessment Index to dental care. J Dent Educ
13. Allison P, Locker D, Jokovic A, Slade G. A cross- 1997;61:37±46.
cultural study of oral health values. J Dent Res 26. Kressin NR, Atchison KA, Miller DR. Comparing the
1999;78:643±9. impact of oral disease in two populations of older
14. Slade GD. Derivation and validation of a short form adults: application of the geriatric oral health assess-
Oral Health Impact Profile. Community Dent Oral ment index. J Public Health Dent 1997;57:224±32.
Epidemiol 1997;25:284±90. 27. Hescot P, Bourgeois D, Doury J. Oral health in 35±44
15. Locker D, Matear D, Stephen M, Lawrence H, Payne year old adults in France. Int Dent J 1997;47:94±9.
B. Comparison of the GOHAI and OHIP-14 as 28. Calabrese JM, Friedman PK, Rose LM, Jones JA.
measures of the oral health-related quality of life of Using the GOHAI to assess oral health status of frail
the elderly. Community Dent Oral Epidemiol 2001;29: homebound elders: reliability, sensitivity, and speci-
373±81. ficity. Spec Care Dent 1999;19:214±9.
16. Atchison KA, Der-Martirosian C, Gift H. Components 29. Tubert-Jeannin S. Comment expliquer la reÂduction de
of self-reported oral health and general health in la preÂvalence carieuse chez les jeunes FrancËais? Inf
racial and ethnic groups. J Public Health Dent Dent 1995;77:825±33.
1998;58:301±7. 30. Dolan TA, Peek CW, Stuck AE, Beck JC. Three-year
17. Atchison KA. The General Oral Health Assessment changes in global oral health rating by elderly dentate
Index (GOHAI). In: Slade GD, editor. Measuring oral adults. Community Dent Oral Epidemiol 1998;26:62±9.

284